Neural Network based Unsupervised Face and Mask Detection in Surveillance Networks

Abstract

In the post-pandemic world, surveillance cameras play a key aspect when it comes to detecting various kinds of security risks. These can range from burglars entering a premises to an individual wearing or not wearing a mask where convention dictates one way versus the other. We are proposing a system that would allow autonomously detecting these security risks with minimal human intervention. We propose using Multi-task Cascaded Convolutional Neural Networks as the face detector, a choice of a complete range of classic image feature extractors, and the Kernel-based Online Anomaly Detection algorithm to identify the potential risk in real time. We tested our proposed framework on three different datasets including real-world settings. Our proposed framework yielded high detection rates with low false alarm rates, in addition to being adaptive, portable, and requiring minimal infrastructure.