Thesis (Ph.D.) - Cairo University - Faculty of Graduate Studies for Statistical - Department of Computer and Information Science

Brain cancer importance emanates from the importance of the brain as an organ and its functions. It has a great effect on the whole human body. Identification brain cancer according to its type, it refers to a multiclass classification problem in the machine learning world. In the real-world, object detection and classification face numerous challenges because the object has a large variation in appearances. Feature extraction is a very important and crucial stage in recognition system. It has been widely used in object recognition, image content analysis and many other applications. Feature extraction is the best way/method to recognize images in the field of medical images. However, the selection of proper feature extraction method is equally important because the classifier output depends on the input features.This research proposes an image classification methodology that automatically classifies human brain magnetic resonance MR images.The research has two components; the first component focused on that human brain is normal or abnormal. If it is abnormal brain the second component appears to decide its type.The proposed methods consist of four main stages: preprocessing, feature extraction, feature reduction and classification, followed by evaluation for each direction. The first component consists of four stages, it starts with noise reduction in MR images. In the second stage, the features related to MRI are obtained using Gabor filter. In the third stage, the features of MRI are reduced to the more essential features using kernel linear discriminator analysis (KLDA). In the last stage, the classification stage, two classifiers have been developed to classify subjects as normal or abnormal MRI human images. Whereas the first classifier is based on Support Vector Machine (SVM), the second classifier is based on K-Nearest Neighbor (KNN) on Euclidean distance

Issued also as CD