![Open Access](https://i-scholar.in/lib/pkp/templates/images/icons/fulltextgreen.png)
![Restricted Access](https://i-scholar.in/lib/pkp/templates/images/icons/fulltextred.png)
![Open Access](https://i-scholar.in/lib/pkp/templates/images/icons/fulltextgreen.png)
![Open Access](https://i-scholar.in/lib/pkp/templates/images/icons/fulltext_open_medium.gif)
![Restricted Access](https://i-scholar.in/lib/pkp/templates/images/icons/fulltextred.png)
![Restricted Access](https://i-scholar.in/lib/pkp/templates/images/icons/fulltext_restricted_medium.gif)
A Significance Test-Based Feature Selection Method for Diabetic Retinopathy Grading
Subscribe/Renew Journal
Diabetic retinopathy is the dangerous eye disease cause the blindness in worldwide. A fundus camera provides digitized data in the form of a fundus image that can be effectively used for the computerized automated detection of diabetic retinopathy. A completely automated screening system for the disease can largely reduces the burden of the speialist and saves cost. In this article, a computerized diabetic retinopathy grading system, which digitally analyses retinal fundus image, is implemented. The proposed system proceeds on four stages. The first stage points toward reducing noise and other disturbances that occur during image acquisition and may lead to false detection of the disease. This is accomplished by using various image processing techniques. Next, different statistical texture features are extracted which serves as the guideline to identify and grade the severity of the disease. The feature extraction stage is realized in two steps. In the first step, the gray level co-occurrence matrix (GLCM) is computed to extract texture images. We used contrast, energy, homogeneity and entropy measures based on GLCM to characterize texture images. Afterward, these images are combined and the statistical properties of their intensity histograms are obtained. The statistical features extracted are the mean, standard deviation, smoothness, third moment, uniformity and entropy which signify the important texture features of the retinal image. In the third stage, the non-parametric statistical hypothesis test, Kruskal-Wallis test, is used to select the statistical significant features that are capable to distinguish among different diabetic retinopathy classes. Finally, the selected features are fed into the K Nearest Neighbor (K-NN) for classification. The performance of the proposed system is evaluated by using DIARETDB0 database. The images in the dataset are classified based on the lesion type (exudates, Microaneurysms and Hemorrhages) exists into four groups. The results show that the system has high ability to correctly detect retinopathy group 2 (red small dots, hemorrhages, hard exudates, soft exudates) against other groups and has good ability to correctly detect diabetic retinopathy of group 3 (red small dots, hemorrhages, hard exudates) against group 4 (normal).
Keywords
Diabetic Retinopathy, Gray Level Co-Occurrence Matrix, K Nearest Neighbor (K-NN), Kruskal-Wallis Test, Texture Features.
User
Subscription
Login to verify subscription
Font Size
Information
![](https://i-scholar.in/public/site/images/abstractview.png)
Abstract Views: 200
![](https://i-scholar.in/public/site/images/pdfview.png)
PDF Views: 3