MOT Challenge

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

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.

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

Benchmark:

MOT20 | MOT17 |

Short name:

BBT

Detector:

Public

Description:

Tracker uses only bounding boxes, no visual information.
Bounding boxes are associated frame-to-frame maximizing the IOU metric.
Re-identification phase projects discovered tracks to past/future tracks within a time window of 50 frames and quantifies the track matching score.
Fragmented tracks with high matching score are linked together, missing bounding boxes interpolated.
The proposed tracker is extremely fast compared to visual-based methods, and more accurate than (similar-to-our-work method) IOU17 thanks to the re-identification phase,
The proposed tracker is used to discover inter-person relationships in videos.

Reference:

MTAP-D-20-01870 Tracking Subjects and Detecting Relationships in Crowded City Videos (under review)

Last submitted:

May 19, 2020 (3 years ago)

Published:

October 15, 2020 at 16:00:36 CET

Submissions:

Project page / code:

https://github.com/elicz/Multi-Subject-Tracking/

Open source:

Yes

Hardware:

Intel(R) Xeon(R) CPU E5-2620, 1 core

Runtime:

121.6 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	47.8	50.0	40.0	449 (19.1)	936 (39.7)	35,552	255,118	54.8	89.7	40.3	40.0	45.3	66.8	43.3	71.0	79.2	2.0	3,676 (67.1)	4,949 (90.3)

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	22.9	31.4	25.1	3	14	189	4,769	26.1	89.9	33.8	18.7	38.1	66.2	19.5	67.3	76.8	0.4	16	28
MOT17-01-FRCNN	26.3	45.2	39.5	9	7	1,744	2,990	53.6	66.5	45.7	34.3	49.5	71.0	43.7	54.2	78.6	3.9	19	48
MOT17-01-SDP	41.5	49.3	38.8	10	4	1,600	2,113	67.2	73.1	37.8	40.0	41.4	67.2	50.9	55.4	77.0	3.6	58	88
MOT17-03-DPM	44.3	40.2	31.6	22	37	5,166	52,733	49.6	91.0	27.9	36.0	29.6	71.2	38.6	70.7	78.7	3.4	355	561
MOT17-03-FRCNN	61.1	60.5	47.3	47	24	3,350	37,221	64.4	95.3	46.6	48.5	52.8	66.6	51.4	76.0	80.1	2.2	130	333
MOT17-03-SDP	77.7	71.1	54.8	90	12	3,819	19,310	81.6	95.7	50.2	60.0	56.1	67.5	64.4	75.5	80.0	2.5	213	637
MOT17-06-DPM	35.6	41.4	31.5	24	124	825	6,612	43.9	86.2	32.3	31.0	40.0	58.9	33.7	66.2	77.4	0.7	151	123
MOT17-06-FRCNN	49.0	51.4	40.8	49	82	800	5,039	57.2	89.4	38.6	43.5	46.5	63.1	47.1	73.6	81.0	0.7	175	189
MOT17-06-SDP	51.3	52.0	40.6	63	88	786	4,773	59.5	89.9	38.3	43.3	45.5	63.8	47.2	71.4	79.8	0.7	184	164
MOT17-07-DPM	30.0	31.4	24.9	4	31	830	10,900	35.5	87.8	24.2	25.7	25.6	72.0	27.3	67.5	77.2	1.7	99	146
MOT17-07-FRCNN	33.2	36.9	30.2	5	14	1,714	9,349	44.7	81.5	29.5	32.0	32.4	66.7	35.4	64.5	77.2	3.4	223	399
MOT17-07-SDP	47.4	41.9	34.7	14	16	1,059	7,672	54.6	89.7	30.9	39.2	32.8	70.9	42.6	70.0	78.6	2.1	147	267
MOT17-08-DPM	21.5	28.4	25.6	8	41	1,232	15,276	27.7	82.6	30.4	21.6	34.6	63.8	23.0	68.6	80.1	2.0	84	96
MOT17-08-FRCNN	25.5	36.9	32.9	13	34	1,823	13,837	34.5	80.0	42.1	25.9	49.9	62.1	28.4	65.8	79.7	2.9	68	109
MOT17-08-SDP	30.7	39.6	33.9	21	27	2,874	11,597	45.1	76.8	36.4	32.2	45.5	53.3	36.6	62.4	78.1	4.6	166	220
MOT17-12-DPM	34.7	42.6	34.4	10	52	567	5,060	41.6	86.4	38.6	30.8	41.4	70.9	33.3	69.1	79.9	0.6	35	43
MOT17-12-FRCNN	37.1	50.6	39.8	13	43	763	4,656	46.3	84.0	46.2	34.4	52.6	70.1	37.9	68.9	80.5	0.8	29	53
MOT17-12-SDP	40.6	50.7	40.7	20	40	952	4,163	52.0	82.6	44.3	37.4	52.0	67.2	42.3	67.2	81.0	1.1	32	52
MOT17-14-DPM	15.4	20.4	16.7	2	120	368	15,092	18.3	90.2	20.2	13.9	21.3	62.9	14.3	70.3	78.2	0.5	185	154
MOT17-14-FRCNN	16.5	29.7	24.3	10	70	2,961	11,752	36.4	69.4	23.8	25.7	26.5	54.2	29.2	55.7	74.2	3.9	722	675
MOT17-14-SDP	30.1	36.1	27.5	12	56	2,130	10,204	44.8	79.5	25.4	30.3	28.2	58.7	34.1	60.5	75.9	2.8	585	564

Raw data:

download

RTAT

BBT

AFN17

BBT: Bounding box based tracking enhanced by track-projection re-identification