An Enhanced Gabor Filter Based on Heat-Diffused Top Hat Transform for Retinal Blood Vessel Segmentation

Abstract

Precise retinal blood vessels segmentation is a vital assignment to diagnose many pathological ailments. Here, an efficient approach has been presented for the segmentation of retinal vessels that includes pre-processing, segmentation, and the post-processing stage. In the pre-processing stage, the retinal images are denoised and restored with the connected vessel lines by utilizing anisotropic diffusion filter. In the next step, retinal images are enhanced by using top hat transform. Further, Gabor filters of various orientations are applied on top hat transformed images to obtain different characteristics of the retinal images. Finally, hysteresis thresholding is applied for the segmentation. The accomplishment of the recommended methodology is inspected with other competitive combined methodologies based on median filter and Gabor filter using different performance indicators. The approach can be used effectively for diagnosis of different ocular disorders, like diabetic retinopathy and glaucoma, which can be followed by different surgical procedures for further treatment.