BEGIN:VCALENDAR VERSION:2.0 PRODID:Linklings LLC BEGIN:VTIMEZONE TZID:Australia/Melbourne X-LIC-LOCATION:Australia/Melbourne BEGIN:DAYLIGHT TZOFFSETFROM:+1000 TZOFFSETTO:+1100 TZNAME:AEDT DTSTART:19721003T020000 RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU END:DAYLIGHT BEGIN:STANDARD DTSTART:19721003T020000 TZOFFSETFROM:+1100 TZOFFSETTO:+1000 TZNAME:AEST RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20240214T070310Z LOCATION:Darling Harbour Theatre\, Level 2 (Convention Centre) DTSTART;TZID=Australia/Melbourne:20231212T093000 DTEND;TZID=Australia/Melbourne:20231212T124500 UID:siggraphasia_SIGGRAPH Asia 2023_sess209@linklings.com SUMMARY:Technical Papers Fast-Forward DESCRIPTION:Technical Papers\n\nA 160 minutes preview session of all Techn ical Papers will also be held on the first day of the event where author(s ) of each paper get less than a minute to wow the attendees with a brief o verview of their work.\n\nAdaptNet: Policy Adaptation for Physics-Based Ch aracter Control\n\nMotivated by human’s ability to adapt skills in the lea rning of new ones, this paper presents AdaptNet, an approach for modifying the latent space of existing policies to allow new behaviors to be quickl y learned from like tasks in comparison to learning from scratch. Building on top of a give...\n\n\nPei Xu (Clemson University, Roblox); Kaixiang Xi e (McGill University); Sheldon Andrews (École de technologie supérieure, R oblox); Paul G. Kry (McGill University); Michael Neff (University of Calif ornia Davis); Morgan McGuire (Roblox, University of Waterloo); Ioannis Kar amouzas (University of California Riverside); and Victor Zordan (Roblox, C lemson University)\n---------------------\nLocally-Adaptive Level-of-Detai l for Hardware-Accelerated Ray Tracing\n\nWe introduce an adaptive level-o f-detail technique for ray tracing triangle meshes that aims to reduce the memory bandwidth used during ray traversal, which can be the bottleneck f or rendering time with large scenes and the primary consumer of energy. We propose a specific data structure for hierarc...\n\n\nJacob Haydel (Unive rsity of Utah); Cem Yuksel (University of Utah, Roblox); and Larry Seiler (Independent)\n---------------------\nA Locality-based Neural Solver for O ptical Motion Capture\n\nWe present a novel locality-based learning method for cleaning and solving optical motion capture data. Given noisy marker data, we propose a new heterogeneous graph neural network which treats mar kers and joints as different types of nodes, and uses graph convolution op erations to extract the local...\n\n\nXiaoyu Pan and Bowen Zheng (State Ke y Laboratory of CAD & CG, Zhejiang University; ZJU-Tencent Game and Intell igent Graphics Innovation Technology Joint Lab); Xinwei Jiang, Guanglong X u, Xianli Gu, and Jingxiang Li (Tencent Games Digital Content Technology C enter); Qilong Kou (Tencent Technology (Shenzhen) Co., LTD); He Wang (Univ ersity College London (UCL)); Tianjia Shao and Kun Zhou (State Key Laborat ory of CAD & CG, Zhejiang University); and Xiaogang Jin (State Key Laborat ory of CAD & CG, Zhejiang University; ZJU-Tencent Game and Intelligent Gra phics Innovation Technology Joint Lab)\n---------------------\nNeural Pack ing: from Visual Sensing to Reinforcement Learning\n\nWe present a novel l earning framework to solve the transport-and-packing (TAP) problem in 3D. It constitutes a full solution pipeline from partial observations of input objects via RGBD sensing and recognition to final box placement, via robo tic motion planning, to arrive at a compact packing in a t...\n\n\nJuzhan Xu (Shenzhen University), Minglun Gong (University of Guelph), Hao Zhang ( Simon Fraser University), and Hui Huang and Ruizhen Hu (Shenzhen Universit y)\n---------------------\nNeural Categorical Priors for Physics-Based Cha racter Control\n\nRecent advances in learning reusable motion priors have demonstrated their effectiveness in generating naturalistic behaviors. In this paper, we propose a new learning framework in this paradigm for contr olling physics-based characters with significantly improved motion quality and diversity over ex...\n\n\nQingxu Zhu, He Zhang, Mengting Lan, and Lei Han (Tencent)\n---------------------\nLitNeRF: Intrinsic Radiance Decompo sition for High-Quality View Synthesis and Relighting of Faces\n\nHigh-fid elity, photorealistic 3D capture of a human face is a long-standing proble m in computer graphics -- the complex material of skin, intricate geometry of hair, and fine scale textural details make it challenging. Traditional techniques rely on very large and expensive capture rigs to reconstru...\ n\n\nKripasindhu Sarkar (Google Inc.); Marcel Bühler and Simon Li (ETH Zür ich, Google Inc.); and Daoye Wang, Delio Vicini, Jérémy Riviere, Yinda Zha ng, Sergio Orts-Escolano, Paulo Gotardo, Thabo Beeler, and Abhimitra Meka (Google Inc.)\n---------------------\nInteraction-Driven Active 3D Reconst ruction with Object Interiors\n\nWe introduce an active 3D reconstruction method, which integrates visual perception, robot-object interaction, and 3D scanning to recover both the exterior and interior geometries of a targ et 3D object. Unlike other works in active vision which focus on optimizin g camera viewpoints to better investi...\n\n\nZihao Yan, Fubao Su, Mingyan g Wang, and Ruizhen Hu (Shenzhen University); Hao Zhang (Simon Fraser Univ ersity); and Hui Huang (Shenzhen University)\n---------------------\nMIPS- Fusion: Multi-Implicit-Submaps for Scalable and Robust Online Neural RGB-D Reconstruction\n\nWe introduce MIPS-Fusion, a robust and scalable online RGB-D reconstruction method based on a novel neural implicit representatio n -- multi-implicit-submap. Different from existing neural RGB-D reconstru ction methods lacking either flexibility with a single neural map or scala bility due to extra sto...\n\n\nYijie Tang (National University of Defense Technology (NUDT)), Jiazhao Zhang (Peking University), Zhinan Yu (Nationa l University of Defense Technology (NUDT)), He Wang (Peking University), a nd Kai Xu (National University of Defense Technology (NUDT))\n------------ ---------\nDifferentiable Rendering of Parametric Geometry\n\nWe propose a n efficient method for differentiable rendering of parametric surfaces and curves, which enables their use in inverse graphics problems. Our central observation is that a representative triangle mesh can be extracted from a continuous parametric object in a differentiable and efficient w...\n\n\ nMarkus Worchel and Marc Alexa (TU Berlin)\n---------------------\nEfficie nt Cone Singularity Construction for Conformal Parameterizations\n\nWe pro pose an efficient method to construct sparse cone singularities under dist ortion-bounded constraints for conformal parameterizations. Central to our algorithm is using the technique of shape derivatives to move cones for d istortion reduction without changing the number of cones. In particular,.. .\n\n\nMo Li, Qing Fang, Zheng Zhang, Ligang Liu, and Xiao-Ming Fu (Univer sity of Science and Technology of China)\n---------------------\nTranspare nt Object Reconstruction via Implicit Differentiable Refraction Rendering\ n\nReconstructing the geometry of transparent objects has been a long-stan ding challenge. Existing methods rely on complex setups, such as manual an notation or darkroom conditions, to obtain object silhouettes and usually require controlled environments with designed patterns to infer ray-backgr ound co...\n\n\nFangzhou Gao, Lianghao Zhang, Li Wang, Jiamin Cheng, and J iawan Zhang (Tianjin University)\n---------------------\nShaDDR: Interacti ve Example-Based Geometry and Texture Generation via 3D Shape Detailizatio n and Differentiable Rendering\n\nWe present ShaDDR, an example-based deep generative neural network which produces a high-resolution textured 3D sh ape through geometry detailization and conditional texture generation appl ied to an input coarse voxel shape. Trained on a small set of detailed and textured exemplar shapes, our method ...\n\n\nQimin Chen, Zhiqin Chen, Ha ng Zhou, and Hao Zhang (Simon Fraser University)\n---------------------\nR MIP: Displacement ray tracing via inversion and oblong bounding\n\nHigh-pe rformance ray tracing of triangle meshes equipped with displacement maps i s a challenging task. Existing methods either rely on pre-tessellation, ta king full advantage of the hardware but with a poor memory quality tradeof f, or use custom displacement-centric acceleration structures, preservi... \n\n\nTheo Thonat, Iliyan Georgiev, François Beaune, and Tamy Boubekeur (A dobe)\n---------------------\nMetaLayer: A Meta-learned BSDF Model for Lay ered Materials\n\nReproducing the appearance of arbitrary layered material s has long been a critical challenge in computer graphics, with regard to the demanding requirements of both physical accuracy and low computation c ost. Recent studies have demonstrated promising results by learning-based representations that i...\n\n\nJie Guo, Zeru Li, and Xueyan He (Nanjing Un iversity); Beibei Wang (Nankai University, Nanjing University of Science a nd Technology); Wenbin Li and Yanwen Guo (Nanjing University); and Ling-Qi Yan (University of California, Santa Barbara)\n---------------------\nHyp erDreamer: Hyper-Realistic 3D Content Generation and Editing from a Single Image\n\n3D content creation from a single image is a long-standing yet h ighly desirable task. Recent advances introduce 2D diffusion priors, yield ing reasonable results. However, existing methods are not hyper-realistic enough for post-generation usage, as users cannot view, render and edit th e resulting 3D...\n\n\nTong Wu and Zhibing Li (The Chinese University of H ong Kong, Shanghai AI Laboratory); Shuai Yang (Shanghai Jiao Tong Universi ty, Shanghai AI Laboratory); Pan Zhang (Shanghai AI Laboratory); Xingang P an (Max Planck Institute for Informatics); Jiaqi Wang (Shanghai AI Laborat ory); Dahua Lin (The Chinese University of Hong Kong, Shanghai AI Laborato ry); and Ziwei Liu (Nanyang Technological University)\n------------------- --\nContent-based Search for Deep Generative Models\n\nThe growing prolife ration of customized and pretrained generative models has made it infeasib le for a user to be fully cognizant of every model in existence. To addres s this need, we introduce the task of content-based model search: given a query and a large set of generative models, finding the mod...\n\n\nDaohan Lu, Sheng-Yu Wang, Nupur Kumari, Rohan Agarwal, and Mia Tang (Carnegie Me llon University); David Bau (Northeastern University); and Jun-Yan Zhu (Ca rnegie Mellon University)\n---------------------\nLiCROM: Linear-Subspace Continuous Reduced Order Modeling with Neural Fields\n\nLinear reduced-ord er modeling (ROM) simplifies complex simulations by approximating the beha vior of a system using a simplified kinematic representation. Typically, R OM\nis trained on input simulations created with a specific spatial discre tization, \nand then serves to accelerate simulations with the...\n\n\nYue Chang (University of Toronto), Peter Yichen Chen (MIT CSAIL), Zhecheng Wa ng (University of Toronto), Maurizio M. Chiaramonte and Kevin Carlberg (Me ta Reality Labs Research), and Eitan Grinspun (University of Toronto)\n--- ------------------\nExample-Based Sampling with Diffusion Models\n\nMuch e ffort has been put into developing samplers with specific properties, such as producing blue noise, low-discrepancy, lattice or Poisson disk samples . These samplers can be slow if they rely on optimization processes, may r ely on a wide range of numerical methods, are not always differentiable... .\n\n\nBastien Doignies (Université Claude Bernard Lyon, CNRS); Nicolas Bo nneel, David Coeurjolly, and Julie Digne (CNRS, LIRIS); Loïs Paulin (Unive rsité Claude Bernard Lyon / CNRS, Adobe); and Jean-Claude Iehl and Victor Ostromoukhov (Université Claude Bernard Lyon, CNRS)\n--------------------- \nEfficient Graphics Representation with Differentiable Indirection\n\nWe introduce differentiable indirection -- a novel learned primitive that emp loys differentiable multi-scale lookup tables as an effective substitute f or traditional compute and data operations across the graphics pipeline. W e demonstrate its flexibility on a number of graphics tasks, i.e., geometr i...\n\n\nSayantan Datta (McGill University, Meta Reality Labs); Carl Mars hall (Meta); Derek Nowrouzezahrai (McGill University, Meta); and Zhao Dong and Zhengqin Li (Meta)\n---------------------\nClose the Design-to-Manufa cturing Gap in Computational Optics with a ’Real2Sim’ Learned Two-Photon N eural Lithography Simulator\n\nWe introduce neural lithography to address the ‘design-to-manufacturing’ gap in computational optics. Computational o ptics with large design degrees of freedom enable advanced functionalities and performance beyond traditional optics. However, the existing design a pproaches often overloo...\n\n\nCheng Zheng (MIT), Guangyuan Zhao (The Chi nese University of Hong Kong), and Peter So (MIT)\n---------------------\n RT-Octree: Accelerate PlenOctree Rendering with Batched Regular Tracking a nd Neural Denoising for Real-time Neural Radiance Fields\n\nNeural Radianc e Fields (NeRF) has demonstrated its ability to generate high-quality synt hesized views. Nonetheless, due to its slow inference speed, there is a ne ed to explore faster inference methods. In this paper, we propose RT-Octre e, which uses batched regular tracking based on PlenOctree with ...\n\n\nZ ixi Shu, Ran Yi, Yuqi Meng, Yutong Wu, and Lizhuang Ma (Shanghai Jiao Tong University)\n---------------------\nPerceptual error optimization for Mon te Carlo animation rendering\n\nIndependently estimating individual pixel values in Monte Carlo rendering results in a perceptually sub-optimal whit e-noise distribution of error in image space. Recent works have shown that perceptual fidelity can be improved significantly by distributing pixel e rror as blue noise instead. Most suc...\n\n\nMiša Korać (Saarland Universi ty, DFKI); Corentin Salaün (Max Planck Institute for Informatics); Iliyan Georgiev (Adobe); Pascal Grittmann (Saarland University); Philipp Slusalle k (Saarland University, DFKI); and Karol Myszkowski and Gurprit Singh (Max Planck Institute for Informatics)\n---------------------\nText-Guided Vec tor Graphics Customization\n\nVector graphics are widely used in digital a rt and valued by designers for their scalability and layer-wise topologica l properties. However, the creation and editing of vector graphics necessi tate creativity and design expertise, leading to a time-consuming process. In this paper, we propose a novel...\n\n\nPeiying Zhang (City University of Hong Kong), Nanxuan Zhao (Adobe Research), and Jing Liao (City Universi ty of Hong Kong)\n---------------------\nA Micrograin BSDF Model for the R endering of Porous Layers\n\nWe introduce a new BSDF model for the renderi ng of porous layers, as found on surfaces covered by dust, rust, dirt, or sprayed paint. \nOur approach is based on a distribution of elliptical opa que micrograins, extending the Trowbridge-Reitz (GGX) distribution [Trowbr idge1975,Walter2007] to handle por...\n\n\nSimon Lucas (Université de Bord eaux, INRIA); Mickael Ribardiere (Université de Poitiers); and Romain Paca nowski and Pascal Barla (INRIA)\n---------------------\nManifold Path Guid ing for Importance Sampling Specular Chains\n\nComplex visual effects such as caustics are often produced by light paths containing multiple consecu tive specular vertices (dubbed specular chains), which pose a challenge to unbiased estimation in Monte Carlo rendering.\nIn this work, we study the light transport behavior within a sub-path that is ...\n\n\nZhimin Fan (N anjing University); Pengpei Hong (University of Utah); Jie Guo (Nanjing Un iversity); Changqing Zou (Zhejiang Lab; State Key Lab of CAD&CG, Zhejiang University); Yanwen Guo (Nanjing University); and Ling-Qi Yan (University of California, Santa Barbara)\n---------------------\nSimpleNeRF: Regulari zing Sparse Input Neural Radiance Fields with Simpler Solutions\n\nNeural Radiance Fields (NeRF) show impressive performance for the photo-realistic free-view rendering of scenes. However, NeRFs require dense sampling of i mages in the given scene, and their performance degrades significantly whe n only a sparse set of views are available. Researchers have found that... \n\n\nNagabhushan Somraj, Adithyan Karanayil, and Rajiv Soundararajan (Ind ian Institute of Science)\n---------------------\nMultiple-bounce Smith Mi crofacet BRDFs using the Invariance Principle\n\nSmith microfacet models a re widely used in computer graphics to represent materials. Traditional mi crofacet models do not consider the multiple bounces on microgeometries, l eading to visible energy missing, especially on rough surfaces. Later, as the equivalence between the microfacets and volume ha...\n\n\nYuang Cui (A nhui Science and Technology University); Gaole Pan and Jian Yang (Nanjing University of Science and Technology); Lei Zhang (The Hong Kong Polytechni c University); Ling-Qi Yan (University of California, University of Califo rnia Santa Barbara); and Beibei Wang (Nankai University, Nanjing Universit y of Science and Technology)\n---------------------\nGarmentCode: Programm ing Parametric Sewing Patterns\n\nGarment modeling is an essential task of the global apparel industry and a core part of digital human modeling. Re alistic representation of garments with valid sewing patterns is key to th eir accurate digital simulation and eventual fabrication. \nHowever, littl e-to-no computational tools provide sup...