Semantic Classification of Brain Tumors in MRI Images using Encoder-Decoder Networks

  • Ahmed Kareem Alzeyadi, Sajjad Ali Ettyem, Ali Abdulhasan Kadhim, Ahmed Raheem Hassan

Abstract

Nowadays, the use of deep learning in scientific, technical and professional fields has become common and the processing of raw and very large data are easily possible by these structures. On the other hand, traditional and shallow methods require the extraction of the examined features, which are challenging and lack high accuracy, but deep structures have mechanisms that can automatically extract these features and satisfy the needs of the extraction problem. With the advent of new medical technologies, awareness of biomechanisms is increasing, and we can treat diseases better than ever before. Deep learning due to computing power and high accuracy has helped a lot in this regard. In this study, the application of deep learning in the classification of brain tumors in medical images has been used. An encryption-decryption network has been used in this path, which according to the obtained results has good accuracy and high speed and optimal memory usage. In this way, the dataset (Brats) provided for brain tumor by (MICCAI) was used. We were able to achieve a dice score of 0.72 and an accuracy of 0.98 and achieve better results with less memory usage and faster than other algorithms in this field.

Published
2021-10-21
How to Cite
Ahmed Kareem Alzeyadi, Sajjad Ali Ettyem, Ali Abdulhasan Kadhim, Ahmed Raheem Hassan. (2021). Semantic Classification of Brain Tumors in MRI Images using Encoder-Decoder Networks. Design Engineering, 6188-6197. Retrieved from http://thedesignengineering.com/index.php/DE/article/view/5580
Section
Articles