A Hough Transform Based Feature Extraction Algorithm for Finger Knuckle Biometric Recognition System

Abstract

Finger Knuckle Print is an emerging biometric trait to recognize one’s identity. In this paper, we have developed a novel method for finger knuckle feature extraction and its representation using Hough transform. Hough transform plays a significant role in locating features like lines, curves etc., present in the digital images by quantifying its collinear points. This paper formulates the Elliptical Hough Transform for feature extraction from the captured digital image of finger knuckle print (FKP). The primary pixel points present in the texture patterns of FKP are transformed into a five dimensional parametric space defined by the parametric representation in order to describe ellipse. The discrete coordinate of the five dimensional coordinate spaces along with its rotational angle are determined and characterized as the parameters of elliptical representation. These parametric representations are the unique feature information obtained from the captured FKP images. Further, this feature information can be used for matching various FKP images in order to identify the individuals. Extensive experimental analysis was carried out to evaluate the performance of the proposed system in terms of accuracy. The obtained results shows the lowest error rate of EER = 0.78%, which is found to be remarkable when compared to the results of existing systems presented in the literature.