A Modified Intuitionistic Fuzzy Clustering Approach for Sclera Segmentation

Abstract

The culmination of technology-driven society has bought individuals in a lot of digital transactions. How legitimate these transactions are, is the question of the hour. Biometric-enabled transactions have gained popularity. Sclera, a new biometricbased recognition system promises to add value to such transactions. However, this recognition is purely based on the effective segmentation of the sclera from the occluded region of the eye. This work proposes a Modified Intuitionistic Fuzzy Clustering approach for the effective segmentation of sclera images. The traditional fuzzy set assumes that the non-membership value is always the complement of the membership value. But in the true sense, this assumption is not always correct because of hesitation. To alleviate the problems of hesitation degree and noise in the images, the Modified Intuitionistic Fuzzy C-Means (MIFCM) is proposed and tested against the Sclera Segmentation and Recognition Benchmarking Competition (SSRBC2016) and Sclera Segmentation Benchmarking Competition (SSBC 2019) dataset. The experimentation results reveal that the proposed work complements the other existing methods and variants of Fuzzy C-Means.