Title: A deep learning approach for brain tumour detection system using convolutional neural networks

Authors: T. Kalaiselvi; S.T. Padmapriya; P. Sriramakrishnan; K. Somasundaram

Addresses: Department of Computer Science and Applications, The Gandhigram Rural Institute (Deemed to be University), Gandhigram, Tamil Nadu, 624 302, India ' Department of Computer Science and Applications, The Gandhigram Rural Institute (Deemed to be University), Gandhigram, Tamil Nadu, 624 302, India ' Department of Computer Applications, Kalasalingam Academy of Research and Education (Deemed to be University), Krishnankoil, Tamil Nadu, 626 126, India ' Department of Computer Science and Applications, The Gandhigram Rural Institute (Deemed to be University), Gandhigram, Tamil Nadu, 624 302, India

Abstract: The proposed work is aimed to develop convolutional neural network (CNN) architecture based computer aided diagnostic system for human brain tumour detection process from magnetic resonance imaging (MRI) volumes. CNN is a class of deep learning networks that are commonly applied to analyse voluminous images. In the proposed method, a CNN model is constructed and trained using MRI volumes of BraTS2013 data. More than 4000 images of normal and tumour slices are used to train the proposed CNN system with 2-layers. The system is tested with about 1000 slices from BraTS and got very accurate results about 90-98% of accuracy. Further, the performance of proposed CNN system is tested by taking a set of clinical MRI volumes of popular hospital. The obtained results are discussed and focused for the future improvement of the proposed system.

Keywords: neural networks; MRI; magnetic resonance imaging; brain tumour; deep learning; tumour detection; CNN; convolutional neural network; BraTS Dataset; activation functions; WBA datasets.

DOI: 10.1504/IJDSDE.2021.120046

International Journal of Dynamical Systems and Differential Equations, 2021 Vol.11 No.5/6, pp.514 - 526

Received: 18 Apr 2019
Accepted: 02 Nov 2019
Published online: 05 Jan 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article