Presentation

Novel View Synthesis of Human Interactions From Sparse Multi-view Videos
SessionRoundtable Session: Neural Pets, People and Avatars
Authors
Qing Shuai
Chen Geng
Qi Fang
Sida Peng
Wenhao Shen
Xiaowei Zhou
Hujun Bao
Event Type
Technical Paper
Keywords
Rendering
Interest Areas
Research & Education
Presentation Types
In Person
Registration Categories
Full Conference Supporter
Full Conference
Exhibitor Additional Full Conference
Exhibitor Full Conference
Recordings
This session WILL NOT be recorded.
TimeThursday, 11 August 20222:49pm - 2:54pm PDT
LocationEast Building, Ballroom A/B
DescriptionThis paper presents a novel system for generating editable, free-viewpoint videos of multiple human performers from very sparse RGB cameras. The system is able to decompose the scene into different layers without instance segmentation as input and reconstruct a layered neural representation to enable novel view synthesis and scene editing.
Authors
Qing Shuai
State Key Lab of CAD and CG, State Key Lab of CAD and CG, Zhejiang University
Chen Geng
State Key Lab of CAD and CG, State Key Lab of CAD and CG, Zhejiang University
Qi Fang
State Key Lab of CAD and CG, State Key Lab of CAD and CG, Zhejiang University
Sida Peng
State Key Lab of CAD and CG, State Key Lab of CAD and CG, Zhejiang University
Wenhao Shen
State Key Lab of CAD and CG, State Key Lab of CAD and CG, Zhejiang University
Xiaowei Zhou
State Key Lab of CAD and CG, State Key Lab of CAD and CG, Zhejiang University
Hujun Bao
State Key Lab of CAD and CG, State Key Lab of CAD and CG, Zhejiang University