{"ID":2870754,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.11605","arxiv_id":"2509.11605","title":"DUAL-VAD: Dual Benchmarks and Anomaly-Focused Sampling for Video Anomaly Detection","abstract":"Video Anomaly Detection (VAD) is critical for surveillance and public safety. However, existing benchmarks are limited to either frame-level or video-level tasks, restricting a holistic view of model generalization. This work first introduces a softmax-based frame allocation strategy that prioritizes anomaly-dense segments while maintaining full-video coverage, enabling balanced sampling across temporal scales. Building on this process, we construct two complementary benchmarks. The image-based benchmark evaluates frame-level reasoning with representative frames, while the video-based benchmark extends to temporally localized segments and incorporates an abnormality scoring task. Experiments on UCF-Crime demonstrate improvements at both the frame and video levels, and ablation studies confirm clear advantages of anomaly-focused sampling over uniform and random baselines.","short_abstract":"Video Anomaly Detection (VAD) is critical for surveillance and public safety. However, existing benchmarks are limited to either frame-level or video-level tasks, restricting a holistic view of model generalization. This work first introduces a softmax-based frame allocation strategy that prioritizes anomaly-dense segm...","url_abs":"https://arxiv.org/abs/2509.11605","url_pdf":"https://arxiv.org/pdf/2509.11605v2","authors":"[\"Seoik Jung\",\"Taekyung Song\",\"Joshua Jordan Daniel\",\"JinYoung Lee\",\"SungJun Lee\"]","published":"2025-09-15T05:48:22Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false}