An Improved Technique For Sclera Image Enhancement

Abstract

The present work is based on the study of contrast enhancement of the sclera images. In this paper, an improved technique for sclera contrast enhancement has been proposed using linear discriminant analysis, recursively separated and weighted histogram equalization with gamma correction (gamma value 0 & 1). The region of interest and shape detection has been applied on eye image and segmented the color using Gray level co-occurrence matrix. Parameters such as peak signal-to-noise ratio, mean square error and absolute mean brightness error are compared for proposed and existing methods to validate the proposed method. The proposed work accuracy of 94% as compared to existing work accuracy of 91% confirm that the efficiency of the proposed method.  The proposed precision value is 0.75 and recall value is 0.65 with high PSNR. Although an attempt has been madeto enhance the sclera images only.In future it can be further extended to extract the red veins and green veins of sclera image.