BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260417T190049Z
LOCATION:West Building\, Rooms 220-222
DTSTART;TZID=America/Los_Angeles:20250813T104500
DTEND;TZID=America/Los_Angeles:20250813T121500
UID:siggraph_SIGGRAPH 2025_sess153@linklings.com
SUMMARY:Avatars
DESCRIPTION:Tiny is not small enough: High quality, low-resource facial an
 imation models through hybrid knowledge distillation\n\nThe goal of this w
 ork is to train lip sync animation models that can run in real-time and on
 -device. We design a two-stage knowledge distillation framework to distill
  large, high-quality models. Our results show that we can train small mode
 ls with low latency and a comparatively small loss in qualit...\n\n\nZhen 
 Han, Mattias Teye, Derek Yadgaroff, and Judith Bütepage (Electronic Arts)\
 n---------------------\nTransforming Unstructured Hair Strands into Proced
 ural Hair Grooms\n\nRecent methods have been developed to reconstruct 3D h
 air strand geometry from images. We introduce an inverse hair grooming pip
 eline to transform these unstructured hair strands into procedural hair gr
 ooms controlled by a small set of guide strands and artist-friendly groomi
 ng operators, enabling e...\n\n\nWesley Chang and Andrew Russell (Universi
 ty of California San Diego); Stephane Grabli, Matt Chiang, Christophe Hery
 , and Doug Roble (Meta); Ravi Ramamoorthi and Tzu-Mao Li (University of Ca
 lifornia San Diego); and Olivier Maury (Meta)\n---------------------\nStre
 amME: Simplify 3D Gaussian Avatar within Live Stream\n\nThe StreamME takes
  live stream video as input to enable rapid 3D head avatar reconstruction.
  It achieves impressive speed, capturing the basic facial appearance withi
 n 10-seconds and reaching high-quality fidelity within 5-minutes. StreamME
  reconstructs facial features through on-the-fly training, a...\n\n\nLuchu
 an Song (University of Rochester, Adobe Research); Yang Zhou, Zhan Xu, Yi 
 Zhou, and Deepali Aneja (Adobe Research); and Chenliang Xu (University of 
 Rochester)\n---------------------\nSqueezeMe: Mobile-Ready Distillation of
  Gaussian Full-Body Avatars\n\nExisting Gaussian Splatting avatars require
  desktop GPUs, limiting mobile device use. SqueezeMe converts these avatar
 s into a lightweight representation, enabling real-time animation and rend
 ering on mobile devices. By distilling the corrective decoder into an effi
 cient linear model, SqueezeMe achie...\n\n\nForrest Iandola, Stanislav Pid
 horskyi, Igor Santesteban, Divam Gupta, Anuj Pahuja, Nemanja Bartolovic, F
 rank Yu, Emanuel Garbin, Tomas Simon, and Shunsuke Saito (Meta)\n---------
 ------------\nAvatars - Interactive Discussion\n\nAfter the summary presen
 tations, attendees will participate in an interactive discussion. Outside 
 the room will be a series of poster boards for authors to gather around wi
 th the audience. Authors are invited to bring any material related to thei
 r paper that could instigate further conversation such...\n\n-------------
 --------\nScaffoldAvatar: High-Fidelity Gaussian Avatars with Patch Expres
 sions\n\nScaffoldAvatar presents a novel approach for reconstructing ultra
 -high fidelity animatable head avatars, which can be rendered in real-time
 . Our method operates on patch-based local expression features and synthes
 izes 3D Gaussians dynamically by leveraging tiny scaffold MLPs. We employ 
 color-based d...\n\n\nShivangi Aneja (Technical University of Munich, Disn
 eyResearch|Studios); Sebastian Weiss, Irene Baeza, Prashanth Chandran, and
  Gaspard Zoss (DisneyResearch|Studios); Matthias Niessner (Technical Unive
 rsity Munich); and Derek Bradley (DisneyResearch|Studios)\n---------------
 ------\nModel See Model Do: Speech-Driven Facial Animation with Style Cont
 rol\n\nModelSeeModelDo presents a speech-driven 3D facial animation method
  using a latent diffusion model conditioned on a reference clip to capture
  nuanced performance styles. A novel "style basis" mechanism extracts key 
 poses to guide generation, achieving expressive, temporally coherent anima
 tions with ...\n\n\nYifang Pan (University of Toronto; Jali Research, Cana
 da); Karan Singh (University of Toronto); and Luiz Gustavo Hafemann (Ubiso
 ft)\n\nInterest Area: Research & Education\n\nRecording: Livestreamed, Not
  Livestreamed, Recorded, Not Recorded\n\nRegistration Category: Full Confe
 rence, Virtual Access, Wednesday\n\nSession Chair: Bernd Bickel (ETH Züric
 h, Google)
END:VEVENT
END:VCALENDAR