Multi-grade brain tumor classification using deep CNN with extensive data augmentation

• Published in: Journal of computational science Vol. 30; pp. 174 - 182
• Main Authors: Sajjad, Muhammad, Khan, Salman, Muhammad, Khan, Wu, Wanqing, Ullah, Amin, Baik, Sung Wook
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2019
• Subjects: Brain tumor; Computer-aided diagnosis; Convolution neural network; Data augmentation; Deep learning; MRI; Multi-grade brain tumor classification; Tumor segmentation; Brain tumor; Deep learning; Multi-grade brain tumor classification; Convolution neural network; MRI; Data augmentation; Tumor segmentation; Computer-aided diagnosis
• ISSN: 1877-7503; 1877-7511
• DOI: 10.1016/j.jocs.2018.12.003

•Proposal of a deep CNN based multi-grade brain tumor classification.•Tumor segmentation using a deep learning approach.•Extensive data augmentation for avoiding lack of data problem for multi-grade tumor classification. Numerous computer-aided diagnosis (CAD) systems have been recently presented in the history of medical imaging to assist radiologists about their patients. For full assistance of radiologists and better analysis of magnetic resonance imaging (MRI), multi-grade classification of brain tumor is an essential procedure. In this paper, we propose a novel convolutional neural network (CNN) based multi-grade brain tumor classification system. Firstly, tumor regions from an MR image are segmented using a deep learning technique. Secondly, extensive data augmentation is employed to effectively train the proposed system, avoiding the lack of data problem when dealing with MRI for multi-grade brain tumor classification. Finally, a pre-trained CNN model is fine-tuned using augmented data for brain tumor grade classification. The proposed system is experimentally evaluated on both augmented and original data and results show its convincing performance compared to existing methods.