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

Short Name: MPNTrack
Complete name: Learning a Neural solver for Multiple Object Track
Description:

Learning a Neural solver for Multiple Object Track

Runtime:

243

Submissions:

1

Benchmark
performance:

MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
47.9943 42.8249 53.3228 33.5782 23.8195 7756 27236 1243 90 1836 75.6815

Detailed
performance:

Video ID MOTA w_MOTA IDF1 MT ML FP FN ID_Sw ID_Sw_DT Frag MOTP
20 77.6949 76.0173 72.5871 69.3878 6.1224 851 1751 59 5 114 79.8156
21 11.4656 7.4656 30.7442 10.0000 23.3333 341 506 41 1 48 67.6419
22 25.7546 19.1682 29.7221 10.6796 24.2718 807 6043 357 16 474 68.4532
23 77.4532 72.0925 60.4776 64.2857 14.2857 85 888 38 6 62 79.4600
24 14.0049 1.4863 20.8329 5.1672 49.5441 1082 10126 342 42 418 69.9653
25 60.7543 57.4632 78.2383 65.5172 0.0000 1248 1086 49 5 88 76.8752
26 33.8441 31.6941 42.4476 41.1765 17.6471 999 1428 34 2 104 74.1537
27 72.4464 67.7220 62.3203 67.7419 4.8387 426 848 126 6 138 76.9309
28 39.5973 31.7015 46.0806 40.0000 15.0000 729 756 45 5 80 71.1093
29 41.3053 39.5011 52.1699 52.0000 20.0000 595 691 27 1 78 73.9894
30 57.8921 57.8921 64.7658 31.1111 2.2222 128 657 50 0 95 73.2991
31 80.7169 80.7169 74.8466 86.6667 4.4444 114 284 27 0 28 77.6751
32 51.3989 50.6432 58.7282 22.5000 25.0000 351 2172 48 1 109 74.4100