| The morphological changes of the retinal blood vessels are closely related to the characterization of certain pathologies. Thus the retinal blood vessels analysis has been widely used in the diagnoses of diseases by ophthalmologists. However, due to the complex morphological properties of the blood vessels and the impacts of uneven illumination, as well as the appearance of pathological areas in the retinal images, none of the existing methods can achieve the most satisfying performance in all aspects so far.To solve retinal blood vessel segmentation problems about the imaging limitations and the complex morphological characteristics of the blood vessels in normal and abnormal images, an automatic method by using the random walks algorithms based on the centerlines is proposed in this study. The proposed method including image segmentation preprocessing, centerlines extraction using the multi-scale and multi-direction morphological method and the divergence of the normalized gradient vector field, automatic seed groups positioning according to the centerlines and the hessian-based directional information of vessels, and vessels segmentation based on the reformed adaptive random walks. And the key steps of the proposed method include the centerline extraction, the labeled seed groups positioning and the weight function reforming of random walks.The Experiments of the proposed method are implemented on the publicly available STARE (the Structured Analysis of the Retina) database. The results are compared to other existing retinal blood vessel segmentation methods with respect to the accuracy, sensitivity and specificity, and the proposed method is proved to be more sensitive in detecting the retinal blood vessels in both normal and pathological areas, which can provide a useful basis for the automatic quantitative analysis of confirmed diagnosis. |
