Aug 2026· International Journal of Computer Vision· Vol 134· 0 citations· 105 references
TL;DR
This paper conducts an extensive analysis of the robustness of 12 leading TAD methods and proposes a recipe with six steps to build a strong TAD baseline, which will play a crucial role in shaping future research on robust video analysis.
This work proposes the first video-language-model post-training technique for mistake detection, which uses a tailored reward function to encourage the model to identify discrepancies between an instruction and the corresponding video, and generalizes especially well to unseen procedures.
Federico Spurio, Olga Zatsarynna, Lars Doorenbos et al.· 1 citation
Temporal action detection (TAD) should remain reliable when videos are collected by different sources, even if the action vocabulary is unchanged. We study this domain-generalization setting with pre-extracted temporal features and no target-domain videos during training. DRTAD models source-related feature factors, mi...
Ling-Wen Meng, Guang-Hui Xi, Jian-Gang Liu et al.· Electronics· 0 citations
STRAND is introduced, a benchmark of human-verified object-centric facts that evaluates intermediate reasoning by decomposing queries into sub-questions, distinguishing genuine temporal understanding from coincidental correctness and an object-centric framework that explicitly constructs and reasons over structured obj...
T. Nguyen, C. Trinh, Khoi M. Le et al.· 0 citations
This work investigates robustness to one-step perturbations applied at a single time step per episode and proposes CARE, which dynamically selects the execution length based on consistency with the previously predicted action chunk, which improves robustness with low computational overhead while preserving clean perfor...
A novel framework termed Motion-Causal Token Selection for cross-dataset action recognition that significantly outperforms existing state-of-the-art domain generalization approaches and enhances cross-dataset transferability but also reduces the computational footprint of vision transformers during inference.
Chloé Caron· International Journal of Com...· 0 citations
SV-WAM is proposed, a surround-view world-action model (WAM) that preserves full six-camera observations while maintaining efficient inference and introduces a differentiable drivable-area compliance regularizer that penalizes vehicle-footprint corners approaching or crossing drivable boundaries, improving planning saf...
Jin-Yang Wang, Shi-Wei Li, Jun-Jian Wang et al.· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.