MOT Challenge

MOT17-14-DPM

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 |

Short name:

UNS20regress

Detector:

Public

Description:

Visual-Spatial input-hiding scheme, plus detection regression/classification pre-processing.

Reference:

F. Bastani, S. He, S. Madden. Self-Supervised Multi-Object Tracking with Cross-input Consistency. In Advances in Neural Information Processing Systems (NeurIPS), 2021.

Last submitted:

January 20, 2021 (3 years ago)

Published:

January 20, 2021 at 17:36:18 CET

Submissions:

Project page / code:

https://favyen.com/uns20/

Open source:

Yes

Hardware:

V100

Runtime:

72.8 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	56.8	58.3	46.4	538 (22.8)	880 (37.4)	11,567	230,645	59.1	96.6	47.9	45.3	51.0	79.0	47.6	77.8	81.4	0.7	1,320 (22.3)	2,061 (34.9)

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	49.9	41.2	5	12	56	3,570	44.7	98.1	49.1	34.7	51.4	80.5	35.7	78.5	81.0	0.1	9	13
MOT17-01-FRCNN	45.6	51.7	42.1	7	10	205	3,296	48.9	93.9	48.8	36.5	51.4	78.1	38.6	74.1	80.4	0.5	11	19
MOT17-01-SDP	46.6	51.8	41.9	8	10	173	3,261	49.4	94.9	47.3	37.2	50.6	77.7	39.1	75.1	80.3	0.4	12	19
MOT17-03-DPM	68.1	63.4	50.5	60	19	1,732	31,549	69.9	97.7	48.0	53.3	50.7	79.1	56.1	78.4	81.4	1.2	134	185
MOT17-03-FRCNN	68.6	65.3	51.1	59	18	1,942	30,808	70.6	97.4	49.0	53.6	51.8	79.3	56.5	78.0	81.2	1.3	123	200
MOT17-03-SDP	73.7	68.2	54.3	75	16	2,813	24,562	76.5	96.6	51.7	57.4	54.8	77.8	61.0	77.0	80.7	1.9	157	264
MOT17-06-DPM	54.3	61.6	48.5	51	94	399	4,933	58.1	94.5	52.5	45.1	59.7	75.2	47.7	77.5	81.9	0.3	54	100
MOT17-06-FRCNN	56.7	60.6	48.8	64	69	707	4,323	63.3	91.3	50.4	47.7	59.1	71.5	51.7	74.6	81.3	0.6	70	140
MOT17-06-SDP	56.9	63.5	50.1	68	71	752	4,254	63.9	90.9	52.7	47.9	60.5	73.6	52.2	74.2	81.4	0.6	72	133
MOT17-07-DPM	44.0	46.9	37.5	6	21	182	9,211	45.5	97.7	39.8	35.6	41.6	80.1	36.8	79.1	81.9	0.4	68	99
MOT17-07-FRCNN	44.7	46.4	37.4	7	19	281	8,982	46.8	96.6	38.9	36.3	40.6	80.7	37.8	77.9	81.6	0.6	72	102
MOT17-07-SDP	47.4	49.2	39.0	10	17	330	8,476	49.8	96.2	40.0	38.4	42.3	79.0	40.2	77.6	81.3	0.7	73	113
MOT17-08-DPM	27.4	31.7	29.5	8	39	96	15,187	28.1	98.4	37.2	23.5	39.1	84.1	24.0	83.9	85.3	0.2	54	62
MOT17-08-FRCNN	26.6	30.9	29.1	8	40	105	15,342	27.4	98.2	37.1	22.9	38.7	84.6	23.4	83.8	85.3	0.2	56	67
MOT17-08-SDP	28.0	32.6	30.1	8	37	177	14,962	29.2	97.2	37.6	24.3	39.4	84.2	24.8	82.8	84.8	0.3	60	76
MOT17-12-DPM	45.1	56.2	45.7	17	52	57	4,686	45.9	98.6	55.5	37.6	58.6	83.0	38.8	83.3	84.8	0.1	18	21
MOT17-12-FRCNN	44.1	54.2	43.8	15	49	36	4,796	44.7	99.1	52.5	36.6	55.4	84.2	37.7	83.6	84.9	0.0	15	17
MOT17-12-SDP	45.2	57.2	46.5	18	49	170	4,563	47.4	96.0	56.5	38.3	59.7	82.5	40.0	81.1	84.4	0.2	19	22
MOT17-14-DPM	34.0	44.2	31.6	12	83	238	11,891	35.7	96.5	36.9	27.3	38.9	76.2	28.2	76.3	80.3	0.3	72	110
MOT17-14-FRCNN	36.1	45.8	33.1	17	79	521	11,203	39.4	93.3	37.2	29.6	39.7	74.0	31.0	73.4	79.5	0.7	81	146
MOT17-14-SDP	37.9	47.5	34.0	15	76	595	10,790	41.6	92.8	37.6	31.0	40.1	73.9	32.7	72.9	79.6	0.8	90	153

Raw data:

download

RTAT

UNS20regress

AFN17

UNS20regress: Self-Supervised Multi-Object Tracking with Cross-input Consistency (NeurIPS 2021)