Hybrid Deep Network for Anomaly Detection

TL;DR

This paper introduces a novel deep CNN model for unsupervised anomaly detection in surveillance videos, leveraging patch position as a classification target to improve spatial localization of anomalies.

Contribution

It is the first to adapt patch position as a classification target in CNNs for anomaly detection, enhancing spatial localization of normal and abnormal events.

Findings

Competitive results on 4 benchmark datasets

Effective spatial localization of anomalies

First to use patch position as classification target

Abstract

In this paper, we propose a deep convolutional neural network (CNN) for anomaly detection in surveillance videos. The model is adapted from a typical auto-encoder working on video patches under the perspective of sparse combination learning. Our CNN focuses on (unsupervisedly) learning common characteristics of normal events with the emphasis of their spatial locations (by supervised losses). To our knowledge, this is the first work that directly adapts the patch position as the target of a classification sub-network. The model is capable to provide a score of anomaly assessment for each video frame. Our experiments were performed on 4 benchmark datasets with various anomalous events and the obtained results were competitive with state-of-the-art studies.