Transfer Learning for Classification of 2D Brain MRI Images and Tumor Segmentation
Onkar Rajesh Mulay1, Hemprasad Yashwant Patil2
1Onkar Rajesh Mulay, is currently a 4th year student pursuing Bachelors of Technology in Biomedical Engineering from Vellore Institute of Technology, Vellore, Tamil Nadu, India.
2Dr. Hemprasad Yashwant Patil, Assistant Professor Vellore Institute of Technology, Vellore, India.
Manuscript received on February 10, 2020. | Revised Manuscript received on February 20, 2020. | Manuscript published on March 30, 2020. | PP: 2016-2019 | Volume-8 Issue-6, March 2020. | Retrieval Number: F8321038620/2020©BEIESP | DOI: 10.35940/ijrte.F8321.038620
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The focus of the paper is to classify the images into tumorous and non-tumorous and then locate the tumor. Amongst many medical imaging applications segmentation of Brain Tumors is an important and arduous task as the data acquired is disrupted due to artifacts being produced and acquisition time being very less, so classifying and finding the exact location of tumor is one of the most important jobs. In the paper, deep learning specifically the convolutional neural network is used to demonstrate its potential for image classification task. As the learning from available dataset will be low, so we use transfer learning [4] approach, as it is a developing AI strategy that overwhelms with the best outcomes on several image classification assignments because the pre-trained models have gained good knowledge about the features by training on a large number of images. Since, medical image datasets are hard to collect so transfer learning (Alexnet) [1] is used. Later on, after successful classification the aim is to find the exact location of the tumor and this is achieved using basics of image processing inspired by well-known technique of Mask R-CNN [9].
Keywords: Alexnet, Classification, Convolutional Neural Network, Transfer Learning.
Scope of the Article: Classification.