Continuous-Time Human Motion Field from Events
Ziyun Wang, Ruijun Zhang, Zi-Yan Liu, Yufu Wang, Kostas Daniilidis · 2024
Read Continuous-Time Human Motion Field from Events by Ziyun Wang, Ruijun Zhang, Zi-Yan Liu, Yufu Wang, Kostas Daniilidis (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-01747.txt.
Description
Ziyun Wang and four co-authors estimate a continuous time human motion field directly from event camera streams. Frame based human mesh recovery suffers aliasing and motion blur from limited temporal resolution. Earlier event methods optimize over fixed pose sets at high frame rates, which grows prohibitively expensive as resolution rises.
A recurrent feed forward network predicts motion in the latent space of possible human motions. The motion field is a time implicit function, so poses can be queried in parallel at arbitrary temporal resolutions. A new hardware synchronized high speed dataset, built with a beam splitter, fills a benchmark gap. On it, joint errors drop by 23.8 percent against prior event methods while compute time falls by 69 percent.
Catalog record
| CREATOR | Ziyun Wang, Ruijun Zhang, Zi-Yan Liu, Yufu Wang, Kostas Daniilidis |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-01747