Retinal Image Analysis with Shearlets

Abstract

In this paper we propose a method for segmenting blood vessels in retinal images based on the shearlet transform. Shearlets are a relatively new directional multi-scale framework for signal analysis, which have been shown effective to enhance signal discontinuities such as edges and corners at multiple scales. The algorithm we propose builds on the idea of enhancing ridgelike structures at different scales by computing the shearlet transform with an appropriate mother function, the mexican hat wavelet. This allows us to detect precisely structures of different widths. We provide an experimental analysis of our approach on a benchmark dataset and we show very good performances in comparison with other multi-resolution methods from the state of the art.