MiCT: Mixed 3D/2D Convolutional Tube for Human Action Recognition.

Yizhou Zhou Xiaoyan Sun Zheng-Jun Zha Wenjun Zeng

Published in: CVPR (2018)

Keyphrases

action recognition
image registration
pose estimation
artificial neural networks
cross sections
databases
learning algorithm
artificial intelligence
image processing
multiscale
video sequences
spatio temporal
d objects
graphical models
sparse coding
deep learning