MOT Challenge

GCRA: Generation Cleaving and Re-connection Association

MOT16-01

MOT16-01

MOT16-03

MOT16-06

MOT16-07

MOT16-08

MOT16-12

MOT16-14

Benchmark:

MOT16 |

Short name:

GCRA

Detector:

Public

Description:

Tracklet Cleaving and Re-connection by Siamese-Bi GRU

Reference:

C. Ma, C. Yang, F. Yang, Y. Zhuang, Z. Zhang, H. Jia, X. Xie. Trajectory Factory: Tracklet Cleaving and Re-connection by Deep Siamese Bi-GRU for Multiple Object Tracking. In ICME, 2018.

Last submitted:

December 14, 2017 (3 years ago)

Published:

December 14, 2017 at 17:48:52 CET

Submissions:

1

Project page / code:

n/a

Open source:

No

Hardware:

NVIDIA TESLA K40

Runtime:

2.8 Hz

Benchmark performance:

Sequence	MOTA	IDF1	HOTA	MT	ML	FP	FN	Rcll	Prcn	AssA	DetA	AssRe	AssPr	DetRe	DetPr	LocA	FAF	ID Sw.	Frag
MOT16	48.2	48.6	37.6	98 (12.9)	312 (41.1)	5,104	88,586	51.4	94.8	36.7	38.9	39.4	74.3	41.0	75.6	80.6	0.9	821 (16.0)	1,117 (21.7)

Detailed performance:

Sequence	MOTA	IDF1	HOTA	MT	ML	FP	FN	Rcll	Prcn	AssA	DetA	AssRe	AssPr	DetRe	DetPr	LocA	FAF	ID Sw.	Frag
MOT16-01	42.5	55.5	39.8	6	10	31	3,633	43.2	98.9	49.7	31.8	51.6	77.4	32.7	74.9	78.2	0.1	11	40
MOT16-03	56.7	51.0	40.0	38	25	2,588	42,348	59.5	96.0	35.8	45.0	38.3	73.4	47.5	76.6	81.0	1.7	301	310
MOT16-06	35.9	44.4	33.2	20	111	823	6,451	44.1	86.1	35.3	31.8	44.6	61.4	34.3	67.0	76.8	0.7	127	209
MOT16-07	44.1	45.9	33.0	6	13	468	8,564	47.5	94.3	31.1	35.3	32.9	76.6	37.0	73.5	78.8	0.9	94	160
MOT16-08	35.4	47.1	37.4	7	22	224	10,505	37.2	96.5	46.3	30.3	48.1	83.3	31.3	81.1	84.1	0.4	75	101
MOT16-12	39.6	48.8	39.1	13	39	528	4,443	46.4	87.9	43.7	35.2	46.4	77.8	38.0	72.0	81.1	0.6	41	51
MOT16-14	28.3	36.0	26.4	8	92	442	12,642	31.6	93.0	29.5	23.8	31.0	71.7	24.7	72.7	78.8	0.6	172	246

Raw data:

GCRA