Hybrid of VGG-16 and FTVT-b16 Models to Enhance Brain Tumors Classification Using MRI Images

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Neuro-oncology

Background:

• Accurate brain tumor classification from MRI is crucial but challenging due to tumor variability and artifacts.
• Existing methods face limitations in capturing complex tumor characteristics.

Purpose of the Study:

• To develop a novel hybrid deep learning framework for enhanced brain tumor classification.
• To improve the precision and reliability of AI-assisted diagnostics in neuro-oncology.

Main Methods:

• A hybrid deep learning framework integrating VGG-16 (CNN) and a fine-tuned Vision Transformer (FTVT-b16).
• Utilized two distinct MRI datasets for comprehensive evaluation, including multi-class and binary classifications.
• Addressed limitations of traditional CNNs and pure ViTs.

Main Results:

• Achieved high classification accuracy, reaching 99.46% and 99.90% on the two tested datasets.
• Demonstrated a robust and interpretable solution suitable for clinical integration.
• Highlighted the effectiveness of hybrid architectures in overcoming individual model weaknesses.

Conclusions:

• The proposed hybrid framework shows significant potential for advancing AI-assisted brain tumor diagnosis.
• Future research will focus on multi-institutional validation and computational optimization for clinical scalability.