Real-Time Vision-Based Moving Object Detection and Tracking

In applications involving surveillance cameras, action recognition, and automotive safety, moving object detection and tracking is commonly a major element. The primary goal of the project is to detect and track the faces of moving objects and people using a static integrated camera capable of detecting the facial attributes of the discovered faces. We describe a comprehensive system for detecting and tracking moving objects that would be centered on a vision system using an image difference algorithm and the feature tracker Kanade Lucas Tomasi. Our project focuses on identifying and locating a person's movements as long as they remain within the recording plan. To accomplish this, the image difference algorithm in MATLAB / Simulink software is being used, and the KLT algorithm is being used to extract features. Our project describes an algorithm to detect the facial features of multiple faces in a given frame. It also keeps the track of the face even when the person tilts his/her head, or moves towards or away from the camera.