LATERN: Test-Time Context-Aware Explainable Video Anomaly Detection

TL;DR

LATERN is a novel framework that improves video anomaly detection by aggregating temporal evidence and providing explainable, event-level decisions using vision-language models.

Contribution

It introduces a context-aware approach with memory and recursive aggregation modules to enhance detection accuracy and explanation coherence in VAD.

Findings

LATERN outperforms existing methods on UCF-Crime and XD-Violence benchmarks.

It produces temporally coherent and semantically grounded explanations.

The framework improves detection accuracy with frozen vision-language models.

Abstract

Vision-language models (VLMs) have recently emerged as a promising paradigm for video anomaly detection (VAD) due to their strong visual reasoning ability and natural language-based explainability. In this paper, we aim to address a key limitation of such pipelines, which perform segment-level inference independently owing to token constraints and reason without structured temporal context, allowing VLMs to interpret anomalies as deviations from evolving video dynamics rather than producing fragmented predictions and explanations. To specify, we propose a context-aware framework named LATERN, which reformulates VAD as a temporal evidence aggregation process. LATERN consists of two complementary modules: Context-Aware Anomaly Scoring (CEA) and Recursive Evidence Aggregation (REA). CEA introduces a novel image-grounded memory mechanism, which selectively chooses historical content via frame diversity and visual-textual alignment as expanded context to help generate reliable anomaly scores. Building upon these scores, REA performs recursive temporal aggregation to identify coherent anomaly intervals and produce event-level decisions and explanations grounded in visual-textual evidence. Extensive experiments on challenging benchmarks, including UCF-Crime and XD-Violence, show that LATERN enhances detection accuracy and explanation consistency for frozen VLMs during test time, while generating temporally coherent and semantically grounded event-level explanations.