A motion-aware feature learned by a temporal network plus attention in MIL ranking yields large gains over prior methods on anomaly detection and action recognition in the UCF Crime dataset.
Temporal Segment Networks: Towards Good Practices for Deep Action Recognition
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Motion-Aware Feature for Improved Video Anomaly Detection
A motion-aware feature learned by a temporal network plus attention in MIL ranking yields large gains over prior methods on anomaly detection and action recognition in the UCF Crime dataset.