DUAL-VAD: Dual Benchmarks and Anomaly-Focused Sampling for Video Anomaly Detection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915495878328320 |
|---|---|
| author | Jung, Seoik Song, Taekyung Daniel, Joshua Jordan Lee, JinYoung Lee, SungJun |
| author_facet | Jung, Seoik Song, Taekyung Daniel, Joshua Jordan Lee, JinYoung Lee, SungJun |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_11605 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DUAL-VAD: Dual Benchmarks and Anomaly-Focused Sampling for Video Anomaly Detection Jung, Seoik Song, Taekyung Daniel, Joshua Jordan Lee, JinYoung Lee, SungJun Computer Vision and Pattern Recognition 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. |
| title | DUAL-VAD: Dual Benchmarks and Anomaly-Focused Sampling for Video Anomaly Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.11605 |