BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260817T171530Z
LOCATION:Hall B5 (1)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241205T172800
DTEND;TZID=Asia/Tokyo:20241205T174000
UID:siggraphasia_SIGGRAPH Asia 2024_sess136_papers_616@linklings.com
SUMMARY:MagicClay: Sculpting Meshes With Generative Neural Fields
DESCRIPTION:Amir Barda (Tel Aviv University), Vladimir Kim (Adobe Research
 ), Noam Aigerman (Université de Montréal), Amit Haim Bermano (Tel Aviv Uni
 versity), and Thibault Groueix (Adobe Research)\n\nThe recent developments
  in neural fields have brought phenomenal capabilities to the field of sha
 pe generation, but they lack crucial properties, such as incremental contr
 ol --- a fundamental requirement for artistic work. Triangular meshes, on 
 the other hand, are the representation of choice for most geometry-related
  tasks, offering efficiency and intuitive control, but do not lend themsel
 ves to neural optimization. \nTo support downstream tasks, previous art ty
 pically proposes a two-step approach, where first, a shape is generated us
 ing neural fields, and then a mesh is extracted for further processing. In
 stead, in this paper, we introduce a hybrid approach that maintains both a
  mesh and a Signed Distance Field (SDF) representations consistently. Usin
 g this representation, we introduce MagicClay --- an artist friendly tool 
 for sculpting regions of a mesh according to textual prompts while keeping
  other regions untouched.\nOur framework carefully and efficiently balance
 s consistency between the representations and regularizations in every ste
 p of the shape optimization. Relying on the mesh representation, we show h
 ow to render the SDF at higher resolutions and faster. In addition, we emp
 loy recent work in differentiable mesh reconstruction to adaptively alloca
 te triangles in the mesh where required, as indicated by the SDF.\nUsing a
 n implemented prototype, we demonstrate superior generated geometry compar
 ed to the state-of-the-art, and novel consistent control, allowing sequent
 ial prompt-based edits to the same mesh for the first time. \nWe will rele
 ase the code upon acceptance.\n\nRegistration Category: Full Access, Full 
 Access Supporter\n\nLanguage Format: English Language\n\nSession Chair: Ma
 nolis Savva (Simon Fraser University)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_616&sess=sess136
END:VEVENT
END:VCALENDAR