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:20260817T171532Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241203T172800
DTEND;TZID=Asia/Tokyo:20241203T174000
UID:siggraphasia_SIGGRAPH Asia 2024_sess111_papers_539@linklings.com
SUMMARY:V^3: Viewing Volumetric Videos on Mobiles via Streamable 2D Dynami
 c Gaussians
DESCRIPTION:Penghao Wang, Zhirui Zhang, Liao Wang, Kaixin Yao, and Siyuan 
 Xie (ShanghaiTech University, NeuDim Inc.); Jingyi Yu (ShanghaiTech Univer
 sity); Minye Wu (KU Leuven); and Lan Xu (ShanghaiTech University)\n\nExper
 iencing high-fidelity volumetric video as seamlessly as 2D videos is a lon
 g-held dream. However, current dynamic 3DGS methods, despite their high re
 ndering quality, face challenges in streaming on mobile devices due to com
 putational and bandwidth constraints. In this paper, we introduce V3 (View
 ing Volumetric Videos), a novel approach that enables high-quality mobile 
 rendering through the streaming of dynamic Gaussians. Our key innovation i
 s to view dynamic 3DGS as 2D videos, facilitating the use of hardware vide
 o codecs. Additionally, we propose a two-stage training strategy to reduce
  storage requirements with rapid training speed. The first stage employs h
 ash encoding and shallow MLP to learn motion, then reduces the number of G
 aussians through pruning to meet the streaming requirements, while the sec
 ond stage fine-tunes other Gaussian attributes using residual entropy loss
  and temporal loss to improve temporal continuity. This strategy, which di
 sentangles motion and appearance, maintains high rendering quality with co
 mpact storage requirements. Meanwhile, we designed a multi-platform player
  to decode and render 2D Gaussian videos. Extensive experiments demonstrat
 e the effectiveness of V3, outperforming other methods by enabling high-qu
 ality rendering and streaming on common devices, which is unseen before. A
 s the first to stream dynamic Gaussians on mobile devices, our companion p
 layer offers users an unprecedented volumetric video experience, including
  smooth scrolling and instant sharing.\n\nRegistration Category: Full Acce
 ss, Full Access Supporter\n\nLanguage Format: English Language\n\nSession 
 Chair: Yifan Peng (The University of Hong Kong)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_539&sess=sess111
END:VEVENT
END:VCALENDAR