Object detection and tracking using iterative division and correlograms

Abstract

This paper presents algorithms for detecting moving objects, tracking them and interpreting their relationships. The algorithms are based on color and texture analysis in HSV space for real time processing. Our goal is to study human interaction by tracking people and objects for surveillance applications. The object detection algorithm is based on color histograms and iteratively divided interest regions for motion detection. The tracking algorithm is based on correlograms which combine spectral and spatial information to match detected objects in consecutive frames.