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:20240214T070243Z
LOCATION:Meeting Room C4.9+C4.10\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T172000
DTEND;TZID=Australia/Melbourne:20231212T173000
UID:siggraphasia_SIGGRAPH Asia 2023_sess162_papers_204@linklings.com
SUMMARY:High-Fidelity and Real-Time Novel View Synthesis for Dynamic Scene
 s
DESCRIPTION:Technical Papers\n\nHaotong Lin (State Key Laboratory of CAD &
  CG, Zhejiang University); Sida Peng (Zhejiang University); and Zhen Xu, T
 ao Xie, Xingyi He, Hujun Bao, and Xiaowei Zhou (State Key Laboratory of CA
 D & CG, Zhejiang University)\n\nThis paper aims to tackle the challenge of
  dynamic view synthesis from multi-view videos. The key observation is tha
 t while previous grid-based methods offer consistent rendering, they fall 
 short in capturing appearance details on a complex dynamic scene, a domain
  where multi-view image-based methods demonstrate the opposite properties.
  To combine the best of two worlds, we introduce a hybrid scene representa
 tion that consists of a grid-based geometry representation and a multi-vie
 w image-based appearance representation. Specifically, the dynamic geometr
 y is encoded as a 4D density function composed of spatiotemporal feature p
 lanes and a small MLP network, which globally models the scene structure a
 nd facilitates the rendering consistency. We represent the scene appearanc
 e by the original multi-view videos and a network that learns to predict t
 he color of a 3D point from image features, instead of totally memorizing 
 the appearance with networks, thereby naturally making the learning of net
 works easier. Our method is evaluated on five dynamic view synthesis datas
 ets including DyNeRF, ZJU-MoCap, NHR, DNA-Rendering and ENeRF-Outdoor data
 sets. The results show that the proposed representation exhibits state-of-
 the-art performance in rendering quality and can be trained quickly, while
  being efficient for real-time rendering with a speed of 79.8 FPS for 512x
 512 images (ZJU-MoCap dataset), on a single RTX 3090 GPU. The code is avai
 lable at https://zju3dv.github.io/im4d.\n\nRegistration Category: Full Acc
 ess\n\nSession Chair: Binh-Son Hua (Trinity College Dublin)
URL:https://asia.siggraph.org/2023/full-program?id=papers_204&sess=sess162
END:VEVENT
END:VCALENDAR