TEGU improves zero-shot temporal action localization by using rich textual information from LLMs and video captions to better distinguish fine-grained actions without any training on labeled data.
Title resolution pending
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
-
Zero-Shot Temporal Action Localization Through Textual Guidance
TEGU improves zero-shot temporal action localization by using rich textual information from LLMs and video captions to better distinguish fine-grained actions without any training on labeled data.