A Quantum Machine Learning Algorithm for Hazelnut Variety Recognition

This study explores a hybrid quantum-classical approach to image classification, applied to hazelnut variety recognition. Using a public dataset, we develop a variational quantum circuit within a hybrid architecture and compare its performance to a classical convolutional neural network. The quantum model is trained using two optimization strategies: ADAM and a genetic algorithm. We analyze their impact on accuracy, convergence, and quantum resource usage. The results highlight the strengths and limitations of quantum-enhanced learning, with gradient-free methods showing particular promise. This work contributes to assessing the practical viability of quantum machine learning for real-world image classification.