MOT Challenge

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

MOT17-14-SDP

MOT17-01-DPM

MOT17-03-DPM

MOT17-06-DPM

MOT17-07-DPM

MOT17-08-DPM

MOT17-12-DPM

MOT17-14-DPM

MOT17-01-FRCNN

MOT17-03-FRCNN

MOT17-06-FRCNN

MOT17-07-FRCNN

MOT17-08-FRCNN

MOT17-12-FRCNN

MOT17-14-FRCNN

MOT17-01-SDP

MOT17-03-SDP

MOT17-06-SDP

MOT17-07-SDP

MOT17-08-SDP

MOT17-12-SDP

MOT17-14-SDP

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:

March 02, 2021 (3 years ago)

Published:

July 15, 2020 at 14:27:00 CET

Submissions:

4

Project page / code:

https://github.com/IPapakis/GCNNMatch

Open source:

Yes

Hardware:

TITAN RTX

Runtime:

1.3 Hz

Benchmark performance:

Sequence	MOTA	IDF1	HOTA	MT	ML	FP	FN	Rcll	Prcn	AssA	DetA	AssRe	AssPr	DetRe	DetPr	LocA	FAF	ID Sw.	Frag
MOT17	57.3	56.3	45.4	575 (24.4)	787 (33.4)	14,100	225,042	60.1	96.0	45.2	45.9	52.5	67.9	48.4	77.4	81.5	0.8	1,911 (31.8)	2,837 (47.2)

Detailed performance:

Sequence	MOTA	IDF1	HOTA	MT	ML	FP	FN	Rcll	Prcn	AssA	DetA	AssRe	AssPr	DetRe	DetPr	LocA	FAF	ID Sw.	Frag
MOT17-01-DPM	43.6	41.4	35.8	5	12	90	3,529	45.3	97.0	37.2	34.5	44.2	64.6	35.9	76.9	80.6	0.2	19	28
MOT17-01-FRCNN	46.2	43.0	37.3	7	11	185	3,268	49.3	94.5	38.1	36.6	45.4	60.2	38.7	74.2	80.2	0.4	19	28
MOT17-01-SDP	46.6	46.0	37.6	8	10	165	3,258	49.5	95.1	38.3	37.1	45.9	62.6	39.0	74.9	80.1	0.4	22	34
MOT17-03-DPM	68.6	63.8	50.7	62	19	1,902	30,803	70.6	97.5	48.0	53.8	53.4	73.0	56.8	78.4	81.7	1.3	118	203
MOT17-03-FRCNN	69.5	66.6	52.0	60	17	1,847	29,976	71.4	97.6	50.1	54.4	55.2	73.4	57.3	78.4	81.5	1.2	112	222
MOT17-03-SDP	74.2	69.9	54.7	80	15	3,113	23,719	77.3	96.3	51.8	58.0	57.7	71.7	61.9	77.0	80.8	2.1	142	304
MOT17-06-DPM	54.2	33.4	31.8	61	82	513	4,774	59.5	93.2	22.3	45.7	55.0	29.9	48.9	76.7	82.4	0.4	107	128
MOT17-06-FRCNN	56.4	36.3	34.1	71	60	814	4,183	64.5	90.3	24.2	48.2	54.7	30.5	52.9	74.1	81.9	0.7	145	177
MOT17-06-SDP	56.2	34.6	32.2	74	63	841	4,162	64.7	90.1	21.7	48.2	53.9	27.7	53.0	73.8	82.0	0.7	159	175
MOT17-07-DPM	45.6	45.8	36.4	8	20	223	8,900	47.3	97.3	36.8	36.3	41.8	61.6	37.8	77.8	81.5	0.4	73	138
MOT17-07-FRCNN	44.9	47.7	37.1	8	17	396	8,835	47.7	95.3	38.3	36.4	41.9	68.1	38.2	76.3	81.1	0.8	83	145
MOT17-07-SDP	46.8	47.2	37.4	10	15	439	8,454	50.0	95.1	37.2	38.0	42.0	61.7	40.1	76.2	81.0	0.9	95	170
MOT17-08-DPM	28.4	32.0	29.4	9	37	190	14,876	29.6	97.0	35.4	24.5	40.5	70.2	25.1	82.4	84.6	0.3	50	67
MOT17-08-FRCNN	27.8	32.5	30.2	8	40	210	14,987	29.1	96.7	38.1	24.1	42.4	74.6	24.7	82.1	84.7	0.3	49	65
MOT17-08-SDP	29.3	33.0	30.8	13	35	232	14,654	30.6	96.5	38.0	25.2	44.0	69.9	25.9	81.5	84.2	0.4	57	85
MOT17-12-DPM	46.6	54.0	44.8	17	43	108	4,496	48.1	97.5	52.0	38.8	58.7	72.1	40.3	81.6	84.0	0.1	20	30
MOT17-12-FRCNN	45.0	51.8	43.7	15	43	121	4,633	46.5	97.1	50.7	37.8	59.0	69.9	39.2	81.7	84.2	0.1	16	25
MOT17-12-SDP	46.3	54.4	45.1	17	43	218	4,423	49.0	95.1	52.5	38.9	59.6	71.2	40.9	79.5	83.6	0.2	15	28
MOT17-14-DPM	33.6	38.8	29.6	12	71	554	11,546	37.5	92.6	31.4	28.1	36.9	59.1	29.5	72.9	79.7	0.7	175	222
MOT17-14-FRCNN	34.1	39.6	29.9	15	71	938	11,037	40.3	88.8	30.7	29.4	37.3	55.0	31.5	69.5	78.9	1.3	201	269
MOT17-14-SDP	36.4	41.4	31.1	15	63	1,001	10,529	43.0	88.8	31.3	31.2	37.0	57.2	33.6	69.4	79.0	1.3	234	294

Raw data:

GNNMatch