Residual Frames with Efficient Pseudo-3D CNN for Human Action Recognition

Re-assign community