A sparse decomposition approach to compressing biomedical signals

Abstract

An approach of compressing biomedical signals was studied in this paper. First of all, we constructed an over-complete dictionary according to characters of compressing signals. Using the orthogonal matching pursuit (OMP) algorithm, sparse decomposition of biomedical signals was performed based on the dictionary. In this work, we used the optimized results of genetic algorithm (GA) as preliminary particles, and the best atoms were found by local search with particle swarm optimization (PSO). With this genetic hybrid particle swarm (GAPSO) approach, the convergence rate (CR) and the root-mean-square error (RMSE) were improved along with less distortion. For MCG signals in mid-length, simulation results showed that the standard error was 2.78%, when the compression ratio was 15%. © 2010 Springer-Verlag.