\n\n\nMaria Korosteleva and Olga Sorkine-Hornung (ETH Zurich)\n---------------------\nC-shells: Deployable Gridshells with Curved Beams\n\nWe introduce a computational pipeline for simulating and designing C-shells, a new class of planar-to-spatial deplo yable linkage structures. A C-shell is composed of curved flexible beams c onnected at rotational joints that can be assembled in a stress-free plana r configuration. When actuated, the e...\n\n\nQuentin Becker, Seiichi Suzu ki, and Yingying Ren (EPFL); Davide Pellis (ISTI - CNR); Julian Panetta (U niversity of California Davis); and Mark Pauly (EPFL)\n------------------- --\nHigh-Fidelity and Real-Time Novel View Synthesis for Dynamic Scenes\n\ nThis paper aims to tackle the challenge of dynamic view synthesis from mu lti-view videos. The key observation is that while previous grid-based met hods offer consistent rendering, they fall short in capturing appearance d etails on a complex dynamic scene, a domain where multi-view image-based m ethod...\n\n\nHaotong Lin (State Key Laboratory of CAD & CG, Zhejiang Univ ersity); Sida Peng (Zhejiang University); and Zhen Xu, Tao Xie, Xingyi He, Hujun Bao, and Xiaowei Zhou (State Key Laboratory of CAD & CG, Zhejiang U niversity)\n---------------------\nBakedAvatar: Baking Neural Fields for R eal-Time Head Avatar Synthesis\n\nSynthesizing photorealistic 4D human hea d avatars from videos is essential for VR/AR, telepresence, and video game applications. Although existing Neural Radiance Fields (NeRF)-based metho ds achieve high-fidelity results, the computational expense limits their u se in real-time applications. To overc...\n\n\nHao-Bin Duan (Beihang Unive rsity); Miao Wang (Beihang University, Zhongguancun Laboratory); Jin-Chuan Shi and Xu-Chuan Chen (Beihang University); and Yan-Pei Cao (Tencent)\n-- -------------------\nThin On-Sensor Nanophotonic Array Cameras\n\nToday's commodity camera systems rely on compound optical systems to map light ori ginating from the scene to positions on the sensor where it gets recorded as an image. To achieve an accurate mapping without optical aberrations, i .e., deviations from Gauss' linear optics model, typical lens systems ...\ n\n\nPraneeth Chakravarthula (Princeton University); Jipeng Sun (Princeton University, Northwestern University); Xiao Li, Chenyang Lei, Gene Chou, a nd Mario Bijelic (Princeton University); Johannes Froesch and Arka Majumda r (University of Washington); and Felix Heide (Princeton University)\n---- -----------------\nPerceptual Requirements for World-Locked Rendering in A R and VR\n\nStereoscopic, head-tracked display systems can show users real istic, world-locked virtual objects and environments. However, discrepanci es between the rendering pipeline and physical viewing conditions can lead to perceived instability in the rendered content resulting in reduced rea lism, immersion,...\n\n\nPhillip Guan, Eric Penner, Joel Hegland, Benjamin Letham, and Douglas Lanman (Meta Reality Labs Research)\n---------------- -----\nPower Plastics: A Hybrid Lagrangian/Eulerian Solver for Mesoscale I nelastic Flows\n\nWe propose a novel hybrid Lagrangian/Eulerian method for simulating inelastic materials that generates high-quality particle distr ibutions with strict volume control. At its core, our approach integrates an updated Lagrangian time discretization of continuum mechanics with the Power Particle-In-Cell...\n\n\nZiyin Qu (University of Pennsylvania, Unive rsity of California Los Angeles); Minchen Li (University of California Los Angeles, Carnegie Mellon University); Yin Yang (University of Utah); Chen fanfu Jiang (University of California Los Angeles); and Fernando de Goes ( Pixar Animation Studios)\n---------------------\nSubspace-Preconditioned G PU Projective Dynamics with Contact for Cloth Simulation\n\nWe propose an efficient cloth simulation method that combines the merits of two drastica lly different numerical procedures, namely the subspace integration and pa rallelizable iterative relaxation. We show those two methods can be organi cally coupled within the framework of projective dynamics (PD), ...\n\n\nX uan Li (UCLA); Yu Fang (UCLA, University of Pennsylvania); Lei Lan (Univer sity of Utah); Huamin Wang (Style 3D Research); Yin Yang (University of Ut ah, Style 3D Research); Minchen Li (UCLA, Carnegie Mellon University); and Chenfanfu Jiang (UCLA, Style 3D Research)\n---------------------\nSeamles sNeRF: Stitching Part NeRFs with Gradient Propagation\n\nNeural Radiance F ields (NeRFs) have emerged as a promising representation for 3D scenes, sp arking a surge in research aimed at extending the editing capabilities in this domain. The task of seamless editing and merging of different NeRFs, similar to the "copy-and-paste" function in 2D image editing,...\n\n\nBing chen Gong and Yuehao Wang (The Chinese University of Hong Kong); Xiaoguang Han (Shenzhen Research Institute of Big Data, the Chinese University of H ong Kong (Shenzhen)); and Qi Dou (The Chinese University of Hong Kong)\n-- -------------------\nThe Shortest Route Is Not Always the Fastest: Probabi lity-Modeled Stereoscopic Eye Movement Completion Time in VR\n\nSpeed and consistency of target-shifting play a crucial role in human ability to per form complex tasks. Shifting our gaze between objects of interest quickly and consistently requires changes both in depth and direction. Gaze change s in depth are driven by slow, inconsistent vergence movements which...\n\ n\nBudmonde Duinkharjav and Benjamin Liang (New York University), Anjul Pa tney and Rachel Brown (NVIDIA Research), and Qi Sun (New York University)\ n---------------------\nDevelopable Quad Meshes and Contact Element Nets\n \nThe property of a surface being developable can be expressed in differen t equivalent ways, by vanishing Gauss curvature, or by the existence of is ometric mappings to planar domains. Computational contributions to this to pic range from special parametrizations to discrete-isometric mappings. Ho wever,...\n\n\nVictor Ceballos Inza and Florian Rist (KAUST), Johannes Wal lner (TU Graz), and Helmut Pottmann (KAUST)\n---------------------\nGeoLat ent: A Geometric Approach to Latent Space Design for Deformable Shape Gene rators\n\nWe study how to optimize the latent space of neural shape genera tors that map latent codes to 3D deformable shapes. The key focus is to lo ok at a deformable shape generator from a differential geometry perspectiv e. We define a Riemannian metric based on as-rigid-as-possible and as-conf ormal-as-possi...\n\n\nHaitao Yang, Bo Sun, Liyan Chen, Amy Pavel, and Qix ing Huang (University of Texas at Austin)\n---------------------\nOnline S cene CAD Recomposition via Autonomous Scanning\n\nAutonomous surface recon struction of 3D scenes has been intensely studied in recent years, however , it is still difficult to accurately reconstruct all the surface details of complex scenes with complicated object relations and severe occlusions, which makes the reconstruction results not suitable f...\n\n\nChanghao Li and Junfu Guo (University of Science and Technology of China), Ruizhen Hu (Shenzhen University), and Ligang Liu (University of Science and Technolo gy of China)\n---------------------\nA Parametric Kinetic Solver for Simul ating Boundary-Dominated Turbulent Flow Phenomena\n\nBoundary layer flow p lays a very important role in shaping the entire flow feature near and beh ind obstacles inside fluids. Thus, boundary treatment methods are crucial for a physically consistent fluid simulation, especially when turbulence o ccurs at a high Reynolds number, in which accurately hand...\n\n\nMengyun Liu and Xiaopei Liu (ShanghaiTech University)\n---------------------\nHand Pose Estimation with Mems-Ultrasonic Sensors\n\nHand tracking is an impor tant aspect of human-computer interaction and has a wide range of applicat ions in extended reality devices. However, current hand motion capture met hods suffer from various limitations. For instance, visual-based hand pose estimation is susceptible to self-occlusion and chan...\n\n\nQiang Zhang, Yuanqiao Lin, Yubin Lin, and Szymon Rusinkiewicz (Princeton University)\n ---------------------\nRepurposing Diffusion Inpainters for Novel View Syn thesis\n\nIn this paper, we present a method for generating consistent nov el views from a single source image. Our approach focuses on maximizing th e reuse of visible pixels from the source view. To achieve this, we use a monocular depth estimator that transfers visible pixels from the source vi ew to the targ...\n\n\nYash Kant (University of Toronto, Snap Inc.); Aliak sandr Siarohin, Michael Vasilkovsky, Riza Alp Guler, Jian Ren, and Sergey Tulyakov (Snap Inc.); and Igor Gilitschenski (University of Toronto)\n---- -----------------\nInput-Dependent Uncorrelated Weighting for Monte Carlo Denoising\n\nImage-space denoising techniques have been widely employed in Monte Carlo rendering, typically blending neighboring pixel estimates usi ng a denoising kernel. It is widely recognized that a kernel should be ada pted to characteristics of the input pixel estimates in order to ensure ro bustness to diver...\n\n\nJonghee Back (Gwangju Institute of Science and T echnology), Binh-Son Hua (Trinity College Dublin), Toshiya Hachisuka (Univ ersity of Waterloo), and Bochang Moon (Gwangju Institute of Science and Te chnology)\n---------------------\nNeural Cache for Monte Carlo Partial Dif ferential Equation Solver\n\nThis paper presents a method that uses neural networks as a caching mechanism to reduce the variance of Monte Carlo Par tial Differential Equation solvers, such as the Walk-on-Spheres algorithm. While these Monte Carlo PDE solvers have the merits of being unbiased and discretization-free, their high ...\n\n\nZilu Li (Cornell University); Gu andao Yang (Cornell University, Stanford University); and Xi Deng, Christo pher De Sa, Bharath Hariharan, and Steve Marschner (Cornell University)\n- --------------------\nAn Implicit Physical Face Model Driven by Expression and Style\n\n3D facial animation is often produced by manipulating facial deformation models (or rigs), that are traditionally parameterized by exp ression controls. A key component that is usually overlooked is expression ``style", as in, how a particular expression is performed. Although it is common to define ...\n\n\nLingchen Yang (ETH Zürich); Gaspard Zoss and Pr ashanth Chandran (The Walt Disney Company (Switzerland) GmbH); Paulo Gotar do (Disney Research Studios, The Walt Disney Company (Switzerland) GmbH); Markus Gross (ETH Zürich, The Walt Disney Company (Switzerland) GmbH); Bar bara Solenthaler (ETH Zürich); Eftychios Sifakis (University of Wisconsin Madison); and Derek Bradley (The Walt Disney Company (Switzerland) GmbH)\n ---------------------\nToRoS: A Topology Optimization Approach for Designi ng Robotic Skins\n\nSoft robotics offers unique advantages in manipulating fragile or deformable\nobjects, human-robot interaction, and exploring in accessible terrain. How-\never, designing soft robots that produce large, targeted deformations is\nchallenging. In this paper, we propose a new met hodology for designing\nsoft...\n\n\nJuan Sebastian Montes Maestre, Ronan Hinchet, Stelian Coros, and Bernhard Thomaszewski (ETH Zürich)\n---------- -----------\nEfficient Visualization of Light Pollution for the Night Sky\ n\nThe artificial light sources make our daily life convenient, but they c ause a serious problem called light pollution. \nWe propose a system for e fficient visualization of the light pollution for the night sky.\nA number of methods have been proposed for rendering the sky, but most of the meth ods focus...\n\n\nYoshinori Dobashi and Naoto Ishikawa (Hokkaido Universit y, Prometech CG Research) and Kei Iwasaki (Saitama University, Prometech C G Research)\n---------------------\nComputational Design of Flexible Plana r Microstructures\n\nMechanical metamaterials enable customizing the elast ic properties of physical objects by altering their fine-scale structure. A broad gamut of effective material properties can be produced even from a single fabrication material by optimizing the geometry of a periodic micr ostructure tiling. Past w...\n\n\nZhan Zhang (University of California Dav is); Christopher Brandt (1000shapes GmbH); Jean Jouve (University Grenoble Alpes Inria, CNRS, Grenoble INP, LJK); Yue Wang and Tian Chen (University of Houston); Mark Pauly (Ecole Polytechnique Fédérale de Lausanne); and J ulian Panetta (University of California Davis)\n---------------------\nNeu ral Point-based Volumetric Avatar: Surface-guided Neural Points for Effici ent and Photorealistic Volumetric Head Avatar\n\nRendering photo-realistic and vividly moving human heads is very important for pleasant and immersi ve experience in AR/VR and video conferencing. However, existing methods u sually struggle to model challenging facial regions (e.g., mouth interior, eyes, hair/beard), resulting in unrealistic and blur...\n\n\nCong Wang (T singhua University); Di Kang, Yan-Pei Cao, Linchao Bao, and Ying Shan (Ten cent); and Song-Hai Zhang (Tsinghua University)\n---------------------\nDR -Occluder: Generating Occluders using Differentiable Rendering\n\nThe targ et of the occluder is to use very few faces to maintain similar occlusion properties of the original 3D model.\nIn this paper, we present DR-Occlude r, a novel coarse-to-fine framework for occluder generation that leverages differentiable rendering to optimize a triangle set to an occluder. Un... \n\n\nJiaxian Wu, Yue Lin, and Dehui Lu (NetEase Games AI Lab)\n---------- -----------\nScene-aware Activity Program Generation with Language Guidanc e\n\nWe address the problem of scene-aware activity program generation, wh ich requires decomposing a given activity task into instructions that can be sequentially performed within a target scene to complete the activity. While existing methods have shown the ability to generate rational or exec utable pr...\n\n\nZejia Su (Shenzhen University), Qingnan Fan (Vivo), Xuel in Chen (Tencent AI Lab), Oliver van Kaick (Carleton University), and Hui Huang and Ruizhen Hu (Shenzhen University)\n---------------------\nHigh-Re solution Volumetric Reconstruction for Clothed Humans\n\nWe present a nove l method for reconstructing clothed humans from a sparse set of, e.g., 1-6 RGB images. We revisit the volumetric approach and demonstrate that bette r performance can be achieved with proper system design. The volumetric re presentation offers significant advantages in leveraging 3D s...\n\n\nSico ng Tang (Simon Fraser University); Guangyuan Wang, Qing Ran, Lingzhi Li, a nd Li Shen (Alibaba); and Ping Tan (Simon Fraser University)\n------------ ---------\nMOCHA: Real-Time Motion Characterization via Context Matching\n \nTransforming neutral, characterless input motions to embody the distinct style of a notable character in real time is highly compelling for charac ter animation. This paper introduces MOCHA, a novel online motion characte rization framework that transfers both motion styles and body proportions from a...\n\n\nDeok-Kyeong Jang (KAIST, MOVIN Inc.); Yuting Ye (Meta); Jun gdam Won (Seoul National University); and Sung-Hee Lee (KAIST)\n---------- -----------\nAnalysis and Synthesis of Digital Dyadic Sequences\n\nWe expl ore the space of matrix-generated $(0, m, 2)$-nets and $(0, 2)$-sequences in base 2, also known as digital dyadic nets and sequences.\nIn computer g raphics, they are arguably leading the competition for use in rendering.\n We provide a complete characterization of the design space and count the . ..\n\n\nAbdalla Ahmed (King Abdullah University of Science and Technology (KAUST)) and Mikhail Skopenkov, Markus Hadwiger, and Peter Wonka (KAUST)\n ---------------------\nLearning Gradient Fields for Scalable and Generaliz able Irregular Packing\n\nThe packing problem, also known as cutting or ne sting, has diverse applications in logistics, manufacturing, layout design , and atlas generation. It involves arranging irregularly shaped pieces to minimize waste while avoiding overlap. Recent advances in machine learnin g, particularly reinforcement ...\n\n\nTianyang Xue (Shandong University), Mingdong Wu (Peking University), Lin Lu and Haoxuan Wang (Shandong Univer sity), and Hao Dong and Baoquan Chen (Peking University)\n---------------- -----\nLight-Efficient Holographic Illumination for Continuous-Wave Time-o f-Flight Imaging\n\nTime-of-flight (TOF) cameras have seen widespread adop tion in recent years across the entire spectrum of commodity devices. Howe ver, these devices are fundamentally limited by their dynamic range, strug gling with saturation from nearby bright objects and noisy depth from fart her darker objects. In t...\n\n\nDorian Chan and Matthew O'Toole (Carnegie Mellon University)\n---------------------\nACE: Adversarial Correspondenc e Embedding for Cross Morphology Motion Retargeting from Human to Nonhuman Characters\n\nMotion retargeting is a promising approach for generating n atural and compelling motions for nonhuman characters. However, it is chal lenging to translate human movements into semantically equivalent motions for target characters with very different morphologies due to ambiguity. T his work presents a...