A Novel Hybrid CNN–Quantum Neural Network Framework with Quantum Acceleration and Error Correction for High-Precision Breast Cancer Classification on AI Edge Devices

Breast cancer remains a critical global health challenge, underscoring the need for early and precise diagnostic methodologies. This paper proposes a novel hybrid deep learning framework that integrates Quantum Neural Networks (QNNs) with the ResNet152 architecture and Transfer Learning to enhance breast cancer classification from medical images. Leveraging the feature extraction power of a pre-trained ResNet152 model, the system encodes refined feature representations into a parameterized quantum circuit utilizing Hadamard, CNOT, and Ry gates. Quantum error correction strategies such as Shor’s Code and surface code are incorporated to preserve quantum state fidelity and reduce decoherence-induced errors, thereby enhancing model robustness. Furthermore, quantum transfer learning accelerates training convergence and enriches pattern recognition capability in high-dimensional medical data. The proposed hybrid model achieves a classification accuracy of 97% and a real-time inference rate of 0.032 frames per second when deployed on an AI edge device, validating its efficacy for point-of-care diagnostics. This study underscores the transformative potential of synergizing quantum computing, edge AI, and deep learning to deliver scalable, interpretable, and high-precision diagnostic tools for clinical applications. The promising results open new avenues for integrating quantum technologies into conventional neural architectures to address the challenges of medical imaging and diagnostic intelligence.