Instance-Aligned Captions for Explainable Video Anomaly Detection

TL;DR

This paper introduces instance-aligned captions for explainable video anomaly detection, linking textual explanations to specific objects and actions for verifiability, and extends benchmark datasets for comprehensive evaluation.

Contribution

It proposes a novel framework for spatially grounded, instance-level explanations in VAD and creates VIEW360+ dataset for benchmarking explainability.

Findings

Current methods lack spatial grounding in explanations.

Instance-aligned captions improve explanation trustworthiness.

Benchmark results reveal limitations of existing LLM- and VLM-based methods.

Abstract

Explainable video anomaly detection (VAD) is crucial for safety-critical applications, yet even with recent progress, much of the research still lacks spatial grounding, making the explanations unverifiable. This limitation is especially pronounced in multi-entity interactions, where existing explainable VAD methods often produce incomplete or visually misaligned descriptions, reducing their trustworthiness. To address these challenges, we introduce instance-aligned captions that link each textual claim to specific object instances with appearance and motion attributes. Our framework captures who caused the anomaly, what each entity was doing, whom it affected, and where the explanationis grounded, enabling verifiable and actionable reasoning. We annotate eight widely used VAD benchmarks and extend the 360-degree egocentric dataset, VIEW360, with 868 additional videos, eight locations, and four new anomaly types, creating VIEW360+, a comprehensive testbed for explainable VAD. Experiments show that our instance-level spatially grounded captions reveal significant limitations in current LLM- and VLM-based methods while providing a robust benchmark for future research in trustworthy and interpretable anomaly detection.