tDetection and quantitative measurement of variations in the retinal blood vessels can helpdiagnose several diseases including diabetic retinopathy. Intrinsic characteristics of abnor-mal retinal images make blood vessel detection difficult. The major problem with traditionalvessel segmentation algorithms is producing false positive vessels in the presence of dia-betic retinopathy lesions. To overcome this problem, a novel scheme for extracting retinalblood vessels based on morphological component analysis (MCA) algorithm is presented inthis paper. MCA was developed based on sparse representation of signals. This algorithmassumes that each signal is a linear combination of several morphologically distinct compo-nents. In the proposed method, the MCA algorithm with appropriate transforms is adoptedto separate vessels and lesions from each other. Afterwards, the Morlet Wavelet Transform isapplied to enhance the retinal vessels. The final vessel map is obtained by adaptive thresh-olding. The performance of the proposed method is measured on the publicly availableDRIVE and STARE datasets and compared with several state-of-the-art methods. An accu-racy of 0.9523 and 0.9590 has been respectively achieved on the DRIVE and STARE datasets,which are not only greater than most methods, but are also superior to the second humanobserver’s performance. The results show that the proposed method can achieve improveddetection in abnormal retinal images and decrease false positive vessels in pathologicalregions compared to other methods. Also, the robustness of the method in the presence ofnoise is shown via experimental result.

Keywords