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

Short Name: TCFM
Complete name: TCFM
Description:

TCFM

Runtime:

1

Submissions:

3

Benchmark
performance:

MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
44.0086 38.7243 40.1714 23.1899 29.0661 5628 30603 2781 92 2140 76.7416

Detailed
performance:

Video ID MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
20 76.5130 74.8354 64.4530 69.3878 6.1224 889 1771 142 5 125 79.1573
21 15.0548 11.0548 24.1427 0.0000 40.0000 118 666 68 1 67 66.6986
22 15.4837 9.7206 17.6959 4.8544 32.0388 786 6841 577 14 499 68.9592
23 77.5647 72.2040 41.5950 57.1429 7.1429 94 811 101 6 86 81.3178
24 7.2891 -4.6334 16.7896 3.0395 58.3587 1055 10966 431 40 359 68.9725
25 56.4065 51.1407 53.2909 22.9885 0.5747 433 1881 333 8 131 78.0115
26 42.3656 40.2156 31.3563 35.2941 23.5294 445 1596 103 2 115 77.2376
27 75.1230 68.8238 43.6349 77.4194 3.2258 320 600 344 8 210 77.9452
28 29.6092 20.1343 32.4471 35.0000 20.0000 680 969 134 6 128 72.4928
29 47.5190 45.7148 39.0428 36.0000 20.0000 285 789 100 1 106 75.4145
30 52.8996 52.8996 43.4913 26.6667 6.6667 82 714 138 0 85 77.2874
31 68.8748 68.8748 50.8976 73.3333 4.4444 125 411 150 0 61 78.8645
32 42.0794 41.3237 42.2484 22.5000 37.5000 316 2588 160 1 168 75.1826