Human-Scene Network: A Novel Baseline with Self-rectifying Loss for Weakly supervised Video Anomaly Detection

TL;DR

This paper introduces a Human-Scene Network with a self-rectifying loss for weakly-supervised video anomaly detection, effectively capturing subtle cues and improving performance on multiple datasets.

Contribution

It proposes a novel Human-Scene Network architecture combined with a self-rectifying loss to enhance weakly-supervised anomaly detection in videos.

Findings

Outperforms state-of-the-art on five of six scenarios

Effective in capturing both subtle and strong cues

Validated on three public datasets

Abstract

Video anomaly detection in surveillance systems with only video-level labels (i.e. weakly-supervised) is challenging. This is due to, (i) the complex integration of human and scene based anomalies comprising of subtle and sharp spatio-temporal cues in real-world scenarios, (ii) non-optimal optimization between normal and anomaly instances under weak supervision. In this paper, we propose a Human-Scene Network to learn discriminative representations by capturing both subtle and strong cues in a dissociative manner. In addition, a self-rectifying loss is also proposed that dynamically computes the pseudo temporal annotations from video-level labels for optimizing the Human-Scene Network effectively. The proposed Human-Scene Network optimized with self-rectifying loss is validated on three publicly available datasets i.e. UCF-Crime, ShanghaiTech and IITB-Corridor, outperforming recently reported state-of-the-art approaches on five out of the six scenarios considered.