\n\n\nTianyu Li (Georgia Institute of Technology), J ungdam Won (Seoul National University), Alexander Clegg (Meta), Jeonghwan Kim (Georgia Institute of Technology), Akshara Rai (Meta), and Sehoon Ha ( Georgia Institute of Technology)\n---------------------\nReconstruction of Machine-Made Shapes from Bitmap Sketches\n\nWe propose a method of recons tructing 3D machine-made shapes from bitmap sketches by separating an inpu t image into individual patches and jointly optimizing their geometry. \nW e rely on two main observations:\n(1) human observers interpret sketches o f man-made shapes as a collection of simple geometr...\n\n\nIvan Puhachov (Universite de Montreal; Huawei Technologies, Canada); Cedric Martens (Uni versite de Montreal); Paul G. Kry (McGill University; Huawei Technologies, Canada); and Mikhail Bessmeltsev (Universite de Montreal)\n-------------- -------\nA Neural Implicit Representation for the Image Stack: Depth, All in Focus, and High Dynamic Range\n\nIn everyday photography, physical limi tations of camera sensors and lenses frequently lead to a variety of degra dations in captured images such as saturation or defocus blur. A common ap proach to overcome these limitations is to resort to image stack fusion, w hich involves capturing multiple images ...\n\n\nChao Wang (Max-Planck-Ins titut für Informatik); Ana Serrano (Universidad de Zaragoza); and Xingang Pan, Bin Chen, Hans-Peter Seidel, Karol Myszkowski, Christian Theobalt, Kr zysztof Wolski, and Thomas Leimkühler (Max-Planck-Institut für Informatik) \n---------------------\nPose and Skeleton-aware Neural IK for Pose and Mo tion Editing\n\nPosing a 3D character for film or game is an iterative and laborious process where many control handles (e.g. joints) need to be man ipulated to achieve a compelling result. Neural Inverse Kinematics (IK) i s a new type of IK that enables sparse control over a 3D character pose, a nd leverages full bo...\n\n\nDhruv Agrawal (ETH Zürich, DisneyResearch|Stu dios); Martin Guay, Jakob Buhmann, and Dominik Borer (DisneyResearch|Studi os); and Robert W. Sumner (DisneyResearch|Studios, ETH Zürich)\n---------- -----------\nAnything to Glyph: Artistic Font Synthesis via Text-to-Image Diffusion Model\n\nThe automatic generation of artistic fonts is a challen ging task that attracts many research interests. Previous methods specific ally focus on glyph or texture style transfer. However, we often come acro ss creative fonts composed of objects in posters or logos. These fonts hav e proven to be a challe...\n\n\nChangShuo Wang, Lei Wu, XiaoLe Liu, and Xi ang Li (Shandong University); Lei Meng (Shandong University, Shandong Rese arch Institute of Industrial Technology); and Xiangxu Meng (Shandong Unive rsity)\n---------------------\nAdaptive Recurrent Frame Prediction with Le arnable Motion Vectors\n\nThe utilization of dedicated ray tracing graphic s cards has contributed to the production of stunning visual effects in re al-time rendering. However, the demand for high frame rates and high resol utions remains a challenge to be addressed. A crucial technique for increa sing frame rate and resolution...\n\n\nZhizhen Wu (State Key Lab of CAD&CG , Zhejiang University); Chenyu Zuo (State Key Lab of CAD&CG, State Key Lab oratory of CAD & CG, Zhejiang University); Yuchi Huo (State Key Lab of CAD &CG, Zhejiang University; Zhejiang Lab); Yazhen Yuan (Tencent); Yifan Peng (The University of Hong Kong (HKU)); Guiyang Pu (China Mobile (Hangzhou) Information Technology Co., Ltd); and Rui Wang and Hujun Bao (State Key La b of CAD&CG, Zhejiang University)\n---------------------\nLearning based 2 D Irregular Shape Packing\n\n2D irregular shape packing is a necessary ste p to arrange UV patches of a 3D model within a texture atlas for memory-ef ficient appearance rendering in computer graphics. Being a joint, combinat orial decision-making problem involving all patch positions and orientatio ns, this problem has well-known N...\n\n\nZeshi Yang and Zherong Pan (Tenc ent America), Manyi Li (Shandong University), and Kui Wu and Xifeng Gao (T encent America)\n---------------------\nDeepBasis: Hand-Held Single-Image SVBRDF Capture via Two-Level Basis Material Model\n\nRecovering spatial-va rying bi-directional reflectance distribution function (SVBRDF) from a sin gle hand-held captured image has been a meaningful but challenging task in computer graphics. Benefiting from the learned data priors, some previous methods can utilize the potential material correlations ...\n\n\nLi Wang, Lianghao Zhang, Fangzhou Gao, and Jiawan Zhang (Tianjin University)\n---- -----------------\nExtended Path Space Manifolds for Physically Based Diff erentiable Rendering\n\nPhysically based differentiable rendering has beco me an increasingly important topic in recent years. A common pipeline comp utes local color derivatives of light paths or pixels with respect to arbi trary scene parameters, and enables optimizing or recovering the scene par ameters through iterative gr...\n\n\nJiankai Xing and Xuejun Hu (Tsinghua University), Fujun Luan (Adobe Research), Ling-Qi Yan (University of Calif ornia Santa Barbara), and Kun Xu (Tsinghua University)\n------------------ ---\nOptimal Design of Robotic Character Kinematics\n\nThe kinematic motio n of a robotic character is defined by its mechanical joints and actuators that restrict the relative motion of its rigid components. Designing robo ts that perform a given target motion as closely as possible with a fixed number of actuated degrees of freedom is challenging, espec...\n\n\nGuirec Maloisel, Christian Schumacher, Espen Knoop, Ruben Grandia, and Moritz Bä cher (Disney Research)\n---------------------\nDigital 3D Smocking Design\ n\nWe develop an optimization-based method to model smocking, a surface em broidery technique that provides decorative geometric texturing while main taining stretch properties of the fabric. During smocking, multiple pairs of points on the fabric are stitched together, creating non-manifold geome tric fe...\n\n\nJing Ren, Aviv Segall, and Olga Sorkine-Hornung (ETH Züric h)\n---------------------\nAmortizing Samples in Physics-Based Inverse Ren dering using ReSTIR\n\nRecently, great progress has been made in physics-b ased differentiable rendering. Existing differentiable rendering technique s typically focus on static scenes, but during inverse rendering—a key app lication for differentiable rendering—the scene is updated dynamically by each gradient s...\n\n\nYu-Chen Wang (University of California Irvine), Ch ris Wyman and Lifan Wu (NVIDIA), and Shuang Zhao (University of California Irvine)\n---------------------\nC·ASE: Learning Conditional Adversarial S kill Embeddings for Physics-based Characters\n\nWe present C·ASE, an effic ient and effective framework that learns conditional Adversarial Skill Emb eddings for physics-based characters. Our physically simulated character c an learn a diverse repertoire of skills while providing controllability in the form of direct manipulation of the skills to be...\n\n\nZhiyang Dou ( The University of Hong Kong (HKU)), Xuelin Chen and Qingnan Fan (Tencent A I Lab), Taku Komura (University of Hong Kong), and Wenping Wang (Texas A&M University)\n---------------------\nTopology Guaranteed B-Spline Surface/ Surface Intersection\n\nThe surface/surface intersection technique serves as one of the most fundamental functions in modern CAD systems. Despite th e long research history and successful applications of surface intersectio n algorithms in various CAD industrial software, challenges still exist in balancing the computational...\n\n\nJieyin Yang and Xiaohong Jia (Key Lab oratory of Mathematics Mechanization, Chinese Academy Of Sciences; Univers ity of Chinese Academy of Sciences) and Dong-Ming Yan (National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Scien ces)\n---------------------\nPerceptually Adaptive Real-Time Tone Mapping\ n\nTone mapping operators aim to remap content to a display's dynamic rang e. Virtual reality is a popular new display modality that has significant differences from other media, making the use of traditional tone mapping t echniques difficult. Moreover, real-time adaptive estimation of tone curve s that ...\n\n\nTaimoor Tariq (Meta, Università della Svizzera italiana) a nd Nathan Matsuda, Eric Penner, Jerry Jia, Douglas Lanman, Ajit Ninan, and Alexandre Chapiro (Meta)\n---------------------\nRerender A Video: Zero-S hot Text-Guided Video-to-Video Translation\n\nLarge text-to-image diffusio n models have exhibited impressive proficiency in generating high-quality images. However, when applying these models to video domain, ensuring temp oral consistency across video frames remains a formidable challenge.\nThis paper proposes a novel zero-shot text-guided video...\n\n\nShuai Yang, Yi fan Zhou, Ziwei Liu, and Chen Change Loy (Nanyang Technological University , Singapore)\n---------------------\nFrom Skin to Skeleton : Towards Biome chanically Accurate 3D Digital Humans\n\nGreat progress has been made in e stimating 3D human pose and shape from images and video by training neural networks to directly regress the parameters of parametric human models li ke SMPL.\nHowever, existing body models have simplified kinematic structur es that do not correspond to accurate joint lo...\n\n\nMarilyn Keller (Max Planck Institute for Intelligent Systems), Keenon Werling (Stanford Unive rsity), Soyong Shin (Max-Planck-Institut für Informatik), Scott Delp (Stan ford), Sergi Pujades (INRIA), Karen Liu (Stanford University), and Michael Black (Max Planck Institute for Intelligent Systems)\n------------------- --\nFast-MSX: Fast Multiple Scattering Approximation\n\nClassical microfac et theory suffers from energy loss on materials with high roughness due to the single bounce assumption of most microfacet models. When roughness is high, there is a large chance of multiple scattering occurring among the microfacets of the surface. Without explicitly modelling for...\n\n\nEnriq ue Rosales (Huawei) and Fatemeh Teimury, Joshua Horacsek, Aria Salari, Xue bin Qin, Adi Bar-Lev, Xiaoqiang Zhe, and Ligang Liu (Huawei Technologies)\ n---------------------\nDifferentiable Dynamic Visible-Light Tomography\n\ nWe propose the first visible-light tomography system for real-time acquis ition and reconstruction of general temporally-varying 3D phenomena. Using a single high-speed camera, a high-performance LED array and optical fibe rs with a total length of 5km, we build a novel acquisition setup with no mecha...\n\n\nKaizhang Kang, Zoubin Bi, Xiang Feng, Yican Dong, Kun Zhou, and Hongzhi Wu (State Key Laboratory of CAD&CG, Zhejiang Univerisity)\n--- ------------------\nMyStyle++: A Controllable Personalized Generative Prio r\n\nIn this paper, we propose an approach to obtain a personalized genera tive prior with explicit control over a set of attributes. We build upon M yStyle, a recently introduced method, that tunes the weights of a pre-trai ned StyleGAN face generator on a few images of an individual. This system allows sy...\n\n\nLibing Zeng (Texas A&M University), Lele Chen and Yi Xu (OPPO US Research Center), and Nima Kalantari (Texas A&M University)\n---- -----------------\nGroundLink: A Dataset Unifying Human Body Movement and Ground Reaction Dynamics\n\nThe physical plausibility of human motions is vital to various applications in the fields including but not limited to g raphics, animation, robotics, vision, biomechanics, and sports science. Wh ile fully simulating human motions with physics is an extreme challenge, w e hypothesize that we can treat ...\n\n\nXingjian Han, Ben Senderling, Sta nley To, Deepak Kumar, and Emily Whiting (Boston University) and Jun Saito (Adobe Research)\n---------------------\nDiffFR: Differentiable SPH-based Fluid-Rigid Coupling for Rigid Body Control\n\nDifferentiable physics sim ulation has shown its efficacy in inverse design problems. Given the perva siveness of the diverse interactions between fluids and solids in life, a differentiable simulator for the inverse design of the motion of rigid obj ects in two-way fluid-rigid coupling is also demande...\n\n\nZhehao Li and Qingyu Xu (University of Science and Technology of China), Xiaohan Ye and Bo Ren (Nankai University), and Ligang Liu (University of Science and Tec hnology of China)\n---------------------\nCommonsense Knowledge-Driven Joi nt Reasoning Approach for Object Retrieval in Virtual Reality\n\nOut-of-re ach object retrieval is an important task in virtual reality (VR). Gesture -based approach, one of the most commonly used approaches, enables bare-ha nd, eyes-free, and direct retrieval by using assigned gestures. However, i t is difficult to retrieve an object from plenty of objects accuratel...\n \n\nHaiyan Jiang (Beijing Institute of Technology; National Key Laboratory of General Artificial Intelligence, Beijing Institute for General Artific ial Intelligence (BIGAI)); Dondong Weng (Beijing Institute of Technology); Xiaonuo Dongye (Beijing Institute of Technology; National Key Laboratory of General Artificial Intelligence, Beijing Institute for General Artifici al Intelligence (BIGAI)); Le Luo (Beijing Institute of Technology); and Zh enliang Zhang (National Key Laboratory of General Artificial Intelligence, Beijing Institute for General Artificial Intelligence (BIGAI))\n--------- ------------\nLock-free Vertex Clustering for Multicore Mesh Reduction\n\n Modern data collection methods can capture representations of 3D objects a t resolutions much greater than they can be discretely rendered as an imag e. To improve the efficiency of storage, transmission, rendering, and edit ing of 3D models constructed from such data, it is beneficial to first emp loy ...\n\n\nNima Fathollahi and Sean Chester (University of Victoria)\n-- -------------------\nHigh Density Ratio Multi-fluid Simulation with Peridy namics\n\nMultiple fluid simulation has raised wide research interest in r ecent years. Despite the impressive successes of current works, simulation of scenes containing mixing or unmixing of high-density-ratio phases usin g particle-based discretizations still remains a challenging task. In this paper, we pro...\n\n\nHan Yan and Bo Ren (Nankai University)\n----------- ----------\nNeural Motion Graph\n\nDeep learning techniques have been empl oyed to design a controllable human motion synthesizer. Despite their pote ntial, however, designing a neural network-based motion synthesis that ena bles flexible user interaction, fine-grained controllability, and the supp ort of new types of motions at reduced ...\n\n\nHongyu Tao, Shuaiying Hou, Changqing Zou, Hujun Bao, and Weiwei Xu (Zhejiang University)\n---------- -----------\nTowards Practical Capture of High-Fidelity Relightable Avatar s\n\nIn this paper, we propose a novel framework, Tracking-free Relightabl e Avatar (TRAvatar), for capturing and reconstructing high-fidelity 3D ava tars. Compared to previous methods, TRAvatar works in a more practical and efficient setting. Specifically, TRAvatar is trained with dynamic image s equences ...\n\n\nHaotian Yang, Mingwu Zheng, Wanquan Feng, and Haibin Hua ng (Kuaishou Technology); Yu-Kun Lai (Cardiff University); and Pengfei Wan , Zhongyuan Wang, and Chongyang Ma (Kuaishou Technology)\n---------------- -----\nWhat is the Best Automated Metric for Text to Motion Generation?\n\ nThere is growing interest in generating skeleton-based human motions from natural language descriptions. While most efforts have focused on develop ing better neural architectures for this task, there has been no significa nt work on determining the proper evaluation metric. Human evaluation is t he ul...\n\n\nJordan Voas, Yili Wang, Qixing Huang, and Raymond Mooney (Un iversity of Texas at Austin)\n---------------------\nRectifying Strip Patt erns\n\nA straight flat strip of inextensible material can be bent into cu rved strips aligned with arbitrary space curves. The large shape variety o f these so-called rectifying strips makes them candidates for shape modeli ng, especially in applications such as architecture where simple elements are preferre...\n\n\nBolun Wang and Hui Wang (King Abdullah University of Science and Technology (KAUST)), Eike Schling (The University of Hong Kong ), and Helmut Pottmann (King Abdullah University of Science and Technology (KAUST))\n---------------------\nConditional Resampled Importance Samplin g and ReSTIR\n\nRecent work on generalized resampled importance sampling ( GRIS) enables importance-sampled Monte Carlo integration with random varia ble weights replacing the usual division by probability density. This enab les very flexible spatiotemporal sample reuse, even if neighboring samples (e.g., light paths)...\n\n\nMarkus Kettunen and Daqi Lin (NVIDIA); Ravi R amamoorthi (NVIDIA, University of California San Diego); Thomas Bashford-R ogers (University of Warwick); and Chris Wyman (NVIDIA)\n----------------- ----\nLearning the Geodesic Embedding with Graph Neural Networks\n\nWe pre sent GeGnn, a learning-based method for computing the approximate geodesic distance between two arbitrary points on discrete polyhedra surfaces with constant time complexity after fast precomputation. Previous relevant met hods either focus on computing the geodesic distance between a single so.. .