Enhanced brain tumour prediction using quantum: a hybrid deep learning approach.

Area of Science:

• Medical Imaging
• Quantum Computing
• Artificial Intelligence

Background:

• Brain tumors require timely and accurate diagnosis for effective treatment.
• Deep learning models like ResNet show promise but face challenges with computational complexity and high-dimensional data.
• Existing methods struggle with scalability and efficiency in processing complex medical imaging data.

Purpose of the Study:

• To develop a hybrid quantum-classical model for enhanced brain tumor diagnosis.
• To leverage quantum transfer learning to overcome limitations of traditional deep learning in medical imaging.
• To improve diagnostic accuracy and computational efficiency in analyzing MRI scans.

Main Methods:

• A hybrid approach combining ResNet for feature extraction and quantum circuits for classification was developed.
• Quantum transfer learning was employed using the PennyLane framework for hybrid quantum-classical integration.
• The model was trained and evaluated on MRI scans for brain tumor detection.

Main Results:

• The hybrid model achieved 95% diagnostic accuracy, a significant improvement over traditional CNN models (82%).
• The proposed method demonstrated superior performance and generalization compared to conventional deep learning approaches.
• The model minimized computational overhead while maximizing diagnostic accuracy.

Conclusions:

• Quantum computing offers a transformative potential for radiology and medical imaging.
• Hybrid quantum-classical models represent a significant advancement in medical diagnostics.
• This approach paves the way for faster, more precise, and scalable diagnostic tools, improving patient outcomes.