Retinal image mosaic base on Genetic Algorithm and automated blood vessel extracting approach

This paper presents an automatic retinal image mosaic method, which is based on the principle of blood vessels resemblance. The method depends on the characteristics of the extracted blood vessels that are segmented by combining mathematical morphology with maximum entropy algorithm. Then, the registration objective function is developed by using the maximal overlap degree of the retinal blood vessels information and its optimal parameters are searched by employing a global optimization algorithm, i.e., genetic algorithm (GA). Finally, the mosaic process of retinal image is implemented through a fade-by-fade method, which can eliminate the mosaic gaps. Experimental results indicate that the proposed approach can efficiently realize the automated retinal image registration and smooth mosaic.