\n\n\nBo Pang (Peking Unversity); Zhongtian Zheng (Peking University); Gu oping Wang (Peking Unversity); and Peng-Shuai Wang (Peking University, Wan gxuan Institute of Computer Technology)\n---------------------\nCLIP-Guide d StyleGAN Inversion for Text-Driven Real Image Editing\n\nResearchers hav e recently begun exploring the use of StyleGAN-based models for real image editing. One particularly interesting application is using natural langua ge descriptions to guide the editing process. Existing approaches for edit ing images using language either resort to instance-level laten...\n\n\nAb dul Basit Anees and Ahmet Canberk Baykal (Koç University), Duygu Ceylan (A dobe Research), Erkut Erdem (Hacettepe University), and Aykut Erdem and De niz Yuret (Koç University)\n---------------------\nReconstructing Close Hu man Interaction from Multiple Views\n\nThis paper addresses the challengin g task of reconstructing the poses of multiple individuals engaged in clos e interactions, captured by multiple calibrated cameras. The difficulty ar ises from the noisy or false 2D keypoint detection due to inter-person occ lusion, the heavy ambiguity to associate ke...\n\n\nQing Shuai (Zhejiang U niversity); Zhiyuan Yu (Department of Mathematics, Hong Kong University of Science and Technology); Zhize Zhou (Capital University of Physical Educa tion and Sports); Lixin Fan and Haijun Yang (WeBank); Can Yang (Department of Mathematics, Hong Kong University of Science and Technology); and Xiao wei Zhou (State Key Laboratory of CAD&CG, Zhejiang Univerisity)\n--------- ------------\nText-Guided Synthesis of Eulerian Cinemagraphs\n\nWe introdu ce Text2Cinemagraph, a fully automated method for creating cinemagraphs f rom text descriptions - an especially challenging task when prompts featur e imaginary elements and artistic styles, given the complexity of interpre ting the semantics and motions of these images. We focus on cinemagr...\n\ n\nAniruddha Mahapatra (Carnegie Mellon University); Aliaksandr Siarohin, Hsin-Ying Lee, and Sergey Tulyakov (Snap Inc.); and Jun-Yan Zhu (Carnegie Mellon University)\n---------------------\nInteractive Story Visualization with Multiple Characters\n\nAccurate Story visualization requires several necessary elements, such as identity consistency across frames, the align ment between plain text and visual content, and a reasonable layout of obj ects in images. Most previous works endeavor to meet these requirements by fitting a text-to-image (T2I) mo...\n\n\nYuan Gong (Tsinghua University); Youxin Pang (MAIS & NLPR, Institute of Automation, Chinese Academy of Sci ences, Beijing, China; School of Artificial Intelligence, University of Ch inese Academy of Sciences); Xiaodong Cun and Menghan Xia (Tencent); Yingqi ng He (Hong Kong University of Science and Technology); Haoxin Chen, Longy ue Wang, Yong Zhang, Xintao Wang, and Ying Shan (Tencent); and Yujiu Yang (Tsinghua University)\n---------------------\nLearning Contact Deformation s with General Collider Descriptors\n\nThis paper presents a learning-base d method for the simulation of rich contact deformations on reduced deform ation models. Previous works learn deformation models for specific pairs o f objects, and we lift this limitation by designing a neural model that su pports general rigid collider shapes. We do...\n\n\nCristian Romero and Da n Casas (Universidad Rey Juan Carlos) and Maurizio Chiaramonte and Miguel A. Otaduy (Meta Reality Labs Research)\n---------------------\nConcept Dec omposition for Visual Exploration and Inspiration\n\nA creative idea is of ten born from transforming, combining, and modifying ideas from existing v isual examples capturing various concepts.\nHowever, one cannot simply cop y the concept as a whole, and inspiration is achieved by examining certain aspects of the concept. Hence, it is often necessary to s...\n\n\nYael Vi nker (Tel Aviv University, Google Research); Andrey Voynov (Google Researc h); Daniel Cohen-Or (Tel Aviv University, Google Research); and Ariel Sham ir (Reichman University)\n---------------------\nKirchhoff-Love Shells wit h Arbitrary Hyperelastic Materials\n\nKirchhoff-Love shells are commonly u sed in many branches of engineering, including in computer graphics, but h ave so far been simulated only under limited nonlinear material options. W e derive the Kirchhoff-Love thin-shell mechanical energy for an arbitrary 3D volumetric hyperelastic material, inclu...\n\n\nJiahao Wen and Jernej B arbic (University of Southern California)\n---------------------\nSelf-Cal ibrating, Fully Differentiable NLOS Inverse Rendering\n\nExisting time-res olved non-line-of-sight (NLOS) imaging methods reconstruct hidden scenes b y inverting the optical paths of indirect illumination measured at visible relay surfaces. These methods are prone to reconstruction artifacts due t o inversion ambiguities and capture noise, which are typicall...\n\n\nKise ok Choi, Inchul Kim, and Dongyoung Choi (Korea Advanced Institute of Scien ce and Technology (KAIST)); Julio Marco and Diego Gutierrez (Universidad d e Zaragoza - I3A); and Min H. Kim (Korea Advanced Institute of Science and Technology (KAIST))\n---------------------\nEfficient Hybrid Zoom using C amera Fusion on Mobile Phones\n\nDSLR cameras can achieve multiple zoom le vels via shifting lens distances or swapping lens types. However, these te chniques are not possible on smartphone devices due to space constraints. Most smartphone manufacturers adopt a hybrid zoom system: commonly a Wide (W) camera at a low zoom level and a ...\n\n\nXiaotong Wu, Wei-Sheng Lai, and Yichang Shih (Google Inc.); Charles Herrmann and Michael Krainin (Goog le Research); Deqing Sun (Google); and Chia-Kai Liang (Google Inc.)\n----- ----------------\nSOL-NeRF: Sunlight Modeling for Outdoor Scene Decomposit ion and Relighting\n\nOutdoor scenes often involve large-scale geometry an d complex unknown lighting conditions, making it difficult to decompose th em into geometry, reflectance and illumination. Recently researchers made attempts to decompose outdoor scenes using Neural Radiance Fields (NeRF) a nd learning-based lighting...\n\n\nJia-Mu Sun and Tong Wu (Institute of Co mputing Technology, Chinese Academy of Sciences; University of Chinese Aca demy of Sciences); Yong-Liang Yang (University of Bath); Yu-Kun Lai (Cardi ff University); and Lin Gao (Institute of Computing Technology, Chinese Ac ademy of Sciences; University of Chinese Academy of Sciences)\n----------- ----------\nMeshes with Spherical Faces\n\nA truly Möbius invariant discre te surface theory must consider meshes where the transformation group acts on all of its elements, including edges and faces. We therefore systemati cally describe so called sphere meshes with spherical faces and circular a rcs as edges. Driven by aspects important for m...\n\n\nMartin Kilian (TU Wien), Anthony Ramos Cisneros (KAUST), Christian Müller (TU Wien), and Hel mut Pottmann (KAUST)\n---------------------\nMatFusion: A Generative Diffu sion Model for SVBRDF Capture\n\nWe formulate SVBRDF estimation from photo graphs as a diffusion task. To model the distribution of spatially varying materials, we first train a novel unconditional SVBRDF diffusion backbone model on a large set of 312,165 synthetic spatially varying material exem plars. This SVBRDF diffusion backbo...\n\n\nSam Sartor and Pieter Peers (College of William & Mary)\n---------------------\nSAILOR: Synergizing Ra diance and Occupancy Fields for Live Human Performance Capture\n\nImmersiv e user experiences in live VR/AR performances require a fast and accurate free-view rendering of the performers. Existing methods are mainly based o n Pixel-aligned Implicit Functions (PIFu) or Neural Radiance Fields (NeRF) . However, while PIFu-based methods usually fail to produce photoreali...\ n\n\nZheng Dong (State Key Laboratory of CAD & CG, Zhejiang University); K e Xu (City University of Hong Kong); Yaoan Gao (State Key Laboratory of CA D & CG, Zhejiang University); Qilin Sun (The Chinese University of Hong Ko ng, Shenzhen); Hujun Bao and Weiwei Xu (State Key Laboratory of CAD & CG, Zhejiang University); and Rynson W.H. Lau (City University of Hong Kong)\n ---------------------\nMultisource Holography\n\nHolographic displays prom ise several benefits including high quality 3D imagery, accurate accommoda tion cues, and compact form-factors. However, holography relies on coheren t illumination which can create undesirable speckle noise in the final ima ge. Although smooth phase holograms can be speckle-fr...\n\n\nGrace Kuo, F lorian Schiffers, Douglas Lanman, Oliver Cossairt, and Nathan Matsuda (Rea lity Labs Research, Meta)\n---------------------\nTowards Garment Sewing P attern Reconstruction from a Single Image\n\nGarment sewing pattern repres ents the intrinsic rest shape of a garment, and is the core for many appli cations like fashion design, virtual try-on, and digital avatars. In this work, we explore the challenging problem of recovering garment sewing patt erns from daily photos for augmenting these appli...\n\n\nLijuan Liu (Sea AI Lab); Xiangyu Xu (Xi'an Jiaotong University, Sea AI Lab); and Zhijie Li n, Jiabin Liang, and Shuicheng Yan (Sea AI Lab)\n---------------------\nSF LSH: Shape-Dependent Soft-Flesh Avatars\n\nWe present a multi-person soft- tissue avatar model. This model maps a body shape descriptor to heterogene ous geometric and mechanical parameters of a soft-tissue model across the body, effectively producing a shape-dependent parametric soft avatar model . The design of the model overcomes two major c...\n\n\nPablo Ramón, Crist ian Romero, Javier Tapia, and Miguel A. Otaduy (Universidad Rey Juan Carlo s)\n---------------------\nNodeGit: Diffing and Merging Node Graphs\n\nThe use of version control is pervasive in collaborative software projects. V ersion control systems are based on two primary operations: diffing two ve rsions to compute the change between them and merging two versions edited concurrently. Recent works provide solutions to diff and merge graphics as s...\n\n\nEduardo Rinaldi (Ubisoft), Davide Sforza (Sapienza University of Rome), and Fabio Pellacini (University of Modena and Reggio Emilia)\n---- -----------------\nLow-Light Image Enhancement with Wavelet-based Diffusio n Models\n\nDiffusion models have achieved promising results in image rest oration tasks, yet suffer from time-consuming, excessive computational res ource consumption, and unstable restoration. To address these issues, we p ropose a robust and efficient Diffusion-based Low-Light image enhancement approach, dubbed...\n\n\nHai Jiang (Sichuan University); Ao Luo and Haoqia ng Fan (Megvii); Songchen Han (Sichuan University); and Shuaicheng Liu (Un iversity of Electronic Science and Technology of China, Megvii)\n--------- ------------\nCompact Neural Graphic Primitives with Learned Hash Probing\ n\nNeural graphics primitives are faster and achieve higher quality when t heir neural networks are augmented by spatial data structures that hold tr ainable features arranged in a grid. However, existing feature grids eithe r come with a large memory footprint (dense or factorized grids, trees, an d hash ...\n\n\nTowaki Takikawa (NVIDIA, University of Toronto); Thomas Mü ller, Merlin Nimier-David, and Alex Evans (NVIDIA); Sanja Fidler (NVIDIA, University of Toronto); Alec Jacobson (University of Toronto, Adobe Resear ch); and Alexander Keller (NVIDIA)\n---------------------\nEditing Motion Graphics Video via Motion Vectorization and Transformation\n\nMotion graph ics videos are widely used in Web design, digital advertising, animated lo gos and film title sequences, to capture a viewer's attention. But editing such video is challenging because the video provides a low-level sequence of pixels and frames rather than higher-level structure such as t...\n\n\ nSharon Zhang and Jiaju Ma (Stanford University); Daniel Ritchie (Brown Un iversity); Jiajun Wu (Stanford University); and Maneesh Agrawala (Stanford University, Roblox)\n---------------------\nQuantum Ray Marching for Refo rmulating Light Transport Simulation\n\nThe use of quantum computers in co mputer graphics has gained interest in recent years, especially for the ap plication to rendering. The current state of the art in quantum rendering relies on Grover's search for finding ray intersections in $O(\sqrt{M})$ f or $M$ primitives. This quantum approach is ...\n\n\nLogan Mosier (Univers ity of Waterloo); Morgan McGuire (Roblox, University of Waterloo); and Tos hiya Hachisuka (University of Waterloo)\n---------------------\nA Hessian- Based Field Deformer for Real-Time Topology-Aware Shape Editing\n\nShape m anipulation is a central research topic in computer graphics. Topology edi ting, such as breaking apart connections, joining disconnected ends, and f illing/opening a topological hole, is generally more challenging than geom etry editing. In this paper, we observe that the saddle points of the s... \n\n\nYunxiao Zhang, Zixiong Wang, Zihan Zhao, and Rui Xu (Shandong Univer sity); Shuangmin Chen (Qingdao University of Science and Technology); Shiq ing Xin (Shandong University); Wenping Wang (Texas A&M University); and Ch anghe Tu (Shandong University)\n---------------------\nNeural Gradient Lea rning and Optimization for Oriented Point Normal Estimation\n\nWe propose Neural Gradient Learning (NGL), a deep learning approach to learn gradient vectors with consistent orientation from 3D point clouds for normal estim ation. It has excellent gradient approximation properties for the underlyi ng geometry of the data. We utilize a simple neural network to para...\n\n \nQing Li (Tsinghua University), Huifang Feng (Xiamen University), Kanle S hi (Kuaishou Technology), Yi Fang (New York University), Yu-Shen Liu (Tsin ghua University), and Zhizhong Han (Wayne State University)\n------------- --------\nDiffusion Posterior Illumination for Ambiguity-aware Inverse Ren dering\n\nInverse rendering, the process of inferring scene properties fro m images, is a challenging inverse problem. The task is ill-posed, as many different scene configurations can give rise to the same image. Most exis ting solutions incorporate priors into the inverse-rendering pipeline to e ncourage plaus...\n\n\nLinjie Lyu (Max-Planck-Institut für Informatik), Ay ush Tewari (MIT CSAIL), Marc Habermann (Max-Planck-Institut für Informatik ), Shunsuke Saito and Michael Zollhöfer (Reality Labs Research), and Thoma s Leimkühler and Christian Theobalt (Max-Planck-Institut für Informatik)\n ---------------------\nSlippage-Preserving Reshaping of Human-Made 3D Cont ent\n\nArtists often need to reshape 3D models of human-made objects by ch anging the relative proportions or scales of different model parts or elem ents while preserving the look and structure of the inputs. Manually resha ping inputs to satisfy these criteria is highly time-consuming; the edit i n our tease...\n\n\nChrystiano Araújo (University of British Columbia); Ni cholas Vining (University of British Columbia, NVIDIA); and Silver Burla, Manuel Ruivo de Oliveira, Enrique Rosales, and Alla Sheffer (University of British Columbia)\n---------------------\nClothCombo: Modeling Inter-Clot h Interaction for Draping Multi-Layered Clothes\n\nWe present ClothCombo, a pipeline to drape arbitrary combinations of clothes on 3D human models w ith varying body shapes and poses. While existing learning-based approache s for draping clothes have shown promising results, multi-layered clothing remains challenging as it is non-trivial to model inte...\n\n\nDohae Lee, Hyun Kang, and In-Kwon Lee (Yonsei University)\n---------------------\nNe ural Field Convolutions by Repeated Differentiation\n\nNeural fields are e volving towards a general-purpose continuous representation for visual com puting. Yet, despite their numerous appealing properties, they are hardly amenable to signal processing. As a remedy, we present a method to perform general continuous convolutions with general continuous si...\n\n\nNtumba Elie Nsampi, Adarsh Djeacoumar, and Hans-Peter Seidel (Max-Planck-Institu t für Informatik); Tobias Ritschel (University College London (UCL)); and Thomas Leimkühler (Max-Planck-Institut für Informatik)\n------------------ ---\nMuscleVAE: Model-Based Controllers of Muscle-Actuated Characters\n\nI n this paper, we present a simulation and control framework for generating biomechanically plausible motion for muscle-actuated characters. We incor porate a fatigue dynamics model, the 3CC-r model, into the widely-adopted Hill-type muscle model to simulate the development and recovery of fatigue in...\n\n\nYusen Feng, Xiyan Xu, and Libin Liu (Peking University)\n----- ----------------\nFluid Simulation on Neural Flow Maps\n\nWe introduce Neu ral Flow Maps, a novel simulation method bridging the emerging paradigm of implicit neural representations with fluid simulation based on the theory of flow maps, to achieve state-of-the-art simulation of inviscid fluid ph enomena. We devise a novel hybrid neural field representation,...