BRAIN TUMOR CLASSIFICATION BY FISHER VECTOR AS FEATURES’ REPRESENTATION AND LOGISTIC REGRESSION AS CLASSIFIER FOR T1-WEIGHTED CONSTRAST-ENHANCED MRI IMAGES

• Main Author: Faqih Al Mubarok, Abdullah
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39514
• Institution: Institut Teknologi Bandung

Brain tumor is a collection of abnormal cells in brain tissues. This final project is intended to design an algorithm for classifying three different types of brain tumor such as meningioma, glioma and pituitary from T1-weighted constrast-enhanced MRI brain images and implement it into web-based application. Meningioma is a brain tumor located adjacent to the meninges or brain membrane. Meanwhile, glioma and pitutiray occur on glia cells and pituitary gland respectively. In this work, the classification is based on several steps such as dense local patches to obtain the lower-order features from all of the segmented area of brain tumor, determining the representation of all of the training data into gaussian mixture models (GMM), determining the higher-order feature from the training and testing data with Fisher Vector, and classification process with linear classifier.