An Efficient Iris Segmentation and Classification of Drunk Person Using Modified Circle Hough Transform

Abstract

With various publications for Iris Recognition (IR) in the past few years, a more attractive research topic is Biometrics. Deformation in the iris pattern is created when the pupil dilates/constricts owing to alcohol consumption probably affecting IR performance. The eye's pupil region was not segmented by the prevailing segmentation schemes along with having low accuracy. An efficient iris segmentation model along with clustering the drunk person utilizing a Modified Circle Hough Transform (MCHT) was proposed in this scheme to tackle those issues. Noise reduction, iris segmentation, pupil segmentation and drink and no-drink person’s clustering are the steps comprised in the proposed model. The input images are gathered from the database initially. The noise is abated from the image by utilizing Median Filtering (MF). The Canny Mathematical Morphology (CMM) was employed to segment the iris part from the noise-free image. Subsequently, the MCHT executed the pupil segmentation as of the segmented iris image. The system’s accuracy along with robustness is elevated by performing this alteration. Consequently, the Matrix Based Clustering (MBC) clustered the drinker and the non-drinker person. The experiential outcomes exhibit that better performance was proved by the proposed method than the top-notch models.