Related Experiment Video
Updated: May 22, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.6K
Learning Architecture for Brain Tumor Classification Based on Deep Convolutional Neural Network: Classic and
Rabei Raad Ali1, Noorayisahbe Mohd Yaacob2, Marwan Harb Alqaryouti3
1Technical Engineering College for Computer and AI, Northern Technical University, Mosul 41000, Iraq.
Diagnostics (Basel, Switzerland)
|March 13, 2025
Summary
Convolutional Neural Networks (CNNs) accurately classify brain tumors in MRI scans. The ResNet50 architecture achieved 99.88% accuracy, significantly aiding diagnosis and treatment planning.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Accurate brain tumor classification from medical images is crucial for patient survival.
- Convolutional Neural Networks (CNNs) are investigated for enhancing diagnostic accuracy.
- Magnetic Resonance Imaging (MRI) datasets are utilized for this classification task.
Purpose of the Study:
- To evaluate the effectiveness of CNNs, specifically Classic layer and ResNet50 architectures, for brain tumor classification.
- To compare the performance of different CNN models on a brain tumor MRI dataset.
- To demonstrate the potential of advanced CNN architectures in improving diagnostic accuracy.
Main Methods:
- A dataset of brain tumor MRI scans (200 × 200 × 1 resolution) was pre-processed.
- Tumors were categorized into Glioma, Meningioma, and Pituitary types.
- Classic layer and ResNet50 CNN architectures were trained and tested using an 80:20 split.
Main Results:
- Both CNN architectures demonstrated high accuracy in classifying brain tumors.
- The Classic layer architecture achieved 94.55% accuracy.
- The ResNet50 architecture achieved a superior accuracy of 99.88%, outperforming previous studies.
Conclusions:
- CNNs, particularly the ResNet50 architecture, are highly effective for brain tumor classification using MRI data.
- These findings highlight the potential of CNNs to assist medical professionals in accurate diagnosis and treatment planning.
- Future research should explore transfer learning to enhance performance, especially with limited annotated data.