\n\n\nYit ong Deng (Dartmouth College), Hong-Xing Yu (Stanford University), Diyang Z hang (Dartmouth College), Jiajun Wu (Stanford University), and Bo Zhu (Dar tmouth College)\n---------------------\nA Physically-inspired Approach to the Simulation of Plant Wilting\n\nPlants are among the most complex objec ts to be modeled in computer graphics. While a large body of work is conce rned with structural modeling and the dynamic reaction to external forces, our work focuses on the dynamic deformation caused by plant internal wilt ing processes. To this end, we motivate...\n\n\nFilippo Maggioli (Sapienza - University of Rome), Jonathan Klein and Torsten Hädrich (KAUST), Emanue le Rodolà (Sapienza - University of Rome), Wojtek Pałubicki (AMU), Sören P irk (CAU), and Dominik L. Michels (KAUST)\n---------------------\nNeural-S ingular-Hessian: Implicit Neural Representation of Unoriented Point Clouds by Enforcing Singular Hessian\n\nNeural implicit representation is a prom ising approach for reconstructing surfaces from point clouds. Existing met hods combine various regularization terms to enforce the learned neural fu nction to possess the properties of a SDF, such as the Eikonal term and La placian energy term. However, when the...\n\n\nZixiong Wang, Yunxiao Zhang , and Rui Xu (Shandong University); Fan Zhang (Shandong Technology and Bus iness University); Peng-Shuai Wang (Peking University); Shuangmin Chen (Qi ngdao University of Science and Technology); Shiqing Xin (Shandong Univers ity); Wenping Wang (Texas A&M University); and Changhe Tu (Shandong Univer sity)\n---------------------\nAvatarStudio: Text-driven Editing of 3D Dyna mic Human Head Avatars\n\nCapturing and editing full head performances ena bles the creation of virtual characters with various applications such as extended reality and media production. The past few years witnessed a stee p rise in the photorealism of human head avatars. Such avatars can be cont rolled through different input...\n\n\nMohit Mendiratta, Xingang Pan, Moha med Elgharib, Kartik Teotia, and Mallikarjun B R (Max Planck Institute for Informatics); Ayush Tewari (MIT CSAIL); Vladislav Golyanik (Max Planck In stitute for Informatics); Adam Kortylewski (Max Planck Institute for Infor matics, University of Freiburg); and Christian Theobalt (Max Planck Instit ute for Informatics)\n---------------------\nHigh-Order Moment-Encoded Kin etic Simulation of Turbulent Flows\n\nKinetic solvers for incompressible f luid simulation were designed to run efficiently on massively parallel arc hitectures such as GPUs. While these lattice Boltzmann solvers have recent ly proven much faster and more accurate than common macroscopic Navier-Sto kes-based solvers in graphics, it is alway...\n\n\nWei Li, Tongtong Wang, Zherong Pan, Xifeng Gao, and Kui Wu (LightSpeed Studios) and Mathieu Desbr un (Inria/X)\n---------------------\nVET: Visual Error Tomography for Poin t Cloud Completion and High-Quality Neural Rendering\n\nIn the last few ye ars, deep neural networks opened the doors for big advances in novel view synthesis. Many of these approaches are based on a (coarse) proxy geometry obtained by structure from motion algorithms. Small deficiencies in this proxy can be fixed by neural rendering, but larger holes or ...\n\n\nLinus Franke, Darius Rückert, and Laura Fink (Friedrich-Alexander Universität E rlangen-Nürnberg); Matthias Innmann (NavVis GmbH); and Marc Stamminger (Fr iedrich-Alexander Universität Erlangen-Nürnberg)\n---------------------\nV iCMA: Visual Control of Multibody Animations\n\nMotion control of large-sc ale, multibody physics animations with contact is difficult. Existing appr oaches, such as those based on optimization, are computationally daunting, and, as the number of interacting objects increases, can fail to find sat isfactory solutions. We present a new, complementary...\n\n\nDoug L. James (Stanford University, NVIDIA) and David I. W. Levin (University of Toront o, NVIDIA)\n---------------------\nDiscovering Fatigued Movements for Virt ual Character Animation\n\nVirtual character animation and movement synthe sis have advanced rapidly during recent years, especially through a combin ation of extensive motion capture datasets and machine learning. A remaini ng challenge is interactively simulating characters that fatigue when perf orming extended motions, which ...\n\n\nNoshaba Cheema (German Research Ce nter for Artificial Intelligence, Max-Planck Institute for Informatics); R ui Xu (German Research Center for Artificial Intelligence, Saarland Univer sity); Nam Hee Kim and Perttu Hämäläinen (Aalto University); Vladislav Gol yanik and Marc Habermann (Max-Planck-Institut für Informatik); Christian T heobalt (Max-Planck-Institut für Informatik, Saarland University); and Phi lipp Slusallek (Saarland University, German Research Center for Artificial Intelligence)\n---------------------\nObject Motion Guided Human Motion S ynthesis\n\nModeling human behaviors in contextual environments has a wide range of applications in character animation, embodied AI, VR/AR, and rob otics. In real-world scenarios, humans frequently interact with the enviro nment and manipulate various objects to complete daily tasks. In this work , we study the p...\n\n\nJiaman Li, Jiajun Wu, and Karen Liu (Stanford Uni versity)\n---------------------\nRobust Zero Level-Set Extraction from Uns igned Distance Fields Based on Double Covering\n\nIn this paper, we propos e a new method, called DoubleCoverUDF, for extracting the zero level-set f rom unsigned distance fields (UDFs). DoubleCoverUDF takes a learned UDF an d a user-specified parameter r (a small positive real number) as input and extracts an iso-surface with an iso-value r using the...\n\n\nFei Hou (In stitute of Software, Chinese Academy of Sciences; University of Chinese Ac ademy of Sciences); Xuhui Chen and Wencheng Wang (Institute of Software, C hinese Academy Of Sciences; University of Chinese Academy of Sciences); Ho ng Qin (Stony Brook University); and Ying He (Nanyang Technological Univer sity)\n---------------------\nTwinTex: Geometry-aware Texture Generation f or Abstracted 3D Architectural Models\n\nCoarse architectural models are o ften generated at scales ranging from individual buildings to scenes for d ownstream applications such as Digital Twin City, Metaverse, LODs, etc. Su ch piece-wise planar models can be abstracted as twins from 3D dense recon structions. However, these models typically l...\n\n\nWeidan Xiong, Hongqi an Zhang, Botao Peng, Ziyu Hu, Yongli Wu, Jianwei Guo, and Hui Huang (Shen zhen University)\n---------------------\nVariational Barycentric Coordinat es\n\nWe propose a variational technique to optimize for generalized baryc entric coordinates that offers additional artistic control compared to exi sting models. Prior work represents barycentric coordinates using meshes o r closed-form formulae, in practice limiting the choice of objective funct ion. In co...\n\n\nAna Dodik (MIT CSAIL); Oded Stein (University of Southe rn California, MIT CSAIL); and Vincent SItzmann and Justin Solomon (MIT CS AIL)\n---------------------\nUVDoc: Neural Grid-based Document Unwarping\n \nRestoring the original, flat appearance of a printed document from casua l photographs of bent and wrinkled pages is a common everyday problem. In this paper we propose a novel method for grid-based single-image document unwarping. Our method performs geometric distortion correction via a fully convo...\n\n\nFloor Verhoeven, Tanguy Magne, and Olga Sorkine-Hornung (ET H Zurich)\n---------------------\nFLARE: Fast Learning of Animatable and R elightable Mesh Avatars\n\nOur goal is to efficiently learn personalized a nimatable 3D head avatars from videos that are geometrically accurate, rea listic, relightable, and compatible with current rendering systems. While 3D meshes enable efficient processing and are highly portable, they lack r ealism in terms of shape and ap...\n\n\nShrisha Bharadwaj (Max Planck Inst itute for Intelligent Systems); Yufeng Zheng (ETH Zürich, Max Planck Insti tute for Intelligent Systems); Otmar Hilliges (ETH Zürich); and Michael Bl ack and Victoria Fernandez Abrevaya (Max Planck Institute for Intelligent Systems)\n---------------------\nART-Owen Scrambling\n\nWe present a novel algorithm for implementing Owen-scrambling, combining the generation and distribution of the scrambling bits in a single self-contained compact pro cess.\nWe employ a context-free grammar to build a binary tree of symbols, and equip each symbol with a scrambling code that affects al...\n\n\nAbda lla G. M. Ahmed (KAUST), Matt Pharr (NVIDIA), and Peter Wonka (KAUST)\n--- ------------------\nCapturing Animation-Ready Isotropic Materials Using Sy stematic Poking\n\nCapturing material properties of real-world elastic sol ids is both challenging and highly relevant to many applications in comput er graphics, robotics and related fields. We give a non-intrusive, in-situ and inexpensive approach to measure the nonlinear elastic energy density function of man-made ma...\n\n\nHuanyu Chen, Danyong Zhao, and Jernej Barb ic (University of Southern California)\n---------------------\nConstructiv e Solid Geometry on Neural Signed Distance Fields\n\nSigned Distance Field s (SDFs) parameterized by neural networks have recently gained popularity as a fundamental geometric representation. However, editing the shape enco ded by a neural SDF remains an open challenge. A tempting approach is to leverage common geometric operators (e.g., boolean operat...\n\n\nZoë Mars chner (Massachusetts Institute of Technology, Carnegie Mellon University); Silvia Sellán (University of Toronto); Hsueh-Ti Derek Liu (Roblox Resear ch); and Alec Jacobson (University of Toronto)\n---------------------\nDop pler Time-of-Flight Rendering\n\nWe introduce Doppler time-of-flight (D-To F) rendering, an extension of ToF rendering for dynamic scenes, with appli cations in simulating D-ToF cameras. D-ToF cameras use high-frequency modu lation for illumination and exposure, and measure the Doppler frequency sh ift to compute the velocity of dynami...\n\n\nJuhyeon Kim and Wojciech Jar osz (Dartmouth College), Ioannis Gkioulekas (Carnegie Mellon University), and Adithya Pediredla (Dartmouth College)\n---------------------\nDrivable Avatar Clothing: Faithful Full-Body Telepresence with Dynamic Clothing Dr iven by Sparse RGB-D Input\n\nClothing is an important part of human appea rance but challenging to model in photorealistic avatars. In this work we present avatars with dynamically moving loose clothing that can be faithfu lly driven by sparse RGB-D inputs as well as body and face motion. We prop ose a Neural Iterative Closest Poi...\n\n\nDonglai Xiang (Carnegie Mellon University/Robotics Institute, Meta Reality Labs Research); Fabian Prada, Zhe Cao, Kaiwen Guo, and Chenglei Wu (Meta Reality Labs Research); Jessica Hodgins (Carnegie Mellon University); and Timur Bagautdinov (Meta Reality Labs Research)\n---------------------\nStable Discrete Bending by Analyti c Eigensystem and Adaptive Orthotropic Geometric Stiffness\n\nIn this pape r, we address two limitations of dihedral angle based discrete bending (DA B) models, i.e. the indefiniteness of their energy Hessian and their vulne rability to geometry degeneracies. To tackle the indefiniteness issue, we present novel analytic expressions for the eigensystem of a DAB en...\n\n\ nZhendong Wang (Style3D Research); Yin Yang (University of Utah, Style3D R esearch); and Huamin Wang (Style3D Research)\n---------------------\nMulti -color Holograms Improve Brightness in Holographic Displays\n\nHolographic displays generate Three-Dimensional (3D) images by displaying single-colo r holograms time-sequentially, each lit by a single-color light source. Ho wever, representing each color one by one limits brightness in holographic displays.\n\nThis paper introduces a new driving scheme for realizin...\n \n\nKoray Kavaklı (Koç University), Liang Shi (Massachusetts Institute of Technology), Hakan Urey (Koç University), Wojciech Matusik (Massachusetts Institute of Technology), and Kaan Akşit (University College London (UCL)) \n---------------------\nPSDR-Room: Single Photo to Scene using Differenti able Rendering\n\nA 3D digital scene composes many components: lights, mat erials and geometries, interacting to reach the desired appearance. Stagin g such a scene is time-consuming and requires both artistic and technical skills. In this work, we propose a system allowing to optimize lighting as well as the pose and ...\n\n\nKai Yan (University of California, Irvine; Adobe Research); Fujun Luan, Miloš Hašan, Thibault Groueix, and Valentin D eschaintre (Adobe Research); and Shuang Zhao (University of California, Ir vine)\n---------------------\nProgressive Shell Quasistatics for Unstructu red Meshes\n\nThin shell structures exhibit complex behaviors critical for modeling and design across wide-ranging applications. To capture their me chanical response requires finely detailed, high-resolution meshes. Corres ponding simulations for predicting equilibria with these meshes are expens ive, whereas coars...\n\n\nJiayi Eris Zhang (Stanford University, Adobe); Jérémie Dumas and Raymond Fei (Adobe); Alec Jacobson (University of Toront o, Adobe); Doug James (Stanford University); and Danny Kaufman (Adobe)\n-- -------------------\nSubspace Mixed Finite Elements for Real-Time Heteroge neous Elastodynamics\n\nReal-time elastodynamic solvers are well-suited fo r the rapid simulation of ho\nmogeneous elastic materials, with high-rates generally enabled by aggressive\nearly termination of timestep solves. Un fortunately, the introduction of strong\ndomain heterogeneities can make t hese solvers slow to converge. ...\n\n\nTy Trusty, Otman Benchekroun, and Eitan Grinspun (University of Toronto); Danny M. Kaufman (Adobe, Universit y of Toronto); and David I.W. Levin (University of Toronto)\n------------- --------\nJoint UV Optimization and Texture Baking\n\nLevel of detail (LOD ) has been widely used in interactive computer graphics. In current indust rial 3D modeling pipelines, artists rely on commercial software to generat e highly detailed models with UV maps, and then bake textures for low-poly counterparts. In these pipelines, each step is performed ...\n\n\nJulian Knodt, Zherong Pan, Kui Wu, and Xifeng Gao (LightSpeed Studios)\n--------- ------------\nDecaf: Monocular Deformation Capture for Face and Hand Inter actions\n\nExisting methods for 3D tracking from monocular RGB videos pred ominantly consider articulated and rigid objects (e.g., two hands or human s interacting with rigid environments). Modelling dense non-rigid object deformations in this setting (e.g., when hand are interacting with a face) , remained larg...\n\n\nSoshi Shimada (Max-Planck-Institut für Informatik; Saarbrücken Research Center for Visual Computing, Interaction and Artific ial Intelligence); Vladislav Golyanik (Max-Planck-Institut für Informatik ); Patrick Pérez (Valeo); and Christian Theobalt (Max-Planck-Institut für Informatik; Saarbrücken Research Center for Visual Computing, Interaction and Artificial Intelligence)\n---------------------\nAuthoring and Simula ting Meandering Rivers\n\nWe present a method for interactively authoring and simulating meandering river networks. Starting from a terrain with an initial low-resolution network encoded as a directed graph, we simulate th e evolution of the path of the different channels using a physically-based migration equation augmented ...\n\n\nAxel Paris (Université Claude Berna rd Lyon, Adobe); Eric Guérin (LIRIS); Pauline Collon (Université de Lorrai ne); and Eric Galin (LIRIS)\n---------------------\nSparse Stress Structur es from Optimal Geometric Measures\n\nIdentifying optimal structural desig ns given loads and constraints is a primary challenge in topology optimiza tion and shape optimization. We propose a novel approach to this problem by finding a minimal tensegrity structure—a network of cables and struts i n equilibrium with a given loading f...\n\n\nDylan Rowe and Albert Chern ( University of California San Diego)\n---------------------\nAdaptive Track ing of a Single-Rigid-Body Character in Various Environments\n\nSince the introduction of DeepMimic [Peng et al. 2018], subsequent research\nhas foc used on expanding the repertoire of simulated motions across various\nscen arios. In this study, we propose an alternative approach for this goal,\na deep reinforcement learning method based on the simulation of a single... \n\n\nTaesoo Kwon, Taehong Gu, Jaewon Ahn, and Yoonsang Lee (Hanyang Unive rsity)\n---------------------\nExplorable Mesh Deformation Subspaces from Unstructured 3D Generative Models\n\nExploring variations of 3D shapes is a time-consuming process in traditional 3D modeling tools. Deep generative models of 3D shapes often feature continuous latent spaces that can, in p rinciple, be used to explore potential variations starting from a set of i nput shapes; in practice, doing so can be...\n\n\nArman Maesumi (Brown Uni versity); Paul Guerrero, Vladimir Kim, and Matthew Fisher (Adobe Inc.); Si ddhartha Chaudhuri (Adobe Inc.; Indian Institute of Technology (IIT), Bomb ay); Noam Aigerman (Adobe Inc.); and Daniel Ritchie (Brown University)\n-- -------------------\nHolographic Near-eye Display with Real-time Embedded Rendering\n\nWe present a wearable full-color holographic augmented realit y headset with binocular vision support and real-time embedded hologram ca lculation. Contrarily to most previously proposed prototypes, our headset employs high-speed amplitude-only microdisplays and embeds a compact and l ightweight electr...