MOT Challenge

GNNMatch: Graph Convolutional Neural Networks for Multi-Object Tracking via Sinkhorn Normalization

MOT20-04

MOT20-04

MOT20-06

MOT20-07

MOT20-08

Benchmark:

MOT17 | MOT16 | MOT20 | MOT15 |

Short name:

GNNMatch

Detector:

Public

Description:

Reference:

I. Papakis, A. Sarkar, A. Karpatne. GCNNMatch: Graph Convolutional Neural Networks for Multi-Object Tracking via Sinkhorn Normalization. In arXiv preprint arXiv:2010.00067, 2020.

Last submitted:

February 23, 2021 (3 years ago)

Published:

February 23, 2021 at 15:37:20 CET

Submissions:

4

Project page / code:

https://github.com/IPapakis/GCNNMatch

Open source:

Yes

Hardware:

TITAN RTX

Runtime:

0.1 Hz

Benchmark performance:

Sequence	MOTA	IDF1	HOTA	MT	ML	FP	FN	Rcll	Prcn	AssA	DetA	AssRe	AssPr	DetRe	DetPr	LocA	FAF	ID Sw.	Frag
MOT20	54.5	49.0	40.2	407 (32.8)	317 (25.5)	9,522	223,611	56.8	96.9	37.0	43.8	44.0	58.7	45.9	78.4	82.0	2.1	2,038 (35.9)	2,456 (43.3)

Detailed performance:

Sequence	MOTA	IDF1	HOTA	MT	ML	FP	FN	Rcll	Prcn	AssA	DetA	AssRe	AssPr	DetRe	DetPr	LocA	FAF	ID Sw.	Frag
MOT20-04	74.7	61.9	49.5	331	54	3,720	64,662	76.4	98.3	41.6	59.1	49.0	63.4	62.1	79.9	82.4	1.8	830	1,041
MOT20-06	32.0	29.1	24.5	31	130	2,666	86,996	34.5	94.5	22.9	26.4	28.0	43.4	27.5	75.3	80.7	2.6	676	810
MOT20-07	52.3	47.5	38.5	31	20	460	15,165	54.2	97.5	36.0	41.7	42.9	59.5	43.8	78.9	82.7	0.8	159	185
MOT20-08	22.8	23.2	20.8	14	113	2,676	56,788	26.7	88.6	21.5	20.3	27.1	43.0	21.3	70.6	80.1	3.3	373	420

Raw data:

GNNMatch