In this work, the student will work with a real dataset of MRI brain images delivered by faculty hospital (FN Brno). Further, she will inspect machine learning classification strategies and implement and apply at least two of them to brain image data. This work examines different class groupings of high-grade glioma, low-grade glioma, meningioma, and metastases based on various characteristics. Compared to standard classification tasks, the main objective here is to check,  whether the extended set of modalites (so called IVIM, DTI) can help in improving the classification accuracy. The output of this work will be a text describing the optimal solution and an open, operational and well documented software.