Retinal Vessels Segmentation Based on Enhancing Multi-scale Line Detection

Medical images are helpful for doctors to detect some dangerous diseases such as diabetes, age-related macular degeneration, hypertension, and arteriosclerosis. Abnormalities in retinal blood vessels cause these severe diseases. Thus, analysis of retinal vascular features is an excellent way to detect these abnormalities that help doctors make suitable treatment plans in the early stage. One of the retinal analysis tasks is retinal blood vessel segmentation which plays an important role because it needs to be done before any measurement can be created. In this paper, we proposed an approach for segmentation in retinal vessels by enhancing multi-scale line detection. The proposed method includes two steps. Firstly, the input image quality is improved using the Semi Translation Invariant Contourlet Transform (STICT). Finally, the result of the first step is applied by enhancing the multi-scale line detection technique (EMLD) to detect retinal blood vessels. As a result, our method can work effectively to segment more vessels. The performance of the proposed method evaluates both quantitatively and qualitatively on publicly available DRIVE dataset with an accuracy average reaching to 0.956. The result of the proposed method is better than the other methods.