UCF-101-24

UCF-101-24 is a video action detection benchmark that contains a subset of 24 action categories from the UCF-101 dataset, used to evaluate models on their ability to classify and localize actions in video sequences.

Papers using UCF-101-24 (4)

TubeLite: Lightweight Multi-Actor Spatio-Temporal Action Detection2026Action tube generation by person query matching for spatio-temporal action detection2025Few-Shot Transformation of Common Actions into Time and Space2021 · 5 citesModeling Spatio-Temporal Human Track Structure for Action Localization2018 · 4 cites