Quantum classifier for recognition and identification of leaf profile features

Quantum-based classifiers and architecture are gaining lots of attention in image representation and cryptography. The proposed algorithm applies a quantum classifier to a computer vision system for leaf recognition which can be applied to a quantum computer. Images from ten species of leaves which are categorised into two groups, namely simple and palmately, are recognised using a quantum classifier. The pixels of images are transformed to qubit states using quantum Fourier transform (QFT) and Hadamard gates. The profile and structural features are extracted by applying 1D-convolution and controlled not (CNOT) gates. A quantum nearest neighbour search classifier is used to find the closest matching leaf based on probability. The results for different levels of image processing are evaluated and compared with the nearest neighbour classifier. The recognition rate of the quantum classifier for the best level of image processing is 97.33%. The recognition rate of the classifier is better than the nearest neighbour classifier and also has a low computation time. Graphical abstract