MPNTrackSeg: Multi-Object Tracking and Segmentation via Neural Message Passing

MOTS20-01


Benchmark:

MOTS |

Short name:

MPNTrackSeg

Description:

Reference:

G. Bras'o, O. Cetintas, L. Leal-Taix'e. Multi-Object Tracking and Segmentation Via Neural Message Passing. In International Journal of Computer Vision, 2022.

Last submitted:

June 30, 2022 (5 months ago)

Published:

June 30, 2022 at 14:53:12 CET

Submissions:

1

Open source:

No

Hardware:

RTX 8000

Runtime:

2.3 Hz

Benchmark performance:

Sequence sMOTSA IDF1 MOTSA MOTSP MODSA MT ML TP FP FN Rcll Prcn ID Sw. Frag
MOTS58.668.873.780.674.3207 (63.1)26 (7.9)25,0361,0597,23377.695.9202 (260.4)635 (818.5)

Detailed performance:

Sequence sMOTSA IDF1 MOTSA MOTSP MODSA MT ML TP FP FN Rcll Prcn ID Sw. Frag
MOTS20-0160.675.979.377.679.7802,59111651583.495.71264
MOTS20-0670.169.886.581.987.414158,92634488891.096.391241
MOTS20-0746.560.460.377.760.714177,9661444,91261.998.259235
MOTS20-1264.678.078.284.278.84445,55345591885.892.44095

Raw data: