QVAD: A Question-Centric Agentic Framework for Efficient and Training-Free Video Anomaly Detection

TL;DR

QVAD introduces a dynamic, question-centric framework that enhances lightweight vision-language models for efficient, training-free video anomaly detection, achieving state-of-the-art results with minimal resources.

Contribution

It proposes a novel prompt-updating mechanism that refines queries iteratively, unlocking the capabilities of smaller models for high-performance VAD without extensive training.

Findings

State-of-the-art performance on UCF-Crime, XD-Violence, and UBNormal datasets.

High inference speed and low memory usage enable deployment on edge devices.

Effective generalization demonstrated on the ComplexVAD dataset.

Abstract

Video Anomaly Detection (VAD) is a fundamental challenge in computer vision, particularly due to the open-set nature of anomalies. While recent training-free approaches utilizing Vision-Language Models (VLMs) have shown promise, they typically rely on massive, resource-intensive foundation models to compensate for the ambiguity of static prompts. We argue that the bottleneck in VAD is not necessarily model capacity, but rather the static nature of inquiry. We propose QVAD, a question-centric agentic framework that treats VLM-LLM interaction as a dynamic dialogue. By iteratively refining queries based on visual context, our LLM agent guides smaller VLMs to produce high-fidelity captions and precise semantic reasoning without parameter updates. This ``prompt-updating" mechanism effectively unlocks the latent capabilities of lightweight models, enabling state-of-the-art performance on UCF-Crime, XD-Violence, and UBNormal using a fraction of the parameters required by competing methods. We further demonstrate exceptional generalizability on the single-scene ComplexVAD dataset. Crucially, QVAD achieves high inference speeds with minimal memory footprints, making advanced VAD capabilities deployable on resource-constrained edge devices.