Multi-Class brain normality and abnormality diagnosis using modified Faster R-CNN.

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Neurological Disorder Diagnosis

Background:

• Brain disorder detection is crucial for timely treatment and preventing permanent damage.
• Manual classification of medical images is labor-intensive, time-consuming, and error-prone.
• Developing automated systems is essential for efficient and accurate diagnosis.

Purpose of the Study:

• To develop and evaluate a novel ResNet50 modified Faster Regions with Convolutional Neural Network (R-CNN) model for brain normality and abnormality classification.
• To accurately detect and classify brain conditions including hemorrhage and hydrocephalus.
• To improve the accuracy and efficiency of brain disorder diagnosis using medical imaging.

Main Methods:

• Comparison of various Machine Learning (ML) and Deep Learning (DL) models, including Artificial Neural Network (ANN), Logistic Regression (LR), Support Vector Machine (SVM), and Convolutional Neural Network (CNN) variants.
• Implementation of a novel ResNet50 modified Faster R-CNN model for image classification.
• Validation of all methods on a novel dataset of normal and abnormal brain images.

Main Results:

• The proposed ResNet50 modified Faster R-CNN model achieved a classification accuracy of 99.75%.
• Logistic Regression (LR) yielded the highest accuracy among ML models at 84.80%.
• DenseNet201 achieved the highest accuracy among CNN models at 85.68%.

Conclusions:

• The proposed AI model demonstrates superior performance in detecting and classifying brain disorders compared to existing ML and DL methods.
• The developed framework can serve as a valuable computer-aided medical decision support system for healthcare professionals.
• Automated analysis of medical images using advanced AI significantly enhances diagnostic accuracy and efficiency.