Brain Magnetic Resonance Imaging Classification Using Deep Learning Architectures with Gender and Age.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Accurate brain tumor diagnosis relies on effective classification of Magnetic Resonance Imaging (MRI) scans.
• Previous research primarily utilized Support Vector Machine (SVM) and AlexNet for binary classification of normal versus abnormal brain MRIs.
• Limitations in existing methods highlight the need for advanced classification techniques.

Purpose of the Study:

• To develop and evaluate deep learning architectures for classifying brain MRI images into normal or abnormal categories.
• To investigate the impact of incorporating demographic attributes like age and gender as higher-level features for enhanced classification accuracy.
• To compare the performance of proposed deep learning models against traditional methods and existing deep learning approaches.

Main Methods:

• Implementation of deep learning architectures including Convolutional Neural Network (CNN)-based techniques, Deep Neural Network (DNN), LeNet, AlexNet, and ResNet.
• Integration of age and gender as additional attributes within the classification models.
• Comparative analysis of proposed models against Support Vector Machine (SVM) and AlexNet using brain MRI datasets.

Main Results:

• The proposed deep learning models demonstrated superior performance compared to SVM and AlexNet in classifying brain MRI images.
• A LeNet-inspired model achieved an overall accuracy of 88%, while a CNN-DNN model reached 80%.
• Best accuracies achieved were 100% (LeNet Inspired Model), 92% (CNN-DNN), 92% (ResNet), and 81% (AlexNet), with SVM at 82%.

Conclusions:

• Deep learning architectures, particularly when incorporating age and gender attributes, offer significant improvements in brain tumor classification accuracy from MRI.
• Age and gender are identified as crucial factors that enhance the precision and clinical relevance of brain tumor analysis.
• The developed deep learning models show promise for more accurate and reliable brain tumor diagnosis, outperforming established methods.