A BrainNet Classification Technique Based on Deep Convolutional Neural Network for Detection of Brain Tumor in FLAIR MRI Images
T.H. Manoj1, M. Gunasekaran2, W.Jaisingh3
1T.H. Manoj*, Research Scholar, Bharathiar University, Coimbatore, India.
2M.Gunasekaran, Assistant Professor of Computer Science, Government Arts College, Dharmapuri, India.
3W.Jaisingh, Assistant Professor (SRG) of Computer Applications, Kumaraguru College of Technology, Coimbatore, India.
Manuscript received on September 22, 2019. | Revised Manuscript received on October 15, 2019. | Manuscript published on October 30, 2019. | PP: 3264-3269 | Volume-9 Issue-1, October 2019 | Retrieval Number: A1424109119/2019©BEIESP | DOI: 10.35940/ijeat.A1424.109119
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: Classification process plays a key role in diagnosing brain tumors. Earlier research works are intended for identifying brain tumors using different classification techniques. The logical gap between the visual representation of data captured by MRI device and the information apparent to the person evaluating poses a key challenge in the medical field. Research in computerized segmentation of tumor is widely gaining popularity nowadays, which may lead to an accurate analysis of MRI images and planned treatment of patients. The recent field of deep learning and neural networks promises to classify images with higher accuracy. This work proposes a new Brain Net classification technique that combines fuzzy c means, morphological operators and CNN to identify image regions that are suspicious. The proposed method is assessed with the help of imaging data obtained from Multimodal Brain Tumor Image Segmentation Challenge (BRATS) 2015 and IXI dataset. The effectiveness of the proposed method is computed with traditional machine learning and Convolutional Neural Networks. Experimental results show that our proposed method outperforms state-of-the-art classification on the BRATS 2015 dataset.
Keywords:: Convolutional Neural Network, Fuzzy C Means, Morphological Operator, Classification, Segmentation, MRI Image.