Brain tumour classification of magnetic resonance images using a novel CNN-based medical image analysis and detection network in comparison to VGG16

Abstract

Aim: This study aims at developing an automatic medical image analysis and detection for accurate classification of brain tumors from MRI dataset. The study implemented our novel MIDNet18 CNN architecture in comparison with the VGG16 CNN architecture for classifying normal brain images from the brain tumor images. Materials and methods: The novel MIDNet-18 CNN architecture comprises 14 convolutional layers, 7 pooling layers, 4 dense layers and 1 classification layer. The dataset used for this study has two classes: Normal Brain MR Images and Brain Tumor MR Images. This binary MRI brain dataset consists of 2918 images as training set, 1458 images as validation set and 212 images as test set. Independent sample size calculated was 7 for each group, keeping GPower at 80%. Result: From the experimental results, the proposed MIDNet18 model obtained 98.7% accuracy. Whereas, the VGG16 model obtained an accuracy of 50%. Hence, the performance of the proposed MIDNet18 model achieved is better than VGG16. Conclusion: The proposed model is proved to be statistically significant with p value <0.001 (Independent sample t-test) than the existing model VGG16.