MOT17Det Results

Click on a measure to sort the table accordingly. See below for a more detailed description.



Benchmark Statistics

TrackerAPMODAMODPFAFTPFPFNRecallPrecisionF1 Hz
AInnoDetV2
1. using public detections
0.89
±0.05
-58.4
±126.7
79.1 29.5 107,733 174,608 6,831 94.0 38.2 54.3 296.0
AInnovation: PC Attention Net
ACF
2. using public detections
0.32
±0.00
18.1
±0.0
72.1 2.8 37,312 16,539 77,252 32.6 69.3 44.3 74.0
P. Dollar, R. Appel, S. Belongie, P. Perona. Fast Feature Pyramids for Object Detection. In TPAMI, 2014.
PA_Det_NJ
3. using public detections
0.89
±0.06
82.1
±20.1
79.3 2.4 108,480 14,417 6,084 94.7 88.3 91.4 59.2
PA_TECH_NJ
YTLAB
4. using public detections
0.89
±0.07
76.7
±13.1
80.2 2.8 104,555 16,685 10,009 91.3 86.2 88.7 22.3
Z. Cai, Q. Fan, R. Feris, N. Vasconcelos. A unified multi-scale deep convolutional neural network for fast object detection. In European Conference on Computer Vision, 2016.
DPM
5. using public detections
0.61
±0.14
31.2
±10.8
75.8 7.1 78,007 42,308 36,557 68.1 64.8 66.4 19.7
P. Felzenszwalb, R. Girshick, D. McAllester, D. Ramanan. Object Detection with Discriminatively Trained Part Based Models. In TPAMI, 2010.
F_ViPeD_B
6. using public detections
0.89
±0.06
-14.4
±115.1
77.4 20.8 106,698 123,194 7,831 93.2 46.4 62.0 14.8
L. Ciampi, N. Messina, F. Falchi, C. Gennaro, G. Amato. Virtual to Real adaptation of Pedestrian Detectors for Smart Cities. In arXiv preprint arXiv:2001.03032, 2020.
YLHDv2
7. using public detections
0.46
±0.11
56.9
±12.5
73.2 2.5 80,093 14,938 34,471 69.9 84.3 76.4 11.8
https://arxiv.org/abs/1612.08242
VDet
8. using public detections
0.44
±0.19
44.7
±19.3
75.7 1.0 56,980 5,765 57,584 49.7 90.8 64.3 5.9
Vitrociset Detection Algorithm
FRCNN
9. using public detections
0.72
±0.13
68.5
±0.0
78.0 1.7 88,601 10,081 25,963 77.3 89.8 83.1 5.1
S. Ren, K. He, R. Girshick, J. Sun. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. In NIPS, 2015.
ISE_Detv2
10. using public detections
0.88
±0.05
67.4
±12.8
79.9 4.9 106,094 28,865 8,470 92.6 78.6 85.0 3.2
MIFD
TrackerAPMODAMODPFAFTPFPFNRecallPrecisionF1 Hz
MHD
11. using public detections
0.49
±0.20
11.2
±42.9
69.9 8.8 64,637 51,801 49,927 56.4 55.5 56.0 3.0
Mobilenet-based Human Detection
ZIZOM
12. using public detections
0.81
±0.05
72.0
±22.0
79.8 2.2 95,414 12,990 19,139 83.3 88.0 85.6 2.4
C. Lin, L. Jiwen, G. Wang, J. Zhou. Graininess-Aware Deep Feature Learning for Pedestrian Detection. In ECCV, 2018.
KDNT
13. using public detections
0.89
±0.07
67.1
±22.4
80.1 4.8 105,473 28,623 9,091 92.1 78.7 84.8 0.8
F. Yu, W. Li, Q. Li, Y. Liu, X. Shi, J. Yan. POI: Multiple Object Tracking with High Performance Detection and Appearance Feature. In BMTT, SenseTime Group Limited, 2016.
SDP
14. using public detections
0.81
±0.12
76.9
±16.2
78.0 1.3 95,699 7,599 18,865 83.5 92.6 87.9 0.6
F. Yang, W. Choi, Y. Lin. Exploit All the Layers: Fast and Accurate CNN Object Detector With Scale Dependent Pooling and Cascaded Rejection Classifiers. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
HDGP
15. using public detections
0.45
±0.20
42.1
±20.1
76.4 1.3 55,680 7,436 58,884 48.6 88.2 62.7 0.6
A. Garcia-Martin, R. Sanchez-Matilla, J. Martinez. Hierarchical detection of persons in groups. In Signal, Image and Video Processing, 2017.
SequencesFramesTrajectoriesBoxes
75919785188076

Difficulty Analysis

Sequence difficulty (from easiest to hardest, measured by average AP)

MOT17-08

MOT17-08

(0.82 AP)

MOT17-03

MOT17-03

(0.78 AP)

MOT17-06

MOT17-06

(0.72 AP)

...

...

MOT17-01

MOT17-01

(0.62 AP)

MOT17-14

MOT17-14

(0.51 AP)


Evaluation Measures

Lower is better. Higher is better.
Measure Better Perfect Description
MOTA higher 100 % Multiple Object Tracking Accuracy [1]. This measure combines three error sources: false positives, missed targets and identity switches.
MOTP higher 100 % Multiple Object Tracking Precision [1]. The misalignment between the annotated and the predicted bounding boxes.
IDF1 higher 100 % ID F1 Score [2]. The ratio of correctly identified detections over the average number of ground-truth and computed detections.
FAF lower 0 The average number of false alarms per frame.
MT higher 100 % Mostly tracked targets. The ratio of ground-truth trajectories that are covered by a track hypothesis for at least 80% of their respective life span.
ML lower 0 % Mostly lost targets. The ratio of ground-truth trajectories that are covered by a track hypothesis for at most 20% of their respective life span.
FP lower 0 The total number of false positives.
FN lower 0 The total number of false negatives (missed targets).
ID Sw. lower 0 The total number of identity switches. Please note that we follow the stricter definition of identity switches as described in [3].
Frag lower 0 The total number of times a trajectory is fragmented (i.e. interrupted during tracking).
Hz higher Inf. Processing speed (in frames per second excluding the detector) on the benchmark. The frequency is provided by the authors and not officially evaluated by the MOTChallenge.

Legend

Symbol Description
online method This is an online (causal) method, i.e. the solution is immediately available with each incoming frame and cannot be changed at any later time.
using public detections This method used the provided detection set as input.
using public detections This method used the provided detection set as input.
new This entry has been submitted or updated less than a week ago.
new This entry has been submitted or updated less than a week ago.

References:


[1] Bernardin, K. & Stiefelhagen, R. Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics. Image and Video Processing, 2008(1):1-10, 2008.
[2] Ristani, E., Solera, F., Zou, R., Cucchiara, R. & Tomasi, C. Performance Measures and a Data Set for Multi-Target, Multi-Camera Tracking. In ECCV workshop on Benchmarking Multi-Target Tracking, 2016.
[3] Li, Y., Huang, C. & Nevatia, R. Learning to associate: HybridBoosted multi-target tracker for crowded scene. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2009.