Deep Convolutional Neural Network for Automated Medical Image Diagnosis Using MRI Dataset

Read full text Summarise with VIRAT
Authors: Prabhakar B, Venkat Swamy G
Cited by: 0
References: 15

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The paper discusses an extensive analysis of deep convolutional neural networks (CNNs) for brain tumor diagnosis using MRI scans. The study aims to investigate the performance of both pre-trained and designed CNNs for brain tumor diagnosis based on MRI scans for multi-class brain tumor classification. The study also aims to evaluate the performance of deep CNNs for brain tumor diagnosis using the BRISC dataset, which contains 6,000 contrast-enhanced T1-weighted MRI scans for brain tumor diagnosis. The study proves that deep convolutional neural networks are capable of achieving the best results for brain tumor diagnosis. The study also proves that the designed CNN model using MRI scans for brain tumor diagnosis achieves better results compared to pre-trained models like ResNet50, VGG16, and Xception. The study also proves that the designed CNN model achieves better results compared to recent studies based on brain tumor diagnosis using MRI scans. The designed CNN model achieves an average ROC AUC of 0.99 and an accuracy of 94%. The study also proves that lightweight models like EfficientNet-b0 (accuracy: 98.36%, parameters: 4.01M) and Tiny-ViT-5M (accuracy: 98.41%, parameters: 5.07M) achieve better results for brain tumor diagnosis. The results clearly show that deep CNNs can perform at a diagnostic level comparable to experts while providing scalability, consistency, and computational efficiency for computer-aided diagnosis systems.

Fields of Study & Subjects

Brain Tumor Detection and Classification AI in cancer detection Glioma Diagnosis and Treatment

Publication details

Journal / source
International Journal of Artificial Intelligence Data Science and Machine Learning
Volume
7
Pages
363–372
Publication date
1 January 2026
Language
EN