A novel boosted ada-boost classifier for MRI-based brain tumour detection

Computer-aided analysis of brain images collected through Magnetic Resonance imaging (MRI) and other tomography techniques is an important and complicated task because of the intrinsic nature of the pictures. Selecting a highly accurate segmentation algorithm is a vital step in digital image processing to detect the tumor, edema, and dead tissues in brain MR images; likewise, selecting the best classification algorithm will identify the brain tumor stages. A high-performance fuzzy c-means (FCM) clustering segmentation technique associated with a boosted Ada-Boost classification algorithm has been designed in this work to get a more accurate result. The particle swarm optimization (PSO) program is adopted in feature selection to reduce the computational time required for the entire process and increase the accuracy of tumor identification. The modification in the collaboration of segmentation, feature extraction, and classification algorithm implemented here raised the accuracy of the brain tumor images of the dataset to 98.9 %. The outcome of the image analysis efficiency was compared with other algorithms like Ada-Boost and RBF-NN.