Automated Segmentation of Retinal Blood Vessels using Optimized Gabor Filter with Local Entropy Thresholding

The paper presents an automated method for segmenting retinal blood vessels using an optimized Gabor filter combined with local entropy thresholding techniques. Retinal blood vessel detection is crucial for diagnosing various eye-related diseases, such as diabetic retinopathy and glaucoma. The proposed approach involves multiple stages, beginning with the loading and preprocessing of retinal images. The images are first converted to grayscale, followed by contrast enhancement using adaptive histogram equalization (adapthisteq) to improve image quality. An optimized Gabor filter is then applied to extract directional features that highlight the blood vessels in the image. Local entropy thresholding is employed to segment the vessels by differentiating the blood vessels from the background. The image is subsequently masked, resulting in a binary image that highlights the retinal blood vessels. Various evaluation metrics, such as accuracy, sensitivity, and specificity, are calculated to assess the effectiveness of the segmentation process. The method demonstrates promising results in accurately detecting and segmenting retinal blood vessels, offering a valuable tool for automated medical image analysis. This approach can significantly aid ophthalmologists in diagnosing retinal diseases by providing clear and precise vessel detection, reducing the time and effort required for manual examination.

Keywords: Dataset, Image Processing Techniques, Segmentation.