Decomposition and compression for ECG and EEG signals with sequence index coding method based on matching pursuit

An efficient compression method is proposed by encoding the sequence index of atoms based on matching pursuit (MP) algorithm with over-complete Gabor dictionary, which has the merit to adjust the compression ratio (CR) according to the practical request with low distortion. It is also combined with genetic algorithm (GA) to reduce the computation complexity. Then, the validity of this method is verified by applying it in the compression of electrocardiography (ECG) and Electroencephalography (EEG) signals. The simulation results show that the CR can be achieved at 18:1 with only 1.06% and 2.15% reconstruction errors on ECG and EEG signals respectively. It has higher CR and less reconstruction errors compared to that of the traditional methods, and noise suppression effect is also presented.