Abnormality Detection in retinal images using Haar wavelet and First order features

Diabetic Retinopathy has become a major cause of blindness. Early detection of diabetic retinopathy can prevent vision loss of human. Many automated systems have been developed to detect diabetic retinopathy. All these mechanisms are computationally expensive. They perform segmentation and then classify these regions into exudates and non-exudates. Proposed method classify quickly image as Diseased or Healthy one. The image is classified based on Haar wavelet and First order statistical features obtained from detailed coefficients. The authors used Diaretdb0 and Diaretdb1 database and performed a comparative study of Decision tree classifier and Knn classier. Authors obtained encouraging results with classification accuracy of 85% with Knn classifier and 75% accuracy with decision tree classifier.