요약
Abstract
1. Introduction
2. Proposed Method
2.1 3D Convolutional Neural Network:
2.2 Fully Connected Block
3 Experimental Results
3.1 UT-I Dataset
3.2 TV-Human Interaction Dataset
3.3 Implementation
3.4 Comparison with State-of-the-art
4. Conclusions and Future Work
Acknowledgments
References