BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030653Z
LOCATION:Meeting Room S221\, Level 2
DTSTART;TZID=Asia/Hong_Kong:20251215T150000
DTEND;TZID=Asia/Hong_Kong:20251215T151100
UID:siggraphasia_SIGGRAPH Asia 2025_sess111_papers_1641@linklings.com
SUMMARY:Variational Neural Surfacing of 3D Sketches
DESCRIPTION:Yutao Zhang (Université de Montréal), Stephanie Wang (Adobe Re
 search), and Mikhail Bessmeltsev (Université de Montréal)\n\n3D sketches a
 re an effective representation of a 3D shape, convenient to create via mod
 ern Virtual or Augmented Reality (VR/AR) interfaces or from 2D sketches. F
 or 3D sketches drawn by designers, human observers can consistently imagin
 e the surface they imply, yet reconstructing such a surface with modern me
 thods remains an open problem. Existing methods either assume a clean, wel
 l-structured 3D curve network (while in reality most 3D sketches are rough
  and unstructured), or make no effort to produce a surface that aligns wit
 h the artist's intent. We propose a novel method that addresses this chall
 enge by designing a system that reconstructs a surface that aligns with hu
 man perception from a clean or rough set of 3D sketches. As the topology o
 f the desired surface is unknown, we use an implicit neural surface repres
 entation, parameterized via its gradient field.\n\nAs suggested by previou
 s perception and modelling literature, human observers tend to imagine the
  surface by interpreting some of the input strokes as \emph{representative
  flow-lines}, related to the lines of curvature, and imagining the surface
  whose curvature agrees with those. Inspired by these observations, we des
 ign a novel loss that finds the surface with the smoothest principal curva
 ture field aligned with the input strokes. Together with approximation and
  piecewise smoothness requirements, we formulate a variational optimizatio
 n that performs robustly on a wide variety of 3D sketches. We validate our
  algorithmic choices via a series of qualitative and quantiative evaluatio
 ns, and comparisons to ground truth surfaces and previous methods.\n\nRegi
 stration Category: Full Access, Full Access Supporter\n\nSession Chair: Li
 gang Liu (University of Science and Technology of China)\n\n
END:VEVENT
END:VCALENDAR