Towards Quantum-inspired Evolutionary Algorithms for Denoising PCG Signals

This document aims to reaffirm the superiority of quantum-inspired evolutionary algorithms(QiEA) over conventional algorithms. The study proposes a novel QiEA with sorted mutation and greedy local restart (SMGLR) algorithm for finding the optimal coefficients of a filter using a multi-objective optimization function. PCG signals sourced from the Washington Hearts Sounds and Murmurs database and the Michigan Heart Sounds database were filtered using this enhanced filter. Comparative analysis was conducted against the proposed algorithm and established techniques such as Wavelet-based denoising and Variable Mode decomposition. Results: The results revealed that the filter designed using QiEA-SMGLR exhibited superior performance in terms of improved Signal-to-Noise Ratio (SNR) in both comparisons. A comprehensive statistical analysis yielded a correlation coefficient of close to 0.99 indicating strong signal integrity. The QiEA-SMGLR algorithm outperforms Wavelet-based and VMD methods by improving the output SNR by 20% to 100%. This variance can be attributed to the intensity of noise introduced. The result reaffirms the hypothesis that the QiEA-SMGLR algorithm is superior in improving signal clarity and precision over traditionally popular techniques. The improvement in the analysis of Phonocardiogram (PCG) signals could help improve patient diagnostics, amidst the fight against global cardiac problems.