Short name:
BrinqTraq_v2
Detector:
Private
Description:
Custom trained yolo_v5 x-large model with input size of 1920x1920 as detector.
Reference:
H. Mahadik, N. Bamra, J. Rynne. Brinq Edge AI SDK v2.0, Arcturus Networks Inc. In , 2023.
Last submitted:
April 04, 2023 (1 year ago)
Published:
April 04, 2023 at 18:02:33 CET
Submissions:
1
Project page / code:
Open source:
No
Hardware:
1.9 GHZ, 2560 CUDA Cores
Runtime:
6.7 Hz
Benchmark performance:
Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
MOT20 | 79.5 | 77.4 | 63.1 | 910 (73.3) | 103 (8.3) | 26,830 | 77,899 | 84.9 | 94.2 | 61.5 | 65.0 | 67.7 | 75.9 | 70.8 | 78.6 | 83.8 | 6.0 | 1,169 (0.0) | 1,748 (0.0) |
Detailed performance:
Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
MOT20-04 | 88.8 | 86.7 | 70.5 | 588 | 19 | 14,173 | 16,117 | 94.1 | 94.8 | 68.5 | 72.7 | 74.7 | 79.0 | 79.3 | 79.8 | 84.4 | 6.8 | 273 | 394 |
MOT20-06 | 67.8 | 64.2 | 52.3 | 141 | 54 | 4,953 | 37,296 | 71.9 | 95.1 | 49.8 | 55.1 | 55.5 | 70.6 | 59.1 | 78.1 | 82.8 | 4.9 | 475 | 735 |
MOT20-07 | 80.7 | 72.7 | 61.9 | 86 | 3 | 2,433 | 3,846 | 88.4 | 92.3 | 57.1 | 67.6 | 63.9 | 73.1 | 75.5 | 78.8 | 85.2 | 4.2 | 111 | 128 |
MOT20-08 | 66.2 | 65.2 | 51.8 | 95 | 27 | 5,271 | 20,640 | 73.4 | 91.5 | 50.2 | 53.5 | 55.9 | 70.1 | 59.3 | 73.9 | 81.9 | 6.5 | 310 | 491 |
Raw data: