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


Short Name: multi human tracker
Complete name: multi human tracker with public detection
Description:

na

Runtime:

na

Submissions:

4

Benchmark
performance:

MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
46.8733 41.5890 49.4597 28.3316 27.7020 5533 29302 2181 92 2679 75.8452

Detailed
performance:


Video ID MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
20 80.3604 78.3473 68.5069 69.3878 6.1224 613 1614 116 6 191 78.7533
21 18.4447 14.4447 31.5978 0.0000 23.3333 197 532 89 1 60 65.8904
22 15.7000 9.5253 21.2962 9.7087 20.3883 1202 6266 715 15 733 68.7774
23 82.3372 76.0830 59.6867 71.4286 14.2857 118 595 79 7 106 80.6110
24 8.9271 -2.1012 18.5472 1.8237 61.0942 743 11182 307 37 375 68.2188
25 62.5494 57.2836 75.2076 48.8506 0.0000 706 1407 161 8 154 73.3626
26 42.5000 40.3500 36.2522 35.2941 23.5294 409 1648 82 2 134 76.4174
27 77.9374 71.6382 62.9404 75.8065 1.6129 225 708 188 8 274 78.2592
28 35.0178 25.5429 39.1985 35.0000 15.0000 630 905 111 6 145 71.7155
29 51.4528 49.6486 45.4131 36.0000 20.0000 241 769 76 1 119 75.5285
30 55.5724 55.5724 60.4527 24.4444 2.2222 89 697 95 0 96 77.3669
31 77.2232 77.2232 73.7176 77.7778 4.4444 72 386 44 0 82 78.7051
32 43.3081 42.5524 50.0785 25.0000 35.0000 288 2593 118 1 210 74.9182