\n\n\nAntonin Gilles and Pierre Le Gargasson (Institut e of Research and Technology b-com) and Grégory Hocquet and Patrick Gioia (Orange Labs, Institute of Research and Technology b-com)\n--------------- ------\nComputational Design of Wiring Layout on Tight Suits with Minimal Motion Resistance\n\nAn increasing number of electronics are directly embe dded on the clothing to monitor human status (skeletal motion or electromy ogram activity) or provide haptic feedback.\nA specific challenge to proto type and fabricate such a clothing is to design the wiring layout, to mini mize the intervention to h...\n\n\nKai Wang, Xiaoyu Xu, Yinping Zheng, Da Zhou, and Shihui Guo (Xiamen University); Yipeng Qin (School of Computer S cience and Informatic, Cardiff University); and Xiaohu Guo (University of Texas at Dallas)\n---------------------\nBézier Spline Simplification Usin g Locally Integrated Error Metrics\n\nInspired by surface mesh simplificat ion methods, we present a technique for reducing the number of Bézier curv es in a vector graphics while maintaining high fidelity. We propose a curv e-to-curve distance metric to repeatedly conduct local segment removal ope rations. By construction, we identify all ...\n\n\nSiqi Wang (New York Uni versity); Chenxi Liu (University of British Columbia); Daniele Panozzo and Denis Zorin (New York University); and Alec Jacobson (University of Toron to, Adobe Research)\n---------------------\nNonlinear Ray Tracing for Disp lacement and Shell Mapping\n\nDisplacement mapping and shell mapping add f ine-scale geometric features to meshes and can significantly enhance the r ealism of an object's surface representation. Both methods generate geomet ry within a layer between a base mesh and its offset mesh called a shell. It is not easy to simultaneously a...\n\n\nShinji Ogaki (ZOZO, Inc.)\n---- -----------------\nSAME: Skeleton-Agnostic Motion Embedding for Character Animation\n\nLearning deep neural networks on human motion data has become common in computer graphics research, but the heterogeneity of available datasets poses challenges for training large-scale networks. \nThis paper presents a framework that allows us to solve various animation tasks in a skeleton-agnostic ...\n\n\nSunmin Lee, Taeho Kang, and Jungnam Park (Seoul National University); Jehee Lee (NC Research, Seoul National University); and Jungdam Won (Seoul National University)\n---------------------\nColla psing Embedded Cell Complexes for Safer Hexahedral Meshing\n\nWe present a set of operators to perform modifications, in particular collapses and sp lits, in volumetric cell complexes which are discretely embedded in a back ground mesh. Topological integrity and geometric embedding validity are ca refully maintained. We apply these operators strategically to bloc...\n\n\ nHendrik Brückler and Marcel Campen (Osnabrück University)\n-------------- -------\nProjective Sampling for Differentiable Rendering of Geometry\n\nD iscontinuous visibility changes at object boundaries remain a persistent s ource of difficulty in the area of differentiable rendering. Left untreate d, they bias computed gradients so severely that even basic optimization t asks fail.\n\nPrior path-space methods addressed this bias by decoupling b ounda...\n\n\nZiyi Zhang, Nicolas Roussel, and Wenzel Jakob (Ecole Polytec hnique Fédérale de Lausanne)\n---------------------\nDepolarized Holograph y with Polarization-multiplexing Metasurface\n\nThe evolution of computer- generated holography (CGH) algorithms has prompted significant improvement s in the performances of holographic displays. Nonetheless, they start to encounter a limited degree of freedom in CGH optimization and physical con straints stemming from the coherent nature of hologr...\n\n\nSeung-Woo Nam , Youngjin Kim, Dongyeon Kim, and Yoonchan Jeong (Seoul National Universit y)\n---------------------\nSecond-Order Finite Elements for Deformable Sur faces\n\nWe present a computational framework for simulating deformable su rfaces with second-order triangular finite elements. Our method develops n umerical schemes for discretizing stretching, shearing, and bending energi es of deformable surfaces in a second-order finite-element setting. In par ticular, we i...\n\n\nQiqin Le (Shanghai Qi Zhi Institute); Yitong Deng (S tanford University); Jiamu Bu (Tsinghua University); Bo Zhu (Georgia Insti tute of Technology); and Tao Du (Tsinghua University, Shanghai Qi Zhi Inst itute)\n---------------------\nK-surfaces: Bézier-Splines Interpolating at Gaussian Curvature Extrema\n\nK-surfaces are an interactive modeling tech nique for Bézier-spline surfaces. Inspired by 𝜅-curves by [Yan et al. 2017 ], each patch provides a single control point that is being interpolated a t a local extremum of Gaussian curvature.\nThe challenge is to solve the i nverse problem of finding th...\n\n\nTobias Djuren, Maximilian Kohlbrenner , and Marc Alexa (Technische Universität Berlin)\n---------------------\nV Mesh: Hybrid Volume-Mesh Representation for Efficient View Synthesis\n\nWi th the emergence of neural radiance fields (NeRFs), view synthesis quality has reached an unprecedented level. Compared to traditional mesh-based as sets, this volumetric representation is more powerful in expressing scene geometry but inevitably suffers from high rendering costs and can hardly b e ...\n\n\nYuan-Chen Guo (Tsinghua University, Tencent); Yan-Pei Cao (Tenc ent); Chen Wang (Tsinghua University); Yu He (Chinese Academy of Sciences) ; Ying Shan (Tencent); and Song-Hai Zhang (Tsinghua University)\n--------- ------------\nIntrinsic Harmonization for Illumination-Aware Image Composi ting\n\nDespite significant advancements in network-based image harmonizat ion techniques, there still exists a domain gap between training pairs and real-world composites encountered during inference. Most existing methods are trained to reverse global edits made on segmented image regions, whic h fail to ac...\n\n\nChris Careaga, S. Mahdi H. Miangoleh, and Yağız Aksoy (Simon Fraser University)\n---------------------\nShadow Harmonization fo r Realistic Compositing\n\nCompositing virtual objects into real backgroun d images requires one to carefully match the scene's camera parameters, su rface geometry, textures, and lighting to obtain plausible renderings.\nRe cent learning approaches have shown many scene properties can be estimated from images, resulting in robus...\n\n\nLucas Valença and Jinsong Zhang ( Université Laval), Michaël Gharbi and Yannick Hold-Geoffroy (Adobe), and J ean-François Lalonde (Université Laval)\n---------------------\nSingle-Ima ge 3D Human Digitization with Shape-guided Diffusion\n\nWe present an appr oach to generate a 360-degree view of a person with a consistent, high-res olution appearance from a single input image. NeRF and its variants typica lly require videos or images from different viewpoints. Most existing appr oaches taking monocular input either rely on ground-truth 3D...\n\n\nBadou r AlBahar (Kuwait University); Shunsuke Saito, Hung-Yu Tseng, Changil Kim, and Johannes Kopf (Meta); and Jia-Bin Huang (University of Maryland)\n--- ------------------\nMetric Optimization in Penner Coordinates\n\nMany para metrization and mapping-related problems in geometry processing can be vie wed as metric optimization problems, i.e., computing a metric minimizing a functional and satisfying a set of constraints, such as flatness. \n\nPen ner coordinates are global coordinates on the space of metrics on meshe... \n\n\nRyan Capouellez and Denis Zorin (New York University)\n------------- --------\nReal-time Height-field Simulation of Sand and Water Mixtures\n\n We propose a height-field-based real-time simulation method for sand and w ater mixtures. Inspired by the shallow-water assumption, our approach exte nds the governing equations to handle two-phase flows of sand and water us ing height fields. Our depth-integrated governing equations can model the elas...\n\n\nHaozhe Su (Rutgers University, LightSpeed Studios); Siyu Zhan g and Zherong Pan (LightSpeed Studios); Mridul Aanjaneya (Rutgers Universi ty); and Xifeng Gao and Kui Wu (LightSpeed Studios)\n--------------------- \nNeRFFaceLighting: Implicit and Disentangled Face Lighting Representation Leveraging Generative Prior in Neural Radiance Fields\n\n3D-aware portrai t lighting control is an emerging and promising domain, thanks to the rece nt advance of generative adversarial networks and neural radiance fields. Existing solutions typically try to decouple the lighting from the geometr y and appearance for disentangled control with an explicit lig...\n\n\nKai wen Jiang (Institute of Computing Technology, Chinese Academy of Sciences; Beijing Jiaotong University); Shu-Yu Chen (Institute of Computing Technol ogy, Chinese Academy of Sciences); Hongbo Fu (School of Creative Media, Ci ty University of Hong Kong); and Lin Gao (Institute of Computing Technolog y, Chinese Academy of Sciences; University of Chinese Academy of Sciences) \n---------------------\nDiscontinuity-Aware 2D Neural Fields\n\nNeural im age representations offer the possibility of high-fidelity, compact storag e, and resolution-independent accuracy, providing an attractive alternativ e to traditional pixel and grid-based representations. \nHowever, coordin ate neural networks fail to capture discontinuities present in the ima...\ n\n\nYash Belhe (University of California San Diego); Michael Gharbi, Matt Fisher, and Iliyan Georgiev (Adobe Inc.); and Ravi Ramamoorthi and Tzu-Ma o Li (University of California San Diego)\n---------------------\nEmotiona l Speech-Driven Animation with Content-Emotion Disentanglement\n\nTo be wi dely adopted, 3D facial avatars need to be animated easily, realistically, and directly, from speech signals. While the best recent methods generate 3D animations that are synchronized with the input audio, they largely ig nore the impact of emotions on facial expressions. Instead, their focu...\ n\n\nRadek Daněček (Max Planck Institute for Intelligent Systems); Kiran C hhatre (KTH Royal Institute of Technology); Shashank Tripathi, Yandong Wen , and Michael Black (Max Planck Institute for Intelligent Systems); and Ti mo Bolkart (Max Planck Institut for Intelligent Systems)\n---------------- -----\nDomain-Agnostic Tuning-Encoder for Fast Personalization of Text-To- Image Models\n\nText-to-image (T2I) personalization allows users to guide the creative image generation process by combining their own visual concep ts in natural language prompts. \nRecently, encoder-based techniques have emerged as a new effective approach for T2I personalization, reducing the need for multiple ima...\n\n\nMoab Arar (Tel-Aviv University); Rinon Gal ( Tel Aviv University, NVIDIA Research); Yuval Atzmon (NVIDIA Research); Gal Chechik (NVIDIA Research, Bar-Ilan University); Daniel Cohen-Or (Tel Aviv University); Ariel Shamir (Reichman University (IDC)); and Amit H. Berman o (Tel Aviv University)\n---------------------\n3D Bézier Guarding: Bounda ry-Conforming Curved Tetrahedral Meshing\n\nWe present a method for the ge neration of higher-order tetrahedral meshes. In contrast to previous metho ds, the curved tetrahedral elements are guaranteed to be free of degenerac ies and inversions while conforming exactly to prescribed piecewise polyno mial surfaces, such as domain boundaries or mate...\n\n\nPayam Khanteimour i and Marcel Campen (Osnabrück University)\n---------------------\nThe eff ect of display capabilities on the gloss consistency between real and virt ual objects\n\nA faithful reproduction of gloss is inherently difficult be cause of the limited dynamic range, peak luminance, and 3D capabilities of display devices. This work investigates how the display capabilities affe ct gloss appearance with respect to a real-world reference object. To this end, we employ an ...\n\n\nBin Chen (Max-Planck-Institut für Informatik); Akshay Jindal (Intel Corporation, University of Cambridge); Michal Piovar či (Institute of Science and Technology Austria); Chao Wang and Hans-Peter Seidel (Max-Planck-Institut für Informatik); Piotr Didyk (Università dell a Svizzera italiana); Karol Myszkowski (Max-Planck-Institut für Informatik ); Ana Serrano (Universidad de Zaragoza); and Rafał K. Mantiuk (University of Cambridge)\n---------------------\nEnhancing Diffusion Models with 3D Perspective Geometry Constraints\n\nWhile perspective is a well-studied to pic in art, it is generally taken for granted in images. However, for the recent wave of high-quality image synthesis methods such as latent diffusi on models, perspective accuracy is not an explicit requirement. Since thes e methods are capable of outputting a wi...\n\n\nRishi Upadhyay and Howard Zhang (University of California, Los Angeles); Yunhao Ba (University of C alifornia, Los Angeles; Sony); Ethan Yang, Blake Gella, and Sicheng Jiang (University of California, Los Angeles); Alex Wong (Yale University); and Achuta Kadambi (University of California, Los Angeles)\n------------------ ---\nAnimating Street View\n\nWe present a system that automatically bring s street view imagery to life by populating it with naturally behaving, an imated pedestrians and vehicles. Our approach is to remove existing people and vehicles from the input image, insert moving objects with proper scal e, angle, motion and appearance, p...\n\n\nMengyi Shan, Brian Curless, Ira Kemelmacher-Shlizerman, and Steve Seitz (University of Washington)\n----- ----------------\nZero-Shot 3D Shape Correspondence\n\nWe propose a novel zero-shot approach to computing correspondences\nbetween 3D shapes. Existi ng approaches mainly focus on isometric and\nnear-isometric shape pairs (e .g., human vs. human), but less attention has\nbeen given to strongly non- isometric and inter-class shape matching (e.g., human vs. cow)...\n\n\nAhm ed Abdelreheem and Abdelrahman Eldesokey (King Abdullah University of Scie nce and Technology (KAUST)), Maks Ovsjanikov (Centre National de la Recher che Scientifique - Laboratoire d'informatique de l'École Polytechnique (LI X)), and Peter Wonka (King Abdullah University of Science and Technology ( KAUST))\n---------------------\nFusing Monocular Images and Sparse IMU Sig nals for Real-time Human Motion Capture\n\nEither RGB images or inertial s ignals have been used for the task of motion capture (mocap), but combinin g them together is a new and interesting topic. We believe that the combin ation is complementary and able to solve the inherent difficulties of usin g one modality input, including occlusions, ext...\n\n\nShaohua Pan, Qi Ma , and Xinyu Yi (Tsinghua University); Weifeng Hu, Xiong Wang, Xingkang ZHO U, and Jijunnan LI (OPPO Research Institute); and Feng Xu (Tsinghua Univer sity)\n---------------------\nMCNeRF: Monte Carlo Rendering and Denoising for Real-Time NeRFs\n\nThe volume rendering step used in Neural Radiance F ields (NeRFs) produces highly photorealistic results, but is inherently sl ow because it evaluates an MLP at a large number of sample points per ray. Previous work has addressed this by either proposing neural scene represe ntations that are faster to...\n\n\nKunal Gupta (UC San Diego); Milos Hasa n, Zexiang Xu, Fujun Luan, Kalyan Sunkavalli, and Xin Sun (Adobe Inc.); Ma nmohan Chandraker (UC San Diego); and Sai Bi (Adobe Inc.)\n--------------- ------\nScaNeRF: Scalable Bundle-Adjusting Neural Radiance Fields for Larg e-Scale Scene Rendering\n\nHigh-quality large-scale scene rendering requir es a scalable representation and accurate camera poses. This research comb ines tile-based hybrid neural fields with parallel distributive optimizati on to improve bundle-adjusting neural radiance fields. The proposed method scales with a divide-and-conqu...\n\n\nXiuchao Wu (State Key Laboratory o f CAD & CG, Zhejiang University); Jiamin Xu (Hangzhou Dianzi Univeristy); Xin Zhang (State Key Laboratory of CAD&CG, Zhejiang Univerisity); Hujun Ba o (State Key Laboratory of CAD & CG, Zhejiang University); Qixing Huang (U niversity of Texas at Austin); Yujun Shen (Ant Group); James Tompkin (Brow n University); and Weiwei Xu (State Key Laboratory of CAD&CG, Zhejiang Uni verisity)\n---------------------\nWarped-Area Reparameterization of Differ ential Path Integrals\n\nPhysics-based differentiable rendering is becomin g increasingly crucial for tasks in inverse rendering and machine learning pipelines. To address discontinuities caused by geometric boundaries and occlusion, two classes of methods have been proposed: 1) the edge sampling methods that directly sample...\n\n\nPeiyu Xu (University of California I rvine), Sai Bangaru (MIT CSAIL), Tzu-Mao Li (University of California San Diego), and Shuang Zhao (University of California Irvine)\n--------------- ------\nEfficient Human Motion Reconstruction from Monocular Videos with P hysical Consistency Loss\n\nVision-only motion reconstruction from monocul ar videos often produces artifacts such as foot sliding and jittery motion s. Existing physics-based methods typically either simplify the problem to focus solely on feet-ground contacts, or reconstruct full-body contacts w ithin a physics simulator, neces...