Deep Convolutional Neural Network based Early Diagnosis of multi class brain tumour classification system

Authors

  • Dr. D. Stalin David , D. Saravanan , Dr.A.Jayachandran

Abstract

Brain cancer is the world's second leading cause of death for both men and women, and in the next several decades, it is expected to become the leading cause of death. It has been shown that the most effective methods of reducing mortality are early detection and treatment of brain cancer. The automatic MR image classification of meningioma has been widely used in cancer diagnosis and treatment. In order to analyze medical images, Convolutionary Neural Networks (CNNs) have quickly become the method of choice. In many practical applications, such as medical image classification, CNNs, to date, have achieved breakthrough performance. In this work, due to the benefits of DL, a deep Convolutionary Neural Network (DCNN) based architecture is presented to diagnose and classify brain tumours and assign grade to them. CAD systems can provide the diagnosis depending on the specific characteristics present in the medical images. A comprehensive method for diagnosing the cancerous region in the MRI images is proposed in the present study. The overall classification of brain tumour accuracy of the proposed DCNN method is 97.73, ResNet50 is 95.62, DenseNet121 is 96.22 and 95.91. Compared to the models, the suggested model DCNN produced better results based on the simulation results

Published

2020-10-17

Issue

Section

Articles