Robust Classification of Primary Brain Tumor in Computer Tomography Images Using K-NN and Linear SVM

Computer Tomography (CT) Images are widely used in the intracranical hematoma and hemorrhage. In this paper we have developed a new approach for automatic classification of brain tumor in CT images. The proposed method consists of four stages namely preprocessing, feature extraction, feature reduction and classification. In the first stage Gaussian filter is applied for noise reduction and to make the image suitable for extracting the features. In the second stage, various texture and intensity based features are extracted for classification. In the next stage principal component analysis (PCA) is used to reduce the dimensionality of the feature space which results in a more efficient and accurate classification. In the classification stage, two classifiers are used for classify the experimental images into normal and abnormal. The first classifier is based on k-nearest neighbour and second is Linear SVM. The obtained experimental are evaluated using the metric similarity index (SI), overlap fraction (OF), and extra fraction (EF). For comparison, the performance of the proposed technique has significantly improved the tumor detection accuracy with other neural network based classifier.