3D Multi-Object Tracking on WILDTRACK ground plane benchmark leaderboard · SOTA2 Research

3D Multi-Object Tracking on WILDTRACK ground plane

78.4IDF1Deep-Occlusion + KSP + ptrack

Evaluation Results

Method	Links
Deep-Occlusion + KSP + ptrack2020.01		78.4	84.4	73.1	72	74	25	2,007	5,830	103	95	72.2	60.3	0.75
Faster-RCNN(VGG16) + MV-GLMB-OC2020.01		76.5	84.5	70	119	118	47	545	1,621	104	81	65.3	71.9	0.72
Faster-RCNN(VGG16) + MS-GLMB2020.01		75.5	76.8	74.3	98	104	82	1,114	1,716	179	116	61.5	65.8	0.88
YOLOv3 + MV-GLMB-OC2020.01		74.3	85	75.9	136	111	37	424	1,333	104	86	69.7	73.2	0.69
YOLOv3 + MS-GLMB2020.01		74.2	79	69.9	116	85	83	841	1,951	139	105	61.9	68.3	0.81
Deep-Occlusion + GLMB2020.01		72.5	82.7	72.2	160	86	39	960	990	107	64	70.1	63.1	0.73