LAUSR

Abstract The Circle Hough Transform (CHT) is one of the popular circle detection algorithm in image processing and machine vision application, favored for its tolerance to noise. Nevertheless, it involves huge computation and excessive memory requirements. Because of its drawbacks, various modifications have been suggested to increase its performances. In this paper, we present a new modification of the CHT method developed for an automatic biometric iris recognition system. The novelty of this method resides on the use of the incremental property in order to reduce the resources requirement and the parallel property for decreasing the computation time. The incremental property is obtained using the approximation of the used trigonometric functions, while the parallel property is achieved by calculating, at the same time, several point coordinates of the circle. The software implementation and the validation have been done in C++ and MATLAB on real images. The errors analysis and the performances of the proposed method against the basic CHT method are presented in this paper.