A New Sclera Segmentation and Vessels Extraction Method for Sclera Recognition

Abstract

As a unique biometric trait, the research of sclera blood vessels have becomes active recently, because the sclera vessels can be captured under visible-wavelength light condition rather than near infrared light condition. However, the performance of sclera identification system degrades a lot due to unsatisfactory sclera segmentation and tedious extraction process. In this paper, we propose a novel sclera segmentation algorithm and an improved sclera vessels extraction method. The accurate and efficient sclera segmentation method is proposed based on improved OSTU algorithm. Besides, the vessels extraction method is put forward by adaptive histogram equalization method and Gabor filters. The experimental results on UBIRIS.vl database prove that the proposed sclera vessels extraction algorithm performs well and the sclera segmentation method has an obvious improvement in terms of efficiency and accuracy over than other segmentation algorithms.