A Medical Image Classification Model based on Quantum-Inspired Genetic Algorithm

Abstract

This study used a Quantum-Inspired Genetic Algorithm (QIGA) to select the proper functionality and reduce the dimensions, classification time, and computational cost of a learning dataset. QIGA reduces the complexity of solutions and improves the selection of the best features. The application of quantum principles, in particular the unpredictability of quantum chromosomes, which are represented by qubits, can help in investigating a significantly more extensive solution space. QIGA offers a novel approach to feature selection in optimization problems. Using principles from quantum computing, this algorithm aims to enhance the efficiency and effectiveness of the feature selection process to increase performance. This indicates that features of both exploration and exploitation are embodied by QIGA without requiring massive amounts of data. Considerable gains in classification accuracy were achieved compared to traditional methods. The dynamic design of the models through the evolutionary mechanism in QIGA enables the optimization process to adapt to varying probabilities produced from the qubit overlay via the quantum rotation gate. This is contrary to traditional methods. The model using QIGA offered a more precise classification than the model optimized by Genetic Algorithms (GA). The proposed method achieved superior performance in terms of classification accuracy, with a score of more than 98%, compared to GA, which achieved a classification accuracy of 94%.