\n\n\nLin Cong and Philipp Ruppel (Univ ersität Hamburg), Yizhou Wang (Peking University), Xiang Pan (Diago Tech C ompany), and Norman Hendrich and Jianwei Zhang (Universität Hamburg)\n---- -----------------\nExtraSS: A Framework for Joint Spatial Super Sampling a nd Frame Extrapolation\n\nWe introduce ExtraSS, a novel framework that com bines spatial super sampling and frame extrapolation to enhance real-time rendering performance. By integrating these techniques, our approach achie ves a balance between performance and quality, generating temporally stabl e and high-quality, high-resol...\n\n\nSongyin Wu (University of Californi a Santa Barbara); Sungye Kim (Intel Corporation); Zheng Zeng (University o f California, Santa Barbara); Deepak Vembar, Sangeeta Jha, and Anton Kapla nyan (Intel Corporation); and Ling-Qi Yan (University of California, Santa Barbara)\n---------------------\nFuseSR: Super Resolution for Real-time R endering through Efficient Multi-resolution Fusion\n\nThe workload of real -time rendering is steeply increasing as the demand for high resolution, h igh refresh rates, and high realism rises, overwhelming most graphics card s. To mitigate this problem, one of the most popular solutions is to rende r images at a low resolution to reduce rendering overhead,...\n\n\nZhihua Zhong (State Key Lab of CAD&CG, Zhejiang University; Zhejiang University C ity College); Jingsen Zhu (State Key Lab of CAD&CG, Zhejiang University); Yuxin Dai (Zhejiang A&F University); Chuankun Zheng (State Key Lab of CAD& CG, Zhejiang University); Guanlin Chen (Zhejiang University City College); Yuchi Huo (Zhejiang Lab; State Key Lab of CAD&CG, Zhejiang University); a nd Hujun Bao and Rui Wang (State Key Lab of CAD&CG, Zhejiang University)\n ---------------------\nConstrained Delaunay Tetrahedrization: A Robust and Practical Approach\n\nWe present a numerically robust algorithm for compu ting the constrained Delaunay tetrahedrization (CDT) of a piecewise-linear complex, which has a 100% success rate on the 4408 valid models in the Th ingy10k dataset.\nWe build on the underlying theory of the well-known tetg en software, but use a float...\n\n\nLorenzo Diazzi (UNIMORE, CNR-IMATI: G ENOVA); Daniele Panozzo (NYU); Amir Vaxman (University of Edinburgh); and Marco Attene (CNR-IMATI: GENOVA)\n---------------------\nAnti-Aliased Neur al Implicit Surfaces with Encoding Level of Detail\n\nWe present LoD-NeuS, an efficient neural representation for high-frequency geometry detail rec overy and anti-aliased novel view rendering. Drawing inspiration from voxe l-based representations with the level of detail (LoD), we introduce a mul ti-scale tri-plane-based scene representation that is capa...\n\n\nYiyu Zh uang (Nanjing University); Qi Zhang and Ying Feng (Tencent); Hao Zhu and Y ao Yao (Nanjing University); Xiaoyu Li, Yan-Pei Cao, and Ying Shan (Tencen t); and Xun Cao (Nanjing University)\n---------------------\nSimultaneous Color Computer Generated Holography\n\nComputer generated holography has l ong been touted as the future of augmented and virtual reality (AR/VR) dis plays, but has yet to be realized in practice. Previous high-quality, colo r holographic displays have made either a 3x sacrifice on frame rate by us ing a sequential color illumination scheme ...\n\n\nEric Markley, Nathan M atsuda, Florian Schiffers, Oliver Coissart, and Grace Kuo (Meta)\n-------- -------------\nInovis: Instant Novel-View Synthesis\n\nNovel-view synthesi s is an ill-posed problem in that it requires inference of previously unse en information. Recently, reviving the traditional field of image-based re ndering, neural methods proved particularly suitable for this interpolatio n/extrapolation task; however, they often require a-priori ...\n\n\nMathia s Harrer and Linus Franke (Friedrich-Alexander-Universität Erlangen-Nürnbe rg); Laura Fink (Friedrich-Alexander-Universität Erlangen-Nürnberg, Fraunh ofer IIS); and Marc Stamminger and Tim Weyrich (Friedrich-Alexander-Univer sität Erlangen-Nürnberg)\n---------------------\n360° Reconstruction From a Single Image Using Space Carved Outpainting\n\nWe introduce POP3D, a nov el framework that creates a full $360^\circ$-view 3D model from a single i mage. POP3D resolves two prominent issues that limit the single-view recon struction. Firstly, POP3D offers substantial generalizability to arbitrary categories, a trait that previous methods struggle t...\n\n\nNuri Ryu, Mi nsu Gong, and Geonung Kim (POSTECH); Joo-Haeng Lee (Pebblous Inc.); and Su nghyun Cho (POSTECH, Pebblous Inc.)\n---------------------\nVASCO: Volume and Surface Co-Decomposition for Hybrid Manufacturing\n\nAdditive and subt ractive hybrid manufacturing (ASHM) involves the alternating use of additi ve and subtractive manufacturing techniques, which provides unique advanta ges for fabricating complex geometries with otherwise inaccessible surface s. However, a significant challenge lies in ensuring tool acc...\n\n\nFanc hao Zhong, Haisen Zhao, Haochen Li, Xin Yan, and Jikai Liu (Shandong Unive rsity); Baoquan Chen (Peking University); and Lin Lu (Shandong University) \n---------------------\nThe Design Space of Kirchhoff Rods\n\nThe Kirchho ff rod model describes the deformation behavior of elastic rods interactin g with boundary conditions. We characterize the set of all equilibrium sta tes admitted by this model, assuming spatially-varying cross sections, and present an algorithm to compute the geometry of a rod that will pr...\n\n \nChristian Hafner (Institute of Science and Technology Austria) and Bernd Bickel (Institute of Science and Technology Austria, Google Research)\n-- -------------------\nSparsePoser: Real-time Full-body Motion Reconstructio n from Sparse Data\n\nAccurate and reliable human motion reconstruction is crucial for creating natural interactions of full-body avatars in Virtual Reality (VR) and entertainment applications. As the Metaverse and social applications gain popularity, users are seeking cost-effective solutions t o create full-body animati...\n\n\nJose Luis Ponton and Haoran Yun (Univer sitat Politècnica de Catalunya (UPC)); Andreas Aristidou (University of Cy prus, CYENS Centre of Excellence); and Carlos Andujar and Nuria Pelechano (Universitat Politècnica de Catalunya (UPC))\n---------------------\nCLIPX Plore: Coupled CLIP and Shape Spaces for 3D Shape Exploration\n\nThis pape r presents CLIPXPlore, a new framework that leverages a vision-language mo del to guide the exploration of the 3D shape space. Many recent methods ha ve been developed to encode 3D shapes into a learned latent shape space to enable generative design and modeling. Yet, existing methods lack ef...\n \n\nJingyu Hu, Ka-Hei Hui, and Zhengzhe Liu (The Chinese University of Hon g Kong); Hao (Richard) Zhang (Simon Fraser University); and Chi-Wing Fu (T he Chinese University of Hong Kong)\n---------------------\nEMS: 3D Eyebro w Modeling from Single-view Images\n\nEyebrows play a critical role in fac ial expression and appearance. Although the 3D digitization of faces is we ll explored, less attention has been drawn to 3D eyebrow modeling. In this work, we propose EMS, the first learning-based framework for single-view 3D eyebrow reconstruction. Following the m...\n\n\nChenghong Li, Leyang Ji n, and Yujian Zheng (The Chinese University of Hong Kong, Shenzhen); Yizho u Yu (The University of Hong Kong); and Xiaoguang Han (The Chinese Univers ity of Hong Kong, Shenzhen)\n---------------------\nNeural Metamaterial Ne tworks for Nonlinear Material Design\n\nNonlinear metamaterials with tailo red mechanical properties have applications in engineering, medicine, robo tics, and beyond. While modeling\ntheir macromechanical behavior is challe nging in itself, finding structure\nparameters that lead to an ideal appro ximation of high-level performance goals\nis a ...\n\n\nYue Li, Stelian Co ros, and Bernhard Thomaszewski (ETH Zürich)\n---------------------\nRepara mCAD: Zero-shot CAD Re-Parameterization for Interactive Manipulation\n\nPa rametric CAD models encode entire families of shapes that should, in princ iple, be easy for designers to explore, but are in practice difficult to m anipulate due to implicit semantic constraints between parameter values. F inding and enforcing these semantic constraints purely from geometry or pr og...\n\n\nMilin Kodnongbua and Benjamin Jones (University of Washington), Maaz Bin Safeer Ahmad and Vladimir Kim (Adobe), and Adriana Schulz (Unive rsity of Washington)\n---------------------\nNeural Spectro-polarimetric F ields\n\nThe spatial radiance distribution modeling of any light ray withi n a scene has been extensively explored for applications, including view s ynthesis. Spectrum and polarization — the wave properties of light — are o ften neglected due to their integration into three RGB spectral bands and t...\n\n\nYoungchan Kim, Wonjoon Jin, Sunghyun Cho, and Seung-Hwan Baek (P OSTECH)\n---------------------\nNeural Stochastic Poisson Surface Reconstr uction\n\nReconstructing a surface from a point cloud is an underdetermine d problem. We propose using a neural network to study and quantify this re construction uncertainty under a Poisson smoothness prior. Our algorithm a ddresses the main limitations of existing work and can be fully integrated into the 3D s...\n\n\nSilvia Sellán (University of Toronto) and Alec Jaco bson (University of Toronto, Adobe Research)\n---------------------\nDiffu sion-based Holistic Texture Rectification and Synthesis\n\nWe present a no vel framework for rectifying occlusions and distortions in degraded textur e samples from natural images. Traditional texture synthesis approaches fo cus on generating textures from pristine samples, which necessitate meticu lous preparation by humans and are often unattainable in most n...\n\n\nGu oqing Hao (University of Tsukuba, National Institute of Advanced Industria l Science and Technology); Satoshi Iizuka (University of Tsukuba); Kensho Hara (National Institute of Advanced Industrial Science); Edgar Simo-Serra (Waseda University); Hirokatsu Kataoka (National Institute of Advanced In dustrial Science); and Kazuhiro Fukui (University of Tsukuba)\n----------- ----------\nDiffusing Colors: Image Colorization with Text Guided Diffusio n\n\nThe colorization of grayscale images is a complex and subjective task with significant challenges. Despite recent progress in employing large-s cale datasets with deep neural networks, difficulties with controllability and visual quality persist. To tackle these issues, we present a novel im age color...\n\n\nNir Zabari, Aharon Azulay, Alexey Gorkor, and Tavi Halpe rin (Lightricks) and Ohad Fried (Reichman University)\n------------------- --\nLiveNVS: Neural View Synthesis on Live RGB-D Streams\n\nExisting real- time RGB-D reconstruction approaches, like Kinect Fusion, lack real-time p hoto-realistic visualization. This is due to noisy, oversmoothed or incomp lete geometry and blurry textures which are fused from imperfect depth map s and camera poses. Recent neural rendering methods can overcome...\n\n\nL aura Fink (Friedrich-Alexander-Universität Erlangen-Nürnberg, Fraunhofer I IS); Darius Rückert and Linus Franke (Friedrich-Alexander-Universität Erla ngen-Nürnberg); Joachim Keinert (Fraunhofer IIS); and Marc Stamminger (Fri edrich-Alexander-Universität Erlangen-Nürnberg)\n---------------------\nVR -NeRF: High-Fidelity Virtualized Walkable Spaces\n\nWe present an end-to-e nd system for the high-fidelity capture, model reconstruction and real-tim e rendering of walkable spaces in virtual reality using neural radiance fi elds. To this end, we designed and built a custom multi-camera rig to dens ely capture walkable spaces in high fidelity with multi-...\n\n\nLinning X u (The Chinese University of Hong Kong, Meta); Vasu Agrawal, William Laney , Tony Garcia, Aayush Bansal, Changil Kim, Samuel Rota Bulò, Lorenzo Porzi , Peter Kontschieder, and Aljaž Božič (Meta); Dahua Lin (The Chinese Unive rsity of Hong Kong); and Michael Zollhoefer and Christian Richardt (Meta)\ n---------------------\nControllable Group Choreography using Contrastive Diffusion\n\nMusic-driven group choreography poses a considerable challeng e but holds significant potential for a wide range of industrial applicati ons. The ability to generate synchronized and visually appealing group dan ce motions that are aligned with music opens up opportunities in many fiel ds such as entert...\n\n\nNhat Le and Tuong Do (AIOZ); Khoa Do (VNUHCM-Uni versity of Science); Hien Nguyen, Erman Tjiputra, and Quang Tran (AIOZ); a nd Anh Nguyen (University of Liverpool)\n---------------------\nAn Implici tly Stable Mixture Model for Dynamic Multi-fluid Simulations\n\nParticle-b ased simulation has become increasingly popular in real-time applications due to its efficiency and adaptability, especially in generating highly dy namic fluid effects. Nevertheless, the swift and stable simulation of inte ractions between distinct fluids continues to pose challenges for cu...\n\ n\nYanrui Xu (University of Groningen, University of Science and Technolog y Beijing); Xiaokun Wang (University of Science and Technology Beijing, Bo urnemouth University); Jiamin Wang, Chongming Song, Tiancheng Wang, and Ya nlan Zhang (University of Science and Technology Beijing); Jian Chang and Jianjun Zhang (Bournemouth University); Jiri Kosinka (University of Gronin gen); Alexandru Telea (Utrecht University); and Xiaojuan Ban (University o f Science and Technology Beijing)\n---------------------\nReal-Time Recons truction of Fluid Flow under Unknown Disturbance\n\nWe present a framework that captures sparse Lagrangian flow information from a volume of real li quid and reconstructs its detailed kinematic information in real time. Our framework can perform flow reconstruction even when the liquid is disturb ed by an object of unknown movement and shape. Through a...\n\n\nKinfung C hu (Tohoku University); Jiawei Huang (Chuzhou University, Void Dimensions) ; and Hidemasa Takana and Yoshifumi Kitamura (Tohoku University)\n-------- -------------\nDROP: Dynamics Responses from Human Motion Prior and Projec tive Dynamics\n\nSynthesizing realistic human movements, dynamically respo nsive to the environment, is a long-standing objective in character animat ion, with applications in computer vision, sports, and healthcare, for mot ion prediction and data augmentation. Recent kinematics-based generative m otion models offer im...\n\n\nYifeng Jiang (Stanford University), Jungdam Won (Seoul National University), Yuting Ye (Meta Reality Labs Research), a nd C. Karen Liu (Stanford University)\n---------------------\nAn Unified $ \lambda$-subdivision Scheme for Quadrilateral Meshes with Optimal Curvatur e Performance in Extraordinary Regions\n\nWe propose an unified $\lambda$- subdivision scheme with a continuous family of tuned subdivisions for quad rilateral meshes. Main subdivision stencil parameters of the unified schem e are represented as spline functions of the subdominant eigenvalue $\lamb da$ of respective subdivision matrices and the...\n\n\nWeiyin Ma, Xu Wang, and Yue Ma (City University of Hong Kong)\n---------------------\nSinMPI: Novel View Synthesis from a Single Image with Expanded Multiplane Images\ n\nSingle-image novel view synthesis is a challenging and ongoing problem that aims to generate an infinite number of consistent views from a single input image. Although significant efforts have been made to advance the q uality of generated novel views, less attention has been paid to the expan sion of...\n\n\nGuo Pu, Peng-Shuai Wang, and Zhouhui Lian (Wangxuan Instit ute of Computer Technology, Peking University)\n---------------------\nAn Architecture and Implementation of Real-Time Sound Propagation Hardware fo r Mobile Devices\n\nThis paper presents a high-performance and low-power h ardware architecture for real-time sound rendering on mobile devices. Trad itional sound rendering algorithms require high-performance CPUs or GPUs f or processing because of its high computational complexities to realize ul tra-realistic 3D audio. ...\n\n\nEUNJAE KIM, SUKWON CHOI, and JIYOUNG KIM (Sejong University, Sejongpia); JAE-HO NAH (Sangmyung Univesrity); WOONAM JUNG (Sejongpia); TAE-HYEONG LEE (Sejong University); YEON-KUG MOON (Korea Electronics Technology Institute); and WOO-CHAN PARK (Sejong University, Sejongpia)\n---------------------\nShrink & Morph: 3D-printed self-shaping shells actuated by a shape memory effect\n\nWhile 3D printing enables the customization and home fabrication of a wide range of shapes, fabricating freeform thin-shells remains challenging. As layers misalign with the cur vature, they incur structural deficiencies, while the curved shells requir e large support structures, typically using more ...\n\n\nDavid Jourdan (I NRIA Nancy - Grand Est, LORIA); Camille Schreck (Inria); and Pierre-Alexan dre Hugron, Jonas Martinez, and Sylvain Lefebvre (INRIA)\n---------------- -----\nSpatiotemporally Consistent HDR Indoor Lighting Estimation\n\nWe pr opose a physically-motivated deep learning framework to solve a general ve rsion of the challenging indoor lighting estimation problem. Given a singl e LDR image with a depth map, our method predicts spatially consistent lig hting at any given image position. Particularly, when the input is an LDR. ..