Dynamic Warping Network for Semantic Video Segmentation

Figure 2

The overall structure of our DWNet framework. FRM denotes the flow refinement module. FG-Conv denotes the flow-guided convolution. Feature consistency loss and prediction consistency loss are both the temporal consistency loss, which improves the temporal consistency of video segmentation. The dotted lines denote that the components are only used in the training phase and will be removed in the inference phase.