Prototypical Cross-Attention Networks for Multiple Object Tracking and Segmentation

Benchmark Model Rank Results
multi-object-tracking-and-segmentation-on-3PCAN#3mMOTSA: 27.4
multi-object-tracking-and-segmentation-on-3QDTrack-mots-fix#4mMOTSA: 23.5
multi-object-tracking-and-segmentation-on-3QDTrack-mots#5mMOTSA: 22.5
multi-object-tracking-and-segmentation-on-3MaskTrackRCNN#6mMOTSA: 12.3
multi-object-tracking-and-segmentation-on-3STEm-Seg#7mMOTSA: 12.2
multi-object-tracking-and-segmentation-on-3SortIoU#8mMOTSA: 10.3
video-instance-segmentation-on-bdd100k-valPCAN#1mMOTSA: 27.4
video-instance-segmentation-on-bdd100k-valQDTrack-mots-fix#2mMOTSA: 23.5
video-instance-segmentation-on-bdd100k-valQDTrack-mots#3mMOTSA: 22.5
video-instance-segmentation-on-bdd100k-valMaskTrackRCNN#4mMOTSA: 12.3
video-instance-segmentation-on-bdd100k-valSTEm-Seg#5mMOTSA: 12.2
video-instance-segmentation-on-bdd100k-valSortIoU#6mMOTSA: 10.3
video-instance-segmentation-on-youtube-vis-1PCAN(ResNet-50)#28mask AP: 36.1AP50: 54.9AP75: 39.4AR1: 36.3AR10: 41.6