Video Image for Security System by Using Chaotic Oscillator for Segmentation

Abstract

There are wide variety of mechanisms available for performing the authenticationfunction. One of them is by using video film to monitor important places. In such systemsthe important task is recognizing human faces which need good segmentation and highfeature extraction. In traditional technique the segmentation can be performed by regiongrowing and shrinking, clustering method, and boundary detection which all take imagepixel by pixel and compare each with neighbors to get the similarity. Chaotic oscillatordepends on synchronization concept between points to determine the active points whichconstruct similar region determining the object. Then RGB color space is used to determineskin color pixel. Skin color model aids the process of separating the face from the scene dueto its skin color. After isolating the human face image which contains many features (eye,eyebrow, nose, etc...) they identify the person. Chaotic oscillator is used to extract theimportant features by removing any point that is not synchronized with its neighbors andgives new image which contains only the main face features and then compare this imagewith stored image in Database of authorized person and make decision on if (he/she) is anauthenticated image or not.