Improved Collaborative Representation Classifier Based on -Regularized for Human Action Recognition

Figure 1

Three DCNNs architecture for a depth action sequence to extract features.