DA-Flow: Dual Attention Normalizing Flow for Skeleton-based Video Anomaly Detection

TL;DR

DA-Flow introduces a lightweight dual attention normalizing flow model that effectively captures cross-dimension relationships in skeleton-based video anomaly detection, achieving state-of-the-art results with minimal parameters.

Contribution

The paper proposes the Dual Attention Module (DAM) and integrates it into a normalizing flow framework for improved skeleton-based video anomaly detection.

Findings

DA-Flow outperforms existing SOTA methods in micro AUC.

The model is robust against noise and negative samples.

Random projection without training can still detect anomalies effectively.

Abstract

Cooperation between temporal convolutional networks (TCN) and graph convolutional networks (GCN) as a processing module has shown promising results in skeleton-based video anomaly detection (SVAD). However, to maintain a lightweight model with low computational and storage complexity, shallow GCN and TCN blocks are constrained by small receptive fields and a lack of cross-dimension interaction capture. To tackle this limitation, we propose a lightweight module called the Dual Attention Module (DAM) for capturing cross-dimension interaction relationships in spatio-temporal skeletal data. It employs the frame attention mechanism to identify the most significant frames and the skeleton attention mechanism to capture broader relationships across fixed partitions with minimal parameters and flops. Furthermore, the proposed Dual Attention Normalizing Flow (DA-Flow) integrates the DAM as a post-processing unit after GCN within the normalizing flow framework. Simulations show that the proposed model is robust against noise and negative samples. Experimental results show that DA-Flow reaches competitive or better performance than the existing state-of-the-art (SOTA) methods in terms of the micro AUC metric with the fewest number of parameters. Moreover, we found that even without training, simply using random projection without dimensionality reduction on skeleton data enables substantial anomaly detection capabilities.