Locality-aware Attention Network with Discriminative Dynamics Learning for Weakly Supervised Anomaly Detection

TL;DR

This paper introduces a novel locality-aware attention network combined with discriminative dynamics learning to improve weakly supervised video anomaly detection by modeling temporal variations and global correlations.

Contribution

It proposes a new Discriminative Dynamics Learning method with dynamics ranking and alignment losses, and a Locality-aware Attention Network for better anomaly localization.

Findings

Significant performance improvements on UCF-Crime and XD-Violence datasets.

Effective modeling of temporal dynamics enhances anomaly detection accuracy.

The method outperforms existing weakly supervised approaches.

Abstract

Video anomaly detection is recently formulated as a multiple instance learning task under weak supervision, in which each video is treated as a bag of snippets to be determined whether contains anomalies. Previous efforts mainly focus on the discrimination of the snippet itself without modeling the temporal dynamics, which refers to the variation of adjacent snippets. Therefore, we propose a Discriminative Dynamics Learning (DDL) method with two objective functions, i.e., dynamics ranking loss and dynamics alignment loss. The former aims to enlarge the score dynamics gap between positive and negative bags while the latter performs temporal alignment of the feature dynamics and score dynamics within the bag. Moreover, a Locality-aware Attention Network (LA-Net) is constructed to capture global correlations and re-calibrate the location preference across snippets, followed by a multilayer perceptron with causal convolution to obtain anomaly scores. Experimental results show that our method achieves significant improvements on two challenging benchmarks, i.e., UCF-Crime and XD-Violence.