DWnet: Deep-Wide Network for 3D Action Recognition
This work addresses action recognition for computer vision applications, but it is incremental as it builds on existing deep and broad learning techniques.
The authors tackled 3D action recognition by proposing DWnet, a deep-wide network combining deep structure with broad learning system, which reduces testing time to near real-time and improves feature capture over existing methods, achieving competitive results on two skeletal datasets.
We propose in this paper a deep-wide network (DWnet) which combines the deep structure with the broad learning system (BLS) to recognize actions. Compared with the deep structure, the novel model saves lots of testing time and almost achieves real-time testing. Furthermore, the DWnet can capture better features than broad learning system can. In terms of methodology, we use pruned hierarchical co-occurrence network (PruHCN) to learn local and global spatial-temporal features. To obtain sufficient global information, BLS is used to expand features extracted by PruHCN. Experiments on two common skeletal datasets demonstrate the advantage of the proposed model on testing time and the effectiveness of the novel model to recognize the action.