Detection of Diabetic Retinopathy from Retinal Fundus Image using Wavelet based Image Segmentation

Abstract

In case of Diabetic Retinopathy, retina is damaged because of fluid leaks from blood vessels into the retina. According to ophthalmologists, some basic features are there to recognize diabetic retinopathy, such as blood vessel area, exudates, hemorrhages, micro-aneurysms and texture. Presence of exudates within the macular region is a main hallmark of diabetic which identifies its detection with a high sensitivity.Hence, detection of exudates is an important diagnostic task that can be determined by means of morphological techniques for better result. Here image segmentation has done to detect Diabetic Retinopathy by extracting exudate from retinal fundus images. Here the final extracted exudates contained the boundaries of proper segmented regions which can identify the severity of disease. Using Wavelet Transform decomposition method and morphological operations the feature like exudate is possible to extract from retinal fundus image to detect Diabetic Retinopathy. Here 4 datasets have been taken such as STARE, DRIVE, DIARETDB1 and DIARETDB0 for extraction of exudate from retinal image in which the proposed method obtained an average value of accuracy at 95.72%, sensitivity at 95.87% and specificity at 96.2% respectively.

Deep Learning and Medical Image Processing Techniques for Diabetic Retinopathy: A Survey of Applications, Challenges, and Future Trends

“…This study used a k-means clustering algorithm to categorize disease symptoms and separate them as clusters at various stages. The study in [ 72 ], proposed a new wavelet-based image segmentation method that involves transforming input images into different orientations while maintaining the standard pixels needed to detect image alterations. To state a clear representation of the specific item, we used morphological approaches to retain specific image segment frontiers.…”

Application of Structural Deformation Monitoring Based on Close‐Range Photogrammetry Technology

“…Among many image segmentation methods, gray threshold segmentation is the most commonly used method. is kind of method mainly determines an appropriate threshold value, then compares the threshold value with the gray value of the image, and divides the image according to the threshold value [30,31]. Image recognition plays an important role in computer vision, artificial intelligence, and other fields and is widely used in motion recognition, face recognition, image processing, and other related research [32].…”

Image Texture Analysis for Medical Image Mining: A Comparative Study Direct to Osteoarthritis Classification using Knee X-ray Image

“…Furthermore, texture analysis applications have been widely applied in the medical research for the detection of various diseases including: tumor heterogeneity [9], [17], [18]; brain tumor [19], [20]; head and neck cancer [21], [22]; emphysema [23], [24]; prostate segmentation [25]- [27], colon cancer [28], [29]; small vessel disease and blood brain barrier [30], breast cancer [31]- [34]; skin cancer [35]- [37] retinal vessel segmentation [38], [39] and lung cancer [40], [41].…”

“…Average One-Dependence Estimators (AODE) is a probabilistic classification learning technique which improves the Naïve Bayesian classifier [54] by addressing the problem of attribute-independence. For instance, in the class y, which has a set of features x1,..., xn, AODE can be used to find the probability of each class y by using the following equation: (38) where P " is represented an estimate of P; F has represented the frequency; m has represented a user specified minimum frequency.…”