RETRACTED ARTICLE: Quantum mesh neural network model in precise image diagnosing

This study introduces QMNN, a quantum mesh neural network model incorporating a Quantum Neural Network Expectation (QNNE) circuit, and evaluates its performance on MNIST and CIFAR-10 datasets. QMNN employs innovative quantum mesh architecture with increased random layers, while the QNNE circuit, inspired by Circuit 15, integrates extra Hadamard gates for enhanced feature extraction. Implemented using PennyLane libraries, the model demonstrates superior accuracy in training and testing on both datasets compared to classical neural networks, emphasizing its effectiveness in image classification. While classic methods like CNNs may achieve higher accuracy, the significance lies in QMNN’s ability, along with the QNNE circuit, to improve accuracy without introducing additional optimization parameters. The observed quantum advantage suggests that strong entanglement between qubits may enhance pixel correlations, highlighting the potential of quantum-inspired models in advancing image diagnostics. Conclusively, this study provides a stepping stone for further research into the synergy between quantum computing and neural networks, showcasing the promising role of QMNN in augmenting image classification accuracy. The results indicate notable improvements, with MNIST testing accuracy reaching 92.80% and CIFAR-10 testing accuracy achieving 36.90%, demonstrating the efficacy of the proposed model.