Papers of the day   All papers

Contact and Human Dynamics from Monocular Video

Comments

AK: Contact and Human Dynamics from Monocular Video pdf: https://arxiv.org/pdf/2007.11678.pdf abs: https://arxiv.org/abs/2007.11678 project page: https://geometry.stanford.edu/projects/human-dynamics-eccv-2020/ https://t.co/2tIAjrHvQD

4 replies, 502 likes


Davis Rempe: Our new #ECCV2020 paper uses learned foot contact estimation and trajectory optimization to reconstruct physically-plausible 3D human motion from video. Check it out! Paper: http://arxiv.org/abs/2007.11678 Video: http://youtu.be/qR9KW6JzXX4 Project: http://geometry.stanford.edu/projects/human-dynamics-eccv-2020/ https://t.co/9zc5PqIf3r

3 replies, 132 likes


Angjoo Kanazawa: video -> 3D human + floor -> physics based optimization = physically-plausible motion, very cool!! Congrats @davrempe @jimei_yang @AaronHertzmann @RubenEVillegas Bryan and Leo!!

2 replies, 96 likes


Tim Soret: Mind-blowing progress. A new neural network to generate motion data from ordinary 2D videos, with solid foot anchorage thanks to contact estimation refined by physics simulation.

4 replies, 89 likes


Ankur Handa: impressive results using a combination of synthetic data, ground plane and physics based optimisation for smooth human motion estimation.

1 replies, 46 likes


Aaron Hertzmann: Our new ECCV 2020 paper:

0 replies, 42 likes


arxiv: Contact and Human Dynamics from Monocular Video. http://arxiv.org/abs/2007.11678 https://t.co/Nqr1jYJqRe

0 replies, 5 likes


Davis Rempe: Come by our #ECCV2020 poster session to learn more! We're live now until 8am PDT (16:00 UTC+1) or later today from 4-6pm PDT (00:00-02:00 UTC+1). https://papers.eccv2020.eu/paper/2918/

0 replies, 5 likes


Content

Found on Jul 24 2020 at https://arxiv.org/pdf/2007.11678.pdf

PDF content of a computer science paper: Contact and Human Dynamics from Monocular Video