{"ID":2847251,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.00524","arxiv_id":"2511.00524","title":"Text-guided Fine-Grained Video Anomaly Understanding","abstract":"Subtle abnormal events in videos often manifest as weak spatio-temporal cues that are easily overlooked by conventional anomaly detection systems. Existing video anomaly detection approaches typically provide coarse binary anomaly decisions without interpretable evidence, while large vision-language models (LVLMs) can produce textual judgments but lack precise localization of subtle visual signals. To address this gap, we propose Text-guided Fine-Grained Video Anomaly Understanding T-VAU, a framework that grounds subtle anomaly evidence into multimodal reasoning. Specifically, we introduce an Anomaly Heatmap Decoder (AHD) that performs visual-textual feature alignment to extract pixel-level spatio-temporal anomaly heatmaps from intermediate visual representations. We further design a Region-aware Anomaly Encoder (RAE) that converts these heatmaps into structured prompt embeddings, enabling the LVLM to perform anomaly detection, localization, and semantic explanation in a unified reasoning pipeline. To support fine-grained supervision, we construct a target-level fine-grained video-text anomaly dataset derived from ShanghaiTech and UBnormal with detailed annotations of object appearance, localization, and motion trajectories. Extensive experiments demonstrate that T-VAU significantly improves anomaly localization and textual reasoning performance on both benchmarks, achieving strong results in BLEU-4 metrics and Yes/No decision accuracy while providing interpretable pixel-level spatio-temporal evidence for anomaly understanding. The code will be available at https://github.com/momiji-bit/T-VAU.","short_abstract":"Subtle abnormal events in videos often manifest as weak spatio-temporal cues that are easily overlooked by conventional anomaly detection systems. Existing video anomaly detection approaches typically provide coarse binary anomaly decisions without interpretable evidence, while large vision-language models (LVLMs) can...","url_abs":"https://arxiv.org/abs/2511.00524","url_pdf":"https://arxiv.org/pdf/2511.00524v3","authors":"[\"Jihao Gu\",\"Kun Li\",\"He Wang\",\"Kaan Akşit\"]","published":"2025-11-01T11:59:23Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Language Model\"]","has_code":false,"code_links":[{"ID":607503,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2847251,"paper_url":"https://arxiv.org/abs/2511.00524","paper_title":"Text-guided Fine-Grained Video Anomaly Understanding","repo_url":"https://github.com/momiji-bit/T-VAU","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}