[MS03] Biomedical Imaging in Disease Diagnosis

[MS03-02] A Novel Line Detector Based Approach for the Efficient Segmentation of the Retinal Vasculature

Varun Makkar¹, Arya Tewary¹, B. V. Rathish Kumar², *Rajesh Kumar Pandey¹ (1. Indian Institute of Technology (BHU) Varanasi (India), 2. Indian Institute of Technology Kanpur (India))

Keywords:

Line detector,Retinal blood vessel segmentation,Fractional filter

Retinal blood vessel segmentation is vital for the early detection of eye diseases like diabetic retinopathy, glaucoma, and hypertension. It enables accurate diagnosis, monitoring, and treatment planning by analyzing vessel structure and abnormalities. Automated segmentation aids ophthalmologists, enhances screening efficiency, and supports large-scale medical image analysis and telemedicine applications. In this work, we present an enhanced line detector-based approach for efficient segmentation of the retinal blood vessels present in the low-contrast regions of the retinal fundus images. Here, first, we denoise the retinal image using a fractional filter. Thereafter, enhanced retinal vasculature is obtained by applying a novel multiscale line detector on the denoised retinal image. Finally, a thresholding operation is applied to obtain the segmented vessels. The proposed algorithm achieves consistent performance with respect to multiple performance metrics on publically available datasets, namely DRIVE and STARE.