Optimized U-net model for precise retinal blood vessel segmentation from colour fundus images.

Retinal blood vessels segmentation plays significant role in the detection and diagnosing of various eye diseases viz. diabetic retinopathy, glaucoma, and hypertensive retinopathy. This paper presents a deep learning-based novel approach to the automated segmentation of retinal vessels present in colour fundus images. The proposed approach utilizes a Convolutional Neural Network (CNN) with modified U-Net architecture and advanced image-processing techniques to accurately extract retinal blood vessels present in fundus images. The data used for training and evaluation includes data available in the benchmark retinal fundus image databases. (DRIVE and CHASEDB1 datasets etc.). The major objective of this work is to improve the early screening and diagnosis of retinal diseases by providing an optimal and reliable blood vessel extraction tool for Doctors. The developed approach demonstrates promising retinal vessel segmentation results with respect to accuracy and efficiency. Further performance is compared with existing works to show the effectiveness and wholesome of the given approach enhances the early screening and diagnosis of retinal diseases, primarily for use in automated screening systems.