Segmentation of Retinal Features Using Hybrid BINI Thresholding in Diabetic Retinopathy Fundus Images

Gaussian Kernel Prompted Fuzzy C Means Algorithm with Multi- Object Contouring Method for Segmenting NPDR Features in Diabetic Retinopathy Fundus Images

“…Shalini et al [20] have proposed a comparison work on the detection of hard exudates in diabetic retinopathy fundus images using the principles of Fuzzy-C Means and K-means algorithm. The method involves techniques like green channel extraction, median filter, Binary thresholding, K-means, Fuzzy-C-Means.…”

“…The fundus retinal input image acquired from the DR database is standardized using the bi-cubic interpolation area method [35] for the resizing of the image. In the first phase, the resized image undergoes preprocessing by using the techniques like Green channel extraction, median filter for image enhancement then Binarized contour tracing (BCT), hybrid BINI Thresholding [36], extended minima transform algorithms are applied to detect the Non-DR features like blood vessels, optic disk, and fovea. In the second phase, the detected Non-DR features are removed from the input image using mathematical arithmetic operation (MAO) and pixel replacement method (PRM).…”

“…This green Channel image is applied with the background subtraction process using numerous image processing techniques to detect and remove the unwanted Non-DR features like Blood Vessel, Optic Disc, and Fovea so that the NPDR features are more projective. The first feature Blood Vessel is detected using the binarized contour tracing (BCT) method and the second feature Optic disc is segmented using BINI Thresholding are shown in columns 3.3 and 3.4, which has been already done, and Notes described in the previous work [36]. The proposed work: detected the third feature called Fovea feature using extended minima transform method, as shown in column 3.5.…”