LPC_MOT: Learning a Proposal Classifier for Multiple Object Tracking

MOT20-08


Benchmark:

MOT20 | MOT17 | MOT16 |

Short name:

LPC_MOT

Detector:

Public

Description:

This approach is devised by Tsinghua University and AIBEE Inc.

Reference:

P. Dai, R. Weng, W. Choi, C. Zhang, Z. He, W. Ding. Learning a Proposal Classifier for Multiple Object tracking. In CVPR (Accepted), 2021.

Last submitted:

July 05, 2020 (1 year ago)

Published:

July 05, 2020 at 09:05:11 CET

Submissions:

2

Project page / code:

Open source:

Yes

Hardware:

GeForce GTX 1080 Ti

Runtime:

0.7 Hz

Benchmark performance:

Sequence MOTA IDF1 HOTA MT ML FP FN Rcll Prcn AssA DetA AssRe AssPr DetRe DetPr LocA FAF ID Sw. Frag
MOT2056.362.549.0424 (34.1)313 (25.2)11,726213,05658.896.352.445.854.781.348.178.782.32.61,562 (26.6)1,865 (31.7)

Detailed performance:

Sequence MOTA IDF1 HOTA MT ML FP FN Rcll Prcn AssA DetA AssRe AssPr DetRe DetPr LocA FAF ID Sw. Frag
MOT20-0475.775.759.2356634,18061,86477.498.157.860.860.382.663.880.883.02.0648670
MOT20-0635.343.233.3351213,50381,89138.393.637.729.539.478.030.875.280.63.5499653
MOT20-0750.858.944.5202922915,92151.998.748.940.651.281.142.180.182.30.4124138
MOT20-0825.837.429.4131003,81453,38031.186.337.123.439.174.724.868.979.34.7291404

Raw data: