We describe a dynamic shape feature method to identify the lesion present in the retina. Diabetic retinopathy [DR] happens when these small blood vessels leak blood and different liquids.  This makes the retinal tissue swell, bringing about overcast or obscured vision. A PC supported screening and reviewing framework depends on the programmed discovery of injuries is required for Screening of diabetic retinopathy. DR show dim lesions, for example, Micro aneurysms (MA) and haemorrhages (HE) and bright lesions, such as, exudates and cotton-fleece spots.  The existing systems utilized normal strategies, for example, adaptive fuzzy thresholding  However, after finding the detected vessels, these lesions are lost and not recovered in resulting handling and gives less characterization precision.. Automatic detection of both microaneurysms , haemorrhages and exudates in color fundus image is described and validated based on new set of shape features called dynamic shape features (Elongation, eccentricity, circularity, solidity, rectangularity) that do not require precise segmentation of the regions to be classified. These features represent the evolution of the shape during image flooding and allow discriminating between lesions and vessel segments. Finally machine learning technique random forest classifier is used to classify the diabetic retinopathy type based on dynamic shape features.
Lesions, fundus, elongation, eccentricity circularity, solidity, rectangularity, Microaneurysms, Haemorrhages, exudates, random forest classifier
J. Asha Jenia Merlin , D.Regi Timna "Dynamic Shape Features Based Retinal Lesions Classification Using Random Forest Classifier" Iconic Research And Engineering Journals Volume 2 Issue 10 2019 Page 87-93
J. Asha Jenia Merlin , D.Regi Timna "Dynamic Shape Features Based Retinal Lesions Classification Using Random Forest Classifier" Iconic Research And Engineering Journals, 2(10)