Publication

Abstract : Uncontrollable growth of cells may lead to brain tumors and may cause permanent damages to the brain or even death. To make early diagnosis and treatment, identifying the position and size of tumors is identified as a tedious and troublesome problem among the existing computer-aided diagnosis systems. Moreover, the progression of tumors may vary among the patients with respect to shape, location, and volume. Therefore, to effectively classify and diagnose the brain tumor images according to severity stages follows the sequence of processing such as pre-processing, segmentation, feature extraction, and classification techniques to carrying out the appropriate treatment. To enhance the performance of brain tumors detection and diagnosis, an adaptive neuro-fuzzybased suggestion system ANFSS is proposed with an effective shape-based feature selection technique. Then, the performance of proposed ANFSS is compared with existing classifier models in terms of brain tumor detection and proposed model achieves 98.8% accuracy in prediction of tumor.