CEAVAD: Contrastive Event Adjudication for Training-Free Video Anomaly Detection

Forum topic · 小凯 · 2026-08-12

Summary

CEAVAD is a training-free video anomaly detection (VAD) method proposed by Wenti Yin, Xiang Wang, and Huaxin Zhang (arXiv:2508.05149). While supervised VAD methods learn anomaly decision boundaries from target-domain annotations, they require substantial in-domain data. Existing training-free approaches rely on pretrained models' semantic knowledge and reasoning to describe visual content, but richer anomaly descriptions only capture hazard resemblance without actually defining abnormality. CEAVAD shifts the inference unit from isolated anomaly concepts to falsifiable event hypotheses. It first constructs hazard-normal event contrasts using public safety knowledge, pairing each hazardous mechanism with a generic normal explanation and a mechanism-specific benign counterpart. It then determines whether a target interval better supports the hazardous or benign competing explanation, producing a correctable contrastive boundary proposal. Finally, it adjudicates whether the hazard hypothesis withstands video evidence, enabling both temporally localized detection and evidence-based explanation. Experiments on three widely used VAD benchmarks show CEAVAD achieves state-of-the-art performance in the training-free paradigm.

Paper Overview

Field: Computer Vision Authors: Wenti Yin, Xiang Wang, Huaxin Zhang arXiv: 2508.05149

Abstract

Video anomaly detection (VAD) aims to identify and temporally localize abnormal events in videos. Supervised methods learn anomaly decision boundaries from target-domain annotations but require substantial in-domain data. Existing training-free methods leverage the rich semantic knowledge and reasoning capabilities of pretrained models to interpret visual content, yet these capabilities do not directly define an anomaly decision criterion: richer anomaly descriptions better capture hazard resemblance without resolving abnormality.

To this end, the authors propose Contrastive Event Adjudication for training-free Video Anomaly Detection (CEAVAD), which shifts the unit of inference from isolated anomaly concepts to falsifiable event hypotheses and establishes an inference-time explanatory boundary through interaction between competing explanations and video evidence.

Method

1. Contrast construction: Using public safety knowledge, CEAVAD builds hazard-normal event contrasts, pairing each hazardous mechanism with a generic normal explanation and a mechanism-specific benign counterpart. 2. Contrastive adjudication: It judges whether the target interval better supports the hazardous explanation or its benign competing explanation, generating a correctable contrastive boundary proposal for the target. 3. Evidence-based verdict: CEAVAD adjudicates between the competing explanations, determining whether the hazard hypothesis can withstand the video evidence, supporting both temporally localized anomaly detection and evidence-based explanation.

Results

Experiments on three widely used VAD benchmarks show that CEAVAD achieves state-of-the-art performance under the training-free paradigm.

---

*Auto-collected on 2026-08-12.*