MRI Brain Tumor Image Classification Using a Combined Feature and Image-Based Classifier

Frontiers in Psychology

S. Meenakshi¹  G. Mathivanan²  A. Veeramuthu²  V. Vijayakumar³  V. Subramaniyaswamy⁴  Ketan Kotecha⁵  Jatinderkumar R. Saini⁶ 

[1] Department of Information Technology, Jeppiaar SRR Engineering College, Chennai, India;Department of Information Technology, School of Computing, Sathyabama Institute of Science and Technology, Chennai, India;School of Computer Science and Engineering, University of New South Wales, Sydney, NSW, Australia;;School of Computing, Shanmugha Arts, Science, Technology &Symbiosis Centre for Applied Artificial Intelligence, Symbiosis International (Deemed University), Pune, India;Symbiosis Institute of Computer Studies and Research, Symbiosis International (Deemed University), Pune, India;

关键词: brain tumor;    classification;    deep neural network;    actual image;    segmented image;    combined feature and image based classifier (CFIC);   

DOI  :  10.3389/fpsyg.2022.848784

来源: DOAJ

【 摘 要 】

Brain tumor classification plays a niche role in medical prognosis and effective treatment process. We have proposed a combined feature and image-based classifier (CFIC) for brain tumor image classification in this study. Carious deep neural network and deep convolutional neural networks (DCNN)-based architectures are proposed for image classification, namely, actual image feature-based classifier (AIFC), segmented image feature-based classifier (SIFC), actual and segmented image feature-based classifier (ASIFC), actual image-based classifier (AIC), segmented image-based classifier (SIC), actual and segmented image-based classifier (ASIC), and finally, CFIC. The Kaggle Brain Tumor Detection 2020 dataset has been used to train and test the proposed classifiers. Among the various classifiers proposed, the CFIC performs better than all other proposed methods. The proposed CFIC method gives significantly better results in terms of sensitivity, specificity, and accuracy with 98.86, 97.14, and 98.97%, respectively, compared with the existing classification methods.

【 授权许可】

Unknown