\n\n\nZhengqin Li (Meta, University of California San Diego); Yu Li and Mikhail Okunev (Meta); Manmohan Chandraker (University of California San D iego); and Zhao Dong (Meta)\n---------------------\nNon-Newtonian ViRheome try via Similarity Analysis\n\nWe estimate the three Herschel–Bulkley para meters (yield stress, power law index, and consistency parameter) for shea r-dependent fluid-like materials possibly with large-scale inclusions, for which rheometers may fail to provide a useful measurement. We perform exp eriments using the unknown ma...\n\n\nMitsuki Hamamichi (Aoyama Gakuin Uni versity); Kentaro Nagasawa and Masato Okada (The University of Tokyo); Ryo hei Seto (Wenzhou Institute, University of Chinese Academy of Sciences; Ou jiang Laboratory (Zhejiang Lab for Regenerative Medicine, Vision and Brain Health)); and Yonghao Yue (Aoyama Gakuin University)\n------------------- --\nFace0: Instantaneously Conditioning a Text-to-Image Model on a Face\n\ nWe present Face0, a novel way to instantaneously condition a text-to-imag e generation model on a face, in sample time, without any optimization pro cedures such as fine-tuning or inversions. We augment a dataset of annotat ed images with embeddings of the included faces and train an image generat ion m...\n\n\nDani Valevski, Danny Lumen, Yossi Matias, and Yaniv Leviatha n (Google Research)\n---------------------\nReach For the Spheres: Tangenc y-aware surface reconstruction of SDFs\n\nSigned distance fields (SDFs) ar e a widely utilized implicit surface representation that has applications in various fields such as computer graphics, computer vision, and applied mathematics. Despite their frequent use, traditional methods such as March ing Cubes and its variants often overlook fund...\n\n\nSilvia Sellán (Univ ersity of Toronto), Christopher Batty (University of Waterloo), and Oded S tein (University of Southern California)\n---------------------\nCamP: Cam era Preconditioning for Neural Radiance Fields\n\nNeural Radiance Fields ( NeRF) can be optimized to obtain high-fidelity 3D scene reconstructions of objects and large-scale scenes. However, NeRFs require accurate camera pa rameters as input --- inaccurate camera parameters result in blurry render ings. Extrinsic and intrinsic camera parameters are us...\n\n\nKeunhong Pa rk, Phillip Henzler, Ben Mildenhall, Jonathan T. Barron, and Ricardo Marti n-Brualla (Google Research)\n---------------------\nAdaptive Shells for Ef ficient Neural Radiance Field Rendering\n\nNeural radiance fields achieve unprecedented quality for novel view synthesis, but their volumetric formu lation remains expensive, requiring a huge number of samples to render hig h-resolution images. Volumetric encodings are essential to represent fuzzy geometry such as foliage and hair, and they ar...\n\n\nZian Wang and Tian chang Shen (NVIDIA, University of Toronto); Merlin Nimier-David and Nichol as Sharp (NVIDIA); Jun Gao (NVIDIA, University of Toronto); Alexander Kell er (NVIDIA); Sanja Fidler (NVIDIA, University of Toronto); and Thomas Müll er and Zan Gojcic (NVIDIA)\n---------------------\nIconShop: Text-Guided V ector Icon Synthesis with Autoregressive Transformers\n\nScalable Vector G raphics (SVG) is a popular vector image format that offers good support fo r interactivity and animation. Despite its appealing characteristics, crea ting custom SVG content can be challenging for users due to the steep lear ning curve required to understand SVG grammars or get familia...\n\n\nRong huan Wu, Wanchao Su, Kede Ma, and Jing Liao (City University of Hong Kong) \n---------------------\nNeural Stress Fields for Reduced-order Elastoplas ticity and Fracture\n\nThe material point method (MPM) is a versatile simu lation framework for large-deformation elastoplasticity and fracture. Howe ver, MPM's long runtime and large memory consumptions render it unsuitable for applications constrained by computation time and memory usage, e.g., virtual reality. To overcom...\n\n\nZeshun Zong and Xuan Li (University of California Los Angeles); Minchen Li (University of California Los Angeles , Carnegie Mellon University); Maurizio M. Chiaramonte (Meta Reality Labs Research); Wojciech Matusik (MIT CSAIL); Eitan Grinspun (University of Tor onto); Kevin Carlberg (Meta Reality Labs Research); Chenfanfu Jiang (Unive rsity of California Los Angeles); and Peter Yichen Chen (MIT CSAIL)\n----- ----------------\nActRay: Online Active Ray Sampling for Radiance Fields\n \nThanks to the high-quality reconstruction and photorealistic rendering, the Neural Radiance Field (NeRF) has garnered extensive attention and has been continuously improved. Despite its high visual quality, the prohibiti ve training time limits its practical application. Although significant ac celera...\n\n\nJiangkai Wu, Liming Liu, Yunpeng Tan, Quanlu Jia, Haodan Zh ang, and Xinggong Zhang (Peking University)\n---------------------\nSLANG. D: Fast, Modular and Differentiable Shader Programming\n\nWe introduce SLA NG.D, a shading language that incorporates first-class automatic different iation support derived from the Slang language. The new shading language a llows us to transform a Direct3D-based path tracer to be fully differentia ble with minor modifications to existing code. SLANG.D enables...\n\n\nSai Praveen Bangaru (MIT CSAIL), Lifan Wu (NVIDIA), Tzu-Mao Li (University of California San Diego), Jacob Munkberg (NVIDIA), Gilbert Bernstein (Univer sity of Washington), Jonathan Ragan-Kelley (MIT CSAIL), Aaron Lefohn (NVID IA), Fredo Durand (MIT CSAIL), and Yong He (NVIDIA)\n--------------------- \nDreamEditor: Text-Driven 3D Scene Editing with Neural Fields\n\nNeural f ields have achieved impressive advancements in view synthesis and scene re construction. However, editing these neural fields remains challenging due to the implicit encoding of geometry and texture information. In this pap er, we propose DreamEditor, a novel framework that enables users to pe...\ n\n\nJingyu Zhuang (Sun Yat-sen University); Chen Wang (University of Penn sylvania, Tsinghua University); Liang Lin (Sun Yat-sen University); Lingji e Liu (University of Pennsylvania); and Guanbin Li (Sun Yat-sen University )\n---------------------\nGroomGen: A High-Quality Generative Hair Model U sing Hierarchical Latent Representations\n\nDespite recent successes in ha ir acquisition that fits a high-dimensional hair model to a specific input subject, generative hair models, which establish general embedding spaces for encoding, editing, and sampling diverse hairstyles, are way less expl ored. In this paper, we present GroomGen, the fi...\n\n\nYuxiao Zhou (ETH Zürich), Menglei Chai and Alessandro Pepe (Google Inc.), Markus Gross (ETH Zürich), and Thabo Beeler (Google Inc.)\n---------------------\nReShader: View-Dependent Highlights for Single Image View-Synthesis\n\nIn recent ye ars, novel view synthesis from a single image has seen significant progres s thanks to the rapid advancements in 3D scene representation and image in painting techniques. While the current approaches are able to synthesize g eometrically consistent novel views, they often do not handle the ...\n\n\ nAvinash Paliwal and Brandon G. Nguyen (Texas A&M University), Andrii Tsar ov (Leia Inc.), and Nima Khademi Kalantari (Texas A&M University)\n------- --------------\nEXIM: A Hybrid Explicit-Implicit Representation for Text-G uided 3D Shape Generation\n\nThis paper presents a new text-guided techniq ue for generating 3D shapes. The technique leverages a hybrid 3D shape rep resentation, combining the strengths of explicit and implicit representati ons. Specifically, the explicit stage controls the generated topology of t he 3D shape and allows local mani...\n\n\nZhengzhe Liu, Jingyu Hu, and Ka- Hei Hui (The Chinese University of Hong Kong); Xiaojuan Qi (The University of Hong Kong); Daniel Cohen-Or (Tel-Aviv University); and Chi-Wing Fu (Th e Chinese University of Hong Kong)\n---------------------\nAn Adaptive Fas t-Multipole-Accelerated Hybrid Boundary Integral Equation Method for Accur ate Diffusion Curves\n\nIn theory, diffusion curves promise complex color gradations for infinite-resolution vector graphics. In practice, existing realizations suffer from poor scaling, discretization artifacts, or insuff icient support for rich boundary conditions. Previous applications of the boundary element method to d...\n\n\nSeungbae Bang (University of Toronto, Amazon); Kirill Serkh (University of Toronto); Oded Stein (University of Southern California, MIT); and Alec Jacobson (University of Toronto, Adobe )\n---------------------\nEgo3DPose: Capturing 3D Cues from Binocular Egoc entric Views\n\nWe present Ego3DPose, a highly accurate binocular egocentr ic 3D pose reconstruction system. The binocular egocentric setup offers pr acticality and usefulness in various applications, however, it remains lar gely under-explored. It has been suffering from low pose estimation accura cy due to viewing di...\n\n\nTaeho Kang and Kyungjin Lee (Seoul National U niversity), Jinrui Zhang (Central South University), and Youngki Lee (Seou l National University)\n---------------------\nNeural Collision Fields for Triangle Primitives\n\nWe present neural collision fields as an alternati ve to contact point sampling in physics simulations.\nOur approach is buil t on top of a novel smoothed integral formulation for the contact surface patches between two triangle meshes. By reformulating collisions as an int egral, we avoid issues of sam...\n\n\nRyan Zesch (Texas A&M University), V ismay Modi (University of Toronto), Shinjiro Sueda (Texas A&M University), and David Levin (University of Toronto)\n---------------------\nTexture A tlas Compression Based on Repeated Content Removal\n\nOptimizing the memor y footprint of 3D models can have a major impact on the user experiences d uring real-time rendering and streaming visualization, where the major mem ory overhead lies in the high-resolution texture data. In this work, we pr opose a robust and automatic pipeline to content-aware, lo...\n\n\nYuzhe L uo (State Key Laboratory of CAD & CG, Zhejiang University; LIGHTSPEED STUD IOS); Xiaogang Jin (State Key Laboratory of CAD & CG, Zhejiang University) ; Zherong Pan and Kui Wu (LIGHTSPEED STUDIOS); Qilong Kou and Xiajun Yang (Tencent Technology (Shenzhen) Co., LTD); and Xifeng Gao (LIGHTSPEED STUDI OS)\n---------------------\nProSpect: Prompt Spectrum for Attribute-Aware Personalization of Diffusion Models\n\nPersonalizing generative models off ers a way to guide image generation with user-provided references. Current personalization methods can invert an object or concept into the textual conditioning space and compose new natural sentences for text-to-image dif fusion models. However, representing and ed...\n\n\nYuxin Zhang (MAIS, Ins titute of Automation, Chinese Academy of Sciences; School of Artificial In telligence, University of Chinese Academy of Sciences); Weiming Dong (MAIS , Institute of Automation, Chinese Academy of Sciences; School of AI,Unive rsity of Chinese Academy of Sciences); Fan Tang (Institute of Computing Te chnology, Chinese Academy of Sciences); Nisha Huang (School of AI,Universi ty of Chinese Academy of Sciences; MAIS, Institute of Automation, Chinese Academy of Sciences); Haibin Huang and Chongyang Ma (Kuaishou Technology); Tong-Yee Lee (National Cheng-Kung University); Oliver Deussen (University of Konstanz); and Changsheng Xu (MAIS, Institute of Automation, Chinese A cademy of Sciences; School of Artificial Intelligence, University of Chine se Academy of Sciences)\n---------------------\nComputational Design of LE GO Sketch Art\n\nThis paper presents computational methods to aid the crea tion of LEGO Sketch models from simple input images. Beyond conventional L EGO mosaics, we aim to improve the expressiveness of LEGO models by utiliz ing LEGO tiles with sloping and rounding edges, together with rectangular bricks, to reproduce ...\n\n\nMingjun Zhou (The Chinese University of Hong Kong); Jiahao Ge (The Chinese University of Hong Kong, Qianzhi Technology Inc.); Hao Xu (Qianzhi Technology Inc.); and Chi-Wing Fu (The Chinese Uni versity of Hong Kong)\n---------------------\nBreak-A-Scene: Extracting Mu ltiple Concepts from a Single Image\n\nText-to-image model personalization aims to introduce a user-provided concept to the model, allowing its synt hesis in diverse contexts. However, current methods primarily focus on the case of learning a single concept from multiple images with variations in backgrounds and poses, and struggle when a...\n\n\nOmri Avrahami (The Heb rew University of Jerusalem), Kfir Aberman (Google Research), Ohad Fried ( Reichman University), Daniel Cohen-Or (Tel Aviv University), and Dani Lisc hinski (The Hebrew University of Jerusalem)\n---------------------\nGANeRF : Leveraging Discriminators to Optimize Neural Radiance Fields\n\nNeural R adiance Fields (NeRF) have shown impressive novel view synthesis results; nonetheless, even thorough recordings yield imperfections in reconstructio ns, for instance due to poorly observed areas or minor lighting changes.\n Our goal is to mitigate these imperfections from various sources with a... \n\n\nBarbara Roessle and Norman Müller (Technical University of Munich); Lorenzo Porzi, Samuel Rota Bulò, and Peter Kontschieder (Meta Reality Labs ); and Matthias Niessner (Technical University of Munich)\n--------------- ------\nA Neural Space-Time Representation for Text-to-Image Personalizati on\n\nA key aspect of text-to-image personalization methods is the manner in which the target concept is represented within the generative process. This choice greatly affects the visual fidelity, downstream editability, a nd disk space needed to store the learned concept. In this paper, we explo re a new t...\n\n\nYuval Alaluf, Elad Richardson, Gal Metzer, and Daniel C ohen-Or (Tel Aviv University)\n---------------------\nAniPortraitGAN: Anim atable 3D Portrait Generation from 2D Image Collections\n\nPrevious animat able 3D-aware GANs for human generation have primarily focused on either t he human head or full body. However, head-only videos are relatively uncom mon in real life, and full body generation typically does not deal with fa cial expression control and still has challenges in generating ...\n\n\nYu e Wu (Hong Kong University of Science and Technology), Sicheng Xu (Microso ft Research Asia), Jianfeng Xiang (Tsinghua University), Fangyun Wei (Micr osoft Research Asia), Qifeng Chen (Hong Kong University of Science and Tec hnology), and Jiaolong Yang and Xin Tong (Microsoft Research Asia)\n------ ---------------\nOpenSVBRDF: A Database of Measured Spatially-Varying Refl ectance\n\nWe present the first large-scale database of measured spatially -varying anisotropic reflectance, consisting of 1,000 high-quality near-pl anar SVBRDFs, spanning 9 material categories such as wood, fabric and meta l. Each sample is captured in 15 minutes, and represented as a set of high -resolution tex...\n\n\nXiaohe Ma, Xianmin Xu, Leyao Zhang, Kun Zhou, and Hongzhi Wu (State Key Laboratory of CAD&CG, Zhejiang Univerisity)\n------- --------------\nJoint Sampling and Optimisation for Inverse Rendering\n\nW hen dealing with difficult inverse problems such as inverse rendering, usi ng Monte Carlo estimated gradients to optimise parameters can slow down co nvergence due to variance. Averaging many gradient samples in each iterati on reduces this variance trivially. However, for problems that require tho usa...\n\n\nMartin Balint, Karol Myszkowski, Hans-Peter Seidel, and Gurpri t Singh (Max Planck Institute for Informatics)\n---------------------\nGAR M-LS: A Gradient-Augmented Reference-Map Method for Level-Set Fluid Simula tion\n\nThis paper presents a novel level-set method by combining gradient augmentation and reference mapping to enable high-fidelity interface trac king and surface tension flow simulation. At the center of our approach is a novel reference-map algorithm to concurrently convect level-set values and gradient...\n\n\nXingqiao Li (School of IST & National Key Lab. of AGI , Peking University); Xingyu Ni (School of CS & National Key Lab. of AGI, Peking University); Bo Zhu (Georgia Institute of Technology, Dartmouth Col lege); Bin Wang (Beijing Institute for General Artificial Intelligence); a nd Baoquan Chen (School of IST & National Key Lab. of AGI, Peking Universi ty)\n---------------------\nShapeSonic: Sonifying Fingertip Interactions f or Non-Visual Virtual Shape Perception\n\nFor sighted users, computer grap hics and virtual reality allow them to model and perceive imaginary object s and worlds. However, these approaches are inaccessible to blind and visu ally impaired (BVI) users, since they primarily rely on visual feedback. T o this end, we introduce ShapeSonic, a system ...\n\n\nJialin Huang (Georg e Mason University), Alexa Siu (Adobe Research), Rana Hanocka (University of Chicago), and Yotam Gingold (George Mason University)\n---------------- -----\nCurl Noise Jittering\n\nWe propose a method for implicitly generati ng blue noise point sets. Our method is based on the observations that cur l noise vector fields are volume-preserving and that jittering can be cons trued as moving points along the streamlines of a vector field. We demonst rate that the volume preservation k...\n\n\nJ. Andreas Bærentzen and Jeppe Revall Frisvad (Technical University of Denmark) and Jonàs Martínez (INRI A)\n\nRegistration Category: Full Access, Enhanced Access, Trade Exhibitor , Experience Hall Exhibitor END:VEVENT END:VCALENDAR