Facial Detection in Surveillance Systems

Abstract

Object Detection is an application of Deep Learning, and Computer Vision to deal with the task of spotting or detecting objects belonging to a certain class in an image or video source. Face Detection is a subset of Object Detection, wherein the task is to find human faces from a given input source, and segregate them for further processing. Face Detection is the first step in Facial recognition which finds its place in applications such as Bio-metrics for Security, Law enforcement, CCTV surveillance and many more. This project titled "Facial Detection in Surveillance Systems" is a module of the project Perimeter Security", which aims to detect whether a surveillance footage contains a human face or not, and verify the detected face with an existing database of authorised faces. If the comparison yields no result, it is agged as an intrusion by the Alert Module, and an alert is sent to the concerned authority. This report covers the research and implementation of the Face Detection module, and the Alert generation Module. It covers an introduction to the project, the current state-of-the-art research of algorithms that make facial detection possible, and an implementation of some of these algorithms. A comparison is made between two of the most accurate algorithms, in order to see which algorithm is best suited for the application. Finally, the report concludes with drawbacks in the current implementation, and future enhancements that can be made for the same.