A training scheme that learns network-flow tracker weights from small perturbations of ground-truth tracks, represented as generalized graph differences, achieves competitive DukeMTMCT MOTA without solver-in-the-loop training.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Multi Target Tracking by Learning from Generalized Graph Differences
A training scheme that learns network-flow tracker weights from small perturbations of ground-truth tracks, represented as generalized graph differences, achieves competitive DukeMTMCT MOTA without solver-in-the-loop training.