The method we haveapplied to detect exudates on human retina is inspired by the work described in35. Since the data set is ofcompletely different characteristic we have changed in various sides. That iswhy we are going to describe every step and the reason behind taking it.
Wehave used MATLAB version 2017a for this project and this detection consists ofthe following steps : (a)Preprocessing the image. (b)Detection of Optic disc and other artifacts. (c) Detection of exudates in terms of opticdisc and artifacts. In the preprocessingstep first we extract intensity constituents from an image. Here we are goingto work with gray-scale images because exudates are mostly visible in such images.We then apply Median Filtering for reducing the noise and apply HistogramEqualization to enhance contrast and brightness. The resulting image helps usto detect optic disc and accordingly exudates.
This works as input image.Exudates are high intensity values as well as optic disk. Therefore in order togo for exudates detection we need to find optic disc and then we need todifferentiate between optic disc and exudates near and inside the optic discarea. To do this we consider that optic disc is the largest and most circularpart in brightest portion of the image. We apply Gray Scale Closing to removeblood vessels in the retina mostly in the optic disc area. Here we take a flatdisc shaped structure element and consider the radius is eight. We thresholdthe image to binaries it and use the resulting image as a mask. Then the maskis inverted by pixels before overlaying into the original image.
We then apply reconstructionby dilation was on the overlaid image. We threshold the image and find thedifference between the original image and the reconstructed image by thealgorithm. Consequently, high intensity optic disc is detected and rests areremoved.In this part we faced abig problem of this approach. At the beginning of the process, vessels wereremoved by the Gray Scale Closing and reconstruction was applied on the imagecreated from the original image. Therefore we are going to reconstruct vesselsin the optic disc area.
But we face a problem is that we are not getting one circularoptic disc. Rather we are actually detecting two or three connected componentsin this step. To solve this problem we applied an addition dilation of thefinal mask. As a result the independent areas are connected together into acircular shape. Here we note that we have already detected artifacts and otherbright spots in the image. That is why if we use too big dilation, it can leadto merge the optic disc with those areas.For the properadditional dilation we have considered a flat disc shaped structured elementwith a radius of four. Since the optic disc and also some bright artifacts aredetected in this process, we have estimated for every component of the mask inorder to distinguish between the features some extra values.
These additionalvalues are termed as scores. Thus we have, Score =area circularity3 Here we have some caseto give attention. Since we have situation that the feature rather than opticdisc can become much larger than optic disc, we needed to give circularity moreimportance.
We take elements of size more than 1100 pixels as an optic disckeeping the rest as artifacts. Here we do not classify small areas which canbecome exudates as artifacts. At this stage after optic disc extraction andartifacts detection we are going to detect exudates. As before, high intensityblood vessels are removed by Grey Scale Closing. Then we go for to get astandard deviation image which shows the main characteristics of nearlyarranged exudates. The resulting imageis being threshold by taking the radius is six. We than remove the outsideshape of the retina and fill the holes by imfill(). We consider threshold toremove optic disc and artifacts.
Finally the result is achieved when we apply athreshold at a level 0.01 between the original and the reconstructed one. Theproduced exudates mask image is overlaid into the main image to get a propervision.