Submission details of Tracktor++

Name Tracktor++
Paper Link https://arxiv.org/pdf/1903.05625.pdf
Code Link https://github.com/phil-bergmann/tracking_wo_bnw
MOTA 0.196963
MOTP 0.269225
IDs 7026
False Positives 79573
False Negatives 681672
Input N/A
Runtime 0.2 s
Environment 1 GPU (Titan X)
Abstract The problem of tracking multiple objects in a video sequence poses several challenging tasks. For tracking-bydetection, these include object re-identification, motion prediction and dealing with occlusions. We present a tracker (without bells and whistles) that accomplishes tracking without specifically targeting any of these tasks, in particular, we perform no training or optimization on tracking data. To this end, we exploit the bounding box regression of an object detector to predict the position of an object in the next frame, thereby converting a detector into a Tracktor. We demonstrate the potential of Tracktor and provide a new state-of-the-art on three multi-object tracking benchmarks by extending it with a straightforward re-identification and camera motion compensation. We then perform an analysis on the performance and failure cases of several state-of-the-art tracking methods in comparison to our Tracktor. Surprisingly, none of the dedicated tracking methods are considerably better in dealing with complex tracking scenarios, namely, small and occluded objects or missing detections. However, our approach tackles most of the easy tracking scenarios. Therefore, we motivate our approach as a new tracking paradigm and point out promising future research directions. Overall, Tracktor yields superior tracking performance than any current tracking method and our analysis exposes remaining and unsolved tracking challenges to inspire future research directions.

Detailed results

Per-sequence results

sequence name MOTA MOTP IDs False Positives False Negatives
cubberly-auditorium-2019-04-22_1 26.13 27.14 91 1459 8224
discovery-walk-2019-02-28_0 18.54 27.58 132 1665 9052
discovery-walk-2019-02-28_1 29.05 27.31 119 1486 8673
food-trucks-2019-02-12_0 33.26 26.89 313 1461 41006
gates-ai-lab-2019-04-17_0 15.17 21.59 226 8327 13352
gates-basement-elevators-2019-01-17_0 46.10 23.79 93 400 6193
gates-foyer-2019-01-17_0 31.11 22.56 101 3952 2966
gates-to-clark-2019-02-28_0 38.14 23.52 45 477 1056
hewlett-class-2019-01-23_0 50.62 22.96 186 451 7528
hewlett-class-2019-01-23_1 65.95 21.14 27 513 1209
huang-2-2019-01-25_1 22.70 25.81 43 1286 3796
huang-intersection-2019-01-22_0 8.54 30.57 295 6525 36067
indoor-coupa-cafe-2019-02-06_0 22.38 26.62 501 3791 46623
lomita-serra-intersection-2019-01-30_0 10.16 29.17 87 3152 15262
meyer-green-2019-03-16_1 7.22 28.64 140 2602 18977
nvidia-aud-2019-01-25_0 16.84 27.88 297 4899 22353
nvidia-aud-2019-04-18_1 39.63 24.25 32 839 3949
nvidia-aud-2019-04-18_2 19.11 28.78 62 3036 5138
outdoor-coupa-cafe-2019-02-06_0 8.91 29.58 365 4461 40865
quarry-road-2019-02-28_0 -11.81 29.94 24 2587 3995
serra-street-2019-01-30_0 3.73 29.88 403 6117 34894
stlc-111-2019-04-19_1 61.07 21.09 40 339 2051
stlc-111-2019-04-19_2 51.40 21.67 35 441 1288
tressider-2019-03-16_2 9.62 31.07 136 2942 21014
tressider-2019-04-26_0 15.49 29.03 1088 5270 94825
tressider-2019-04-26_1 22.67 26.23 1018 3666 144200
tressider-2019-04-26_3 17.98 28.83 1127 7429 87116
total 19.70 26.92 7026 79573 681672