CLIP-AE uses CLIP features, audio, cross-attention fusion, and two self-supervised losses to improve unsupervised temporal action localization, with new top scores on THUMOS14 and ActivityNet v1.2.
Two-stream consensus network for weakly-supervised temporal action local- ization,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
CLIP-AE: CLIP-assisted Cross-view Audio-Visual Enhancement for Unsupervised Temporal Action Localization
CLIP-AE uses CLIP features, audio, cross-attention fusion, and two self-supervised losses to improve unsupervised temporal action localization, with new top scores on THUMOS14 and ActivityNet v1.2.