Human eye detection of color images based on morphological segmentation using modified Harris corner detector

Abstract

This paper presents a novel method for human eye detection of color images based on modified Harris corner detection algorithm. The face detection from human face images is first performed using Cb component of YCbCr color space to detect eyes. Secondly, Harris corner detector is applied on a morphologically pre-processed extracted human face to detect eyes using only two arbitrary corner points (one corner point for one eye). This study is found to yield 88% eye detection rate which is comparable with the same obtained by other standard algorithm.