The paper proposes ECVT, a transformer that uses LVLM-generated text prompts to guide action localization, and reports state-of-the-art results that it itself describes as fabricated.
Visual features of interme- diate complexity and their use in classification,
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
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Multi-Level LVLM Guidance for Untrimmed Video Action Recognition
The paper proposes ECVT, a transformer that uses LVLM-generated text prompts to guide action localization, and reports state-of-the-art results that it itself describes as fabricated.