Fig. The semantic categories of PanopticVideo-300 dataset. All fixations in our set are collected by WinDB.
- Video Clips (300):
Training set: [240 clips]https://pan.baidu.com/s/1LeiX-p9YsAhrqTd2jq0Dfw;
Testing set: [60 clips]https://pan.baidu.com/s/1LeiX-p9YsAhrqTd2jq0Dfw.