From Coarse to Fine: Hierarchical Multi-scale Temporal Information Modeling via Sub-group Convolution for Video Action Recognition.

Fengwen Cheng, Huicheng Zheng, Zehua Liu

Browse the full IJCNN paper archive.