Brain Tumor Segmentation Using U-Net++ and Classification by Convolutional Neural Network

  • Malathi M, SaranyaE.R, Alex M, Kathiravan S

Abstract

The progressive models for medical MRI images are modifications of U-Net++ and Fully Convolutional Network (FCN). The representations have main benefits like using the CNN classification and find the accurate depth of the brain tumour. Resulting the proposed work compare to all other classification, this CNN method gives accurate feature values. For segmentation the proposed work uses U-Net++ algorithm only because this is the greatest segmentation algorithm methods compared to all other segmentation methods. After segmentation the features will be extracted from each image, finally each image of the feature will be compared to dataset after that depends upon the input image feature that will be classified using CNN classifier. For the input image, the work using the MRI images because comparing to CT images this gives clear images of the brain. In previous brain tumour methodology the only segmentation part will be done, so for finding the depth and type of tumour the doctor must know about the deep concept of the brain tumour, but in proposed image, don't want to know the deep concept of the brain tumour because the CNN classifier itself decide the depth of the brain tumour and the type of the tumour.

Published
2020-04-18
How to Cite
Malathi M, SaranyaE.R, Alex M, Kathiravan S. (2020). Brain Tumor Segmentation Using U-Net++ and Classification by Convolutional Neural Network. International Journal of Control and Automation, 13(02), 553 - 563. Retrieved from https://sersc.org/journals/index.php/IJCA/article/view/10186
Section
Articles