An XAI-enhanced efficientNetB0 framework for precision brain tumor detection in MRI imaging.

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Machine Learning for Diagnostics

Background:

• Accurate brain tumor diagnosis from MRI is vital for treatment planning.
• Traditional diagnosis relies on radiologist expertise, but AI offers improved accuracy.
• Lack of AI transparency hinders clinical adoption.

Purpose of the Study:

• To develop an AI model for accurate and interpretable brain tumor classification.
• To integrate explainable AI techniques with deep learning models for medical image analysis.
• To enhance diagnostic reliability and facilitate clinical adoption of AI in neuro-oncology.

Main Methods:

• Utilized the EfficientNetB0 deep learning architecture for brain tumor classification.
• Integrated explainable AI (XAI) techniques, specifically Grad-CAM visualization.
• Applied the model to classify MRI scans across four categories: Glioma, Meningioma, No Tumor, and Pituitary.

Main Results:

• Achieved a high classification accuracy of 98.72% for brain tumors.
• Demonstrated precision and recall rates exceeding 97% across all tumor categories.
• Validated XAI through Grad-CAM heatmaps, showing alignment with diagnostic markers.

Conclusions:

• The AI-enhanced EfficientNetB0 framework with XAI significantly improves brain tumor classification accuracy.
• The model provides clear visual insights, increasing diagnostic reliability and trust.
• This approach shows substantial potential for clinical adoption in medical diagnostics.