b
Discover
Search
About
Rethinking Motion Representation: Residual Frames with 3D ConvNets for Better Action Recognition
2020
·
arXiv