A novel approach to automatic object detection using intensity pattern recognition in YCbCr Color space

Abstract

Object detection, is an important element of various computer vision areas, such as image retrieval, shot detection, video surveillance, etc. which require automatic segmentation or location of regions of interest in images which can be further analyzed by more computationally demanding techniques to produce a correct interpretation. This paper explores the possibility of performing this task of object detection in any general environment using the intensity pattern information in individual channels of YCbCr Color space. It was observed that prominent objects could be easily and efficiently detected by the algorithm irrespective of the environment they are kept in. The efficiency of the algorithm boasts the possibilities of further simplistic models for the invigorating task as of image detection for complicated computer vision applications.