Title: Application of deep learning approach for detecting brain tumour in MR images

Authors: Jyoti Agarwal; Manoj Kumar; Anuj Rani; Sunil Gupta

Addresses: Department of Computer Science and Engineering, Graphic Era Deemed to be University, Dehradun, India ' School of Computer Science, University of Petroleum and Energy Studies, Dehradun, India ' Department of Computer Science, G L Bajaj Institute of Technology and Management, G. Noida, India ' School of Computer Science, University of Petroleum and Energy Studies, Dehradun, India

Abstract: A tumour is an abnormal mass of tissues, which consume normal body cells, kill them, and continue to increase in size. For detection of infected tumour area and lesions, magnetic resonance imaging has been used widely in medical field. Image processing and machine learning is also used widely for brain tumour detection and segmentation, but they are not the most appropriate ones, therefore methods involving deep learning are also proposed for the same. In this paper, six traditional machine learning classification algorithms are compared. Afterwards, convolutional neural network is implemented using Keras and TensorFlow in python. Two different CNN based models VGG16 and DenseNet available in Keras trained on imagenet dataset is also used. The dataset contains in total 253 images, which were later augmented to train the model better. From results, it was analysed that deep learning algorithms yield better results than the traditional ML classification algorithms.

Keywords: brain tumour; CNN; deep learning; model; pooling; DenseNet; TensorFlow.

DOI: 10.1504/IJCIS.2023.132210

International Journal of Critical Infrastructures, 2023 Vol.19 No.4, pp.340 - 353

Received: 21 Jul 2021
Accepted: 19 Aug 2021
Published online: 13 Jul 2023 *

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