Video Anomaly Detection with Semantics-Aware Information Bottleneck

TL;DR

This paper introduces SIB-VAD, a semi-supervised video anomaly detection framework that uses adaptive information bottleneck filtering and semantic-aware multimodal modeling to improve detection accuracy and scene adaptability.

Contribution

It proposes a novel feature filtering method replacing memory modules with an adaptive information bottleneck approach and enhances semantic understanding through multimodal prediction.

Findings

Outperforms existing methods on benchmark datasets.

Effectively filters abnormal features using the information bottleneck.

Improves detection robustness with multimodal semantic modeling.

Abstract

Semi-supervised video anomaly detection methods face two critical challenges: (1) Strong generalization blurs the boundary between normal and abnormal patterns. Although existing approaches attempt to alleviate this issue using memory modules, their rigid prototype-matching process limits adaptability to diverse scenarios; (2) Relying solely on low-level appearance and motion cues makes it difficult to perceive high-level semantic anomalies in complex scenes. To address these limitations, we propose SIB-VAD, a novel framework based on adaptive information bottleneck filtering and semantic-aware enhancement. We propose the Sparse Feature Filtering Module (SFFM) to replace traditional memory modules. It compresses normal features directly into a low-dimensional manifold based on the information bottleneck principle and uses an adaptive routing mechanism to dynamically select the most suitable normal bottleneck subspace. Trained only on normal data, SFFMs only learn normal low-dimensional manifolds, while abnormal features deviate and are effectively filtered. Unlike memory modules, SFFM directly removes abnormal information and adaptively handles scene variations. To improve semantic awareness, we further design a multimodal prediction framework that jointly models appearance, motion, and semantics. Through multimodal consistency constraints and joint error computation, it achieves more robust VAD performance. Experimental results validate the effectiveness of our feature filtering paradigm based on semantics-aware information bottleneck. Project page at https://qzfm.github.io/sib_vad_project_page/