Kornea hastalıklarında klinik kararı geliştirmek için topografik parametrelere dayalı bilgisayar destekli tanı

Computer-Aided Diagnosis is an essential topic in the medical image. it is a sophisticated procedure in medicine that assists physicians in the interpretation of medical images. A Human cornea is the front see-through shield of the eye. It refracts light onto the retina to induce vision. Therefore, any defect in the cornea may lead to vision disturbance. This deficiency is estimated by sets of topographical images measured and assessed by ophthalmologists. Consequently, an important priority is the early and accurate diagnosis of diseases that may affect corneal integrity through the use of machine learning algorithms. Corneal images produced by a Pentacam device can be subjected to rotation or some distortion during acquisition; therefore, accurate diagnosis requires the use of local features in the image. Accordingly, new algorithms proposed in this work to overcome these challenges and improve cornel conditions diagnosing. Firstly, a SWFT algorithm suggested to extract the local features from the corneal images. Wavelet transform used to produce images with different scales instead of using the Difference of Gaussians (DoG) as in the standard SIFT algorithm. Secondly, IG-GLCM algorithm proposed to overcome the drawback of GLCM algorithm known as a time-consuming defect. In IG-GLCM the image gradient is measured in different directions then apply the GLCM to generated images. Thirdly, investigate the use of SIFT with multi-scale subbands of wavelet transform. Finally, new algorithm called Local Information Pattern descriptor suggested to overcome the lack of local binary patterns that loss of information from the image and solve image rotation issue. The LIP based on utilizing the sub-image center intensity for estimating neighbors' weights that can use to calculate what so-called contrast Based Centre (CBC) value, as well as local pattern (LP). The Naive Bayes, KNN, decision tree, and SVM employed as classifiers. The proposed model is trained and tested successfully on a collected dataset which comprises 4848 images of different maps. Keywords: Computer-aided diagnosis, feature extraction, machine learning, support vector machines, local Binary Pattern, GLCM