Challenge and Dataset on Large-scale Human-centric Video Analysis in Complex Events (HiEve)


Short Name: refine bytetrack
Complete name: bytetrack using yolov5
Description:

refine bytetrack using yolov5

Runtime:

4fps

Submissions:

1

Benchmark
performance:

MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
46.8590 42.4363 55.9546 18.9927 33.5782 4472 32010 544 77 1516 77.6597

Detailed
performance:


Video ID MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
20 64.8449 64.5094 76.2603 53.0612 18.3673 1541 2629 24 1 123 78.6311
21 5.7827 1.7827 23.0828 0.0000 56.6667 124 813 8 1 22 61.2602
22 30.8849 25.9451 33.1847 11.6505 37.8641 286 6242 181 12 428 73.2299
23 66.0125 60.6518 60.2967 50.0000 14.2857 476 1019 29 6 86 76.6335
24 14.4740 5.2341 22.4115 2.4316 63.5258 250 11147 90 31 183 73.9071
25 36.9071 31.6413 62.1084 0.5747 3.4483 914 2908 9 8 115 75.3089
26 51.8548 49.7048 43.4541 41.1765 41.1765 46 1719 26 2 71 81.1857
27 76.5401 71.0283 71.2131 66.1290 8.0645 127 1007 58 7 165 81.6905
28 56.4548 45.4007 61.1459 20.0000 40.0000 161 916 26 7 77 76.0626
29 64.9084 63.1042 62.3002 56.0000 16.0000 80 676 29 1 79 80.0414
30 58.0938 58.0938 69.1855 24.4444 2.2222 41 778 12 0 53 78.4353
31 83.6207 83.6207 88.4141 86.6667 4.4444 81 277 3 0 18 81.1948
32 57.0321 56.2764 59.0316 27.5000 27.5000 345 1879 49 1 96 76.6739