MOTS Results

Download all results for this benchmark

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


Benchmark Statistics

TrackersMOTSAIDF1MOTSAMOTSPMODSAMTMLTPFPFNRcllPrcnID Sw.FragHz
ReMOTS_
1. using undisclosed detections
70.4
±3.6
75.0 84.4 84.0 85.1 248 (75.6)12 (3.7)28,270 819 3,999 87.6 97.2 231 (263.7)579 (660.9)0.3
F. Yang, X. Chang, C. Dang, Z. Zheng, S. Sakti, S. Nakamura, Y. Wu. ReMOTS: Refining Multi-Object Tracking and Segmentation. In , 2020.
VFS_MOTS
2. online method
69.7
±4.1
68.2 83.5 84.2 84.6 253 (77.1)9 (2.7)28,087 800 4,182 87.0 97.2 342 (392.9)625 (718.1)7.6
COSTA_st
3.
69.5
±3.9
70.3 83.3 84.2 84.6 253 (77.1)9 (2.7)28,093 806 4,176 87.1 97.2 421 (483.6)717 (823.6)2.1
GMPHD_MAF
4. online method
69.4
±3.0
66.4 83.3 84.2 84.8 249 (75.9)11 (3.4)28,284 935 3,985 87.7 96.8 484 (552.2)711 (811.2)2.6
Y. Song, Y. Yoon, K. Yoon, M. Jeon, S. Lee, W. Pedrycz. Online Multi-Object Tracking and Segmentation with GMPHD Filter and Mask-based Affinity Fusion. In [Online]. Available: arXiv:2009.00100, 2021.
UniTrack
5.
68.9
±3.8
67.2 82.6 84.2 84.5 253 (77.1)9 (2.7)28,084 801 4,185 87.0 97.2 622 (714.7)825 (947.9)7.6
ZXPointTrack
6. online method using undisclosed detections
62.3
±4.9
42.9 76.8 82.3 78.4 186 (56.7)41 (12.5)26,276 963 5,993 81.4 96.5 541 (664.4)868 (1,066.0)4.6
TBD
CCP
7. online method
59.3
±7.6
58.1 75.5 80.3 77.0 204 (62.2)32 (9.8)26,610 1,759 5,659 82.5 93.8 484 (586.9)645 (782.2)10.5
TBD
SORTS_ReID
8. online method using public detections
55.8
±5.1
65.8 69.1 81.9 70.0 107 (32.6)52 (15.9)23,671 1,076 8,598 73.4 95.7 304 (414.4)576 (785.2)36.4
Not yet published
SORTS
9. online method using undisclosed detections
55.0
±5.0
57.3 68.3 81.9 70.0 107 (32.6)52 (15.9)23,671 1,076 8,598 73.4 95.7 552 (752.5)577 (786.6)54.5
M. Ahrnbom, M. Nilsson, H. Ard"o. Real-time and Online Segmentation Multi-target Tracking with Track Revival Re-identification.. In VISIGRAPP (5: VISAPP), 2021.
Struct_MOTS
10.
55.0
±15.1
64.0 64.2 85.9 65.2 165 (50.3)52 (15.9)21,215 166 11,054 65.7 99.2 330 (501.9)609 (926.3)8.7
TrackersMOTSAIDF1MOTSAMOTSPMODSAMTMLTPFPFNRcllPrcnID Sw.FragHz
UBVision
11. using undisclosed detections
52.8
±9.8
58.3 67.4 79.9 69.2 162 (49.4)54 (16.5)23,465 1,142 8,804 72.7 95.4 568 (781.1)702 (965.4)10.1
TrackRCNN
12. using public detections
40.6
±13.8
42.4 55.2 76.1 56.9 127 (38.7)71 (21.6)19,628 1,261 12,641 60.8 94.0 567 (932.2)868 (1,427.0)2.0
P. Voigtlaender, M. Krause, A. O\usep, J. Luiten, B. Sekar, A. Geiger, B. Leibe. MOTS: Multi-Object Tracking and Segmentation. In CVPR, 2019.
SequencesFramesTrajectoriesBoxes
4304432832269

Difficulty Analysis

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

MOTS20-06

MOTS20-06

(68.5 sMOTSA)

MOTS20-12

MOTS20-12

(65.7 sMOTSA)

MOTS20-01

MOTS20-01

(61.4 sMOTSA)

MOTS20-07

MOTS20-07

(52.2 sMOTSA)

Evaluation Measures

Lower is better. Higher is better.
Measure Better Perfect Description
sMOTSA higher 100%Mask-based Soft Multi-Object Tracking Accuracy (+/- denotes standard deviation across all sequences) [1]. Soft version of MOTSA, accumulates the mask overlaps of true positives instead of only counting how many masks reach an IoU of more than 0.5.
IDF1 higher 100%ID F1 Score [2]. The ratio of correctly identified detections over the average number of ground-truth and computed detections.
MOTSA higher 100%Mask-based Multi-Object Tracking Accuracy (+/- denotes standard deviation across all sequences) [1]. Variant of MOTA, evaluated based on mask overlap (mask IoU)
MOTSP higher 100%Mask-overlap based variant of Multi-Object Tracking Precision. [1]. Variant of MOTP, evaluated based on mask IoU instead of bounding box IoU.
MODSA higher 100%Mask-overlap based Multi-Object Detection Accuracy [1]. Variant of MODA, evaluated based on the mask overlap (mask IoU).
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.
TP higher #GTThe total number of true positives.
FP lower 0The total number of false positives.
FN lower 0The total number of false negatives (missed targets).
Rcll higher 100%Ratio of correct detections to total number of GT boxes.
Prcn higher 100%Ratio of TP / (TP+FP).
ID Sw. lower 0Number of Identity Switches (ID switch ratio = #ID switches / recall) [3]. Please note that we follow the stricter definition of identity switches as described in the reference
Frag lower 0The 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 private detections This method used a private detection set as input.
new This entry has been submitted or updated less than a week ago.

References:


[1] Voigtlaender, P., Krause, M., Osep, A., Luiten, J., Sekar, B.B.G., Geiger, A. & Leibe, B. MOTS: Multi-Object Tracking and Segmentation. In Proceedings IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2019.
[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.



Contact