On the use of Image Enhancement Technique towards Robust Sclera Segmentation

Abstract

This paper explores image enhancement techniques for robust sclera segmentation. The sclera segmentation is a primary step in sclera based human recognition. The performance of sclera recognition system depends on the accurate sclera segmentation. Unfortunately, the results of sclera segmentation deteriorate due to the acquisition of blurred or dark nature of input image. To improve the sclera segmentation results, this paper explores two well-known image enhancement techniques i.e., Histogram Equalization (HE) and Contrast Limited Adaptive Histogram Equalization (CLAHE). The efficacy of the image enhancement techniques are evaluated using segmentation results obtained by K-Means and Fuzzy C-Means. The experimentations are carried out on standard benchmark Sclera Segmentation and Recognition Benchmarking Competition (SSRBC2015) dataset. The Experimental results reveal that the image enhancement method improves the sclera segmentation results to the best extent.