Automatic brain tumor detection in Magnetic Resonance Images

Abstract

Automatic detection of brain tumor is a difficult task due to variations in type, size, location and shape of tumors. In this paper, a multi-modality framework for automatic tumor detection is presented, fusing different Magnetic Resonance Imaging modalities including T1-weighted, T2-weighted, and T1 with gadolinium contrast agent. The intensity, shape deformation, symmetry, and texture features were extracted from each image. The AdaBoost classifier was used to select the most discriminative features and to segment the tumor region. Multi-modal MR images with simulated tumor have been used as the ground truth for training and validation of the detection method. Preliminary results on simulated and patient MRI show 100% successful tumor detection with average accuracy of 90.11%.