Submission details of DeepSORT

Name DeepSORT
Paper Link https://arxiv.org/abs/1703.07402
Code Link https://github.com/nwojke/deep_sort
MOTA 0.232031
MOTP 0.24579
IDs 5296
False Positives 78947
False Negatives 650478
Input N/A
Runtime 0.025 s
Environment 1 GPU (Titan X)
Abstract Simple Online and Realtime Tracking (SORT) is a pragmatic approach to multiple object tracking with a focus on simple, effective algorithms. In this paper, we integrate appearance information to improve the performance of SORT. Due to this extension we are able to track objects through longer periods of occlusions, effectively reducing the number of identity switches. In spirit of the original framework we place much of the computational complexity into an offline pre-training stage where we learn a deep association metric on a large-scale person re-identification dataset. During online application, we establish measurement-to-track associations using nearest neighbor queries in visual appearance space. Experimental evaluation shows that our extensions reduce the number of identity switches by 45%, achieving overall competitive performance at high frame rates.

Detailed results

Per-sequence results

sequence name MOTA MOTP IDs False Positives False Negatives
cubberly-auditorium-2019-04-22_1 29.17 26.00 92 1564 7715
discovery-walk-2019-02-28_0 23.57 24.62 111 1167 8901
discovery-walk-2019-02-28_1 34.11 23.46 87 1228 8231
food-trucks-2019-02-12_0 36.84 22.91 236 1131 39119
gates-ai-lab-2019-04-17_0 12.00 20.16 180 10276 12267
gates-basement-elevators-2019-01-17_0 48.73 21.32 88 429 5842
gates-foyer-2019-01-17_0 34.14 21.02 39 3886 2785
gates-to-clark-2019-02-28_0 38.65 20.81 29 441 1095
hewlett-class-2019-01-23_0 55.11 20.63 136 758 6528
hewlett-class-2019-01-23_1 83.16 20.16 13 168 684
huang-2-2019-01-25_1 27.35 23.95 33 1272 3512
huang-intersection-2019-01-22_0 12.85 27.63 175 6313 34375
indoor-coupa-cafe-2019-02-06_0 23.33 24.94 405 4688 45202
lomita-serra-intersection-2019-01-30_0 18.98 26.63 33 2095 14557
meyer-green-2019-03-16_1 11.82 25.71 99 2224 18319
nvidia-aud-2019-01-25_0 20.71 27.53 227 4908 21131
nvidia-aud-2019-04-18_1 35.42 20.26 24 1113 4019
nvidia-aud-2019-04-18_2 19.73 27.95 58 3339 4776
outdoor-coupa-cafe-2019-02-06_0 10.53 26.95 257 4352 40270
quarry-road-2019-02-28_0 -4.54 27.15 23 2337 3816
serra-street-2019-01-30_0 8.01 27.09 138 5056 34377
stlc-111-2019-04-19_1 61.61 19.66 38 359 1999
stlc-111-2019-04-19_2 53.28 20.82 24 433 1239
tressider-2019-03-16_2 14.21 24.94 109 2706 20052
tressider-2019-04-26_0 17.99 25.99 872 5273 92037
tressider-2019-04-26_1 27.08 24.10 868 3645 135879
tressider-2019-04-26_3 22.47 26.87 902 7786 81751
total 23.20 24.58 5296 78947 650478