MOT Challenge

Video not available.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Benchmark:

MOT17Det | MOT20Det |

Short name:

SGT_det

Description:

Reference:

J. Hyun, M. Kang, D. Wee, D. Yeung. Detection recovery in online multi-object tracking with sparse graph tracker. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2023.

Last submitted:

October 28, 2021 (3 years ago)

Published:

November 14, 2021 at 16:04:23 CET

Submissions:

Project page / code:

https://github.com/HYUNJS/SGT

Open source:

Yes

Hardware:

1 V100, 12 CPUs

Runtime:

14.3 Hz

Benchmark performance:

Sequence	AP	MODA	MODP	FAF	TP	FP	FN	Rcll	Prcn	F1
MOT20Det	0.90	84.3	80.6	5.0	311,920	22,183	31,604	90.8	93.4	92.1

Detailed performance:

Sequence	AP	MODA	MODP	FAF	TP	FP	FN	Rcll	Prcn	F1
MOT20-04	0.91	89.1	80.3	6.1	218,304	12,720	12,425	94.6	94.5	94.6
MOT20-06	0.72	72.2	81.2	4.2	50,282	4,185	13,607	78.7	92.3	85.0
MOT20-07	0.91	85.4	83.7	3.3	15,840	1,917	458	97.2	89.2	93.0
MOT20-08	0.81	74.0	80.1	4.2	27,494	3,361	5,114	84.3	89.1	86.6

Raw data:

download

SW_YoloX_det

SGT_det

ViPeD20

SGT_det: Sparse Graph Tracker for detection