Applying multi-object tracking to fuse softmax probabilities across frames in camera trap data yields weighted F1-score gains of 5.1%, 3.1%, and 2.0% over standalone classifiers on three datasets.
Boost- Track: boosting the similarity measure and detection con- fidence for improved multiple object tracking.Machine Vision and Applications, 35(3), 2024
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Multi-Object Tracking Consistently Improves Wildlife Inference
Applying multi-object tracking to fuse softmax probabilities across frames in camera trap data yields weighted F1-score gains of 5.1%, 3.1%, and 2.0% over standalone classifiers on three datasets.