Title :
Arithmetic coding for ECG data compression
Author :
Passariello, G. ; Gavidia, L. ; Rodríguez, F. ; Condado, J. ; Ruesta, V. ; Roux, C. ; Mora, F.
Author_Institution :
Grupo de Bioingenieria y Biofisia Aplicada, Univ. Simon Bolivar, Caracas, Venezuela
Abstract :
An electrocardiogram (ECG) entropy coding algorithm, arithmetic coding (AC), a lossless technique, has been used in a hybrid data compression arrangement where lossy techniques, turning point (TP) followed by the fan method, have previously been used. The evaluation was performed using the MIT/BIH arrhythmia database (360 Hz of sampling frequency and truncated to 8-bit resolution). An average compression rate of 10.88:1 (7.71:1 for the combination of TP and FAN, and 1.41:1 for AC) with an average percentage RMS difference (APRD) of 1.21% for normal and arrhythmic recordings was obtained. A clinical evaluation of the algorithm was performed using two groups of specialists, obtaining a 99% diagnostic sensitivity which shows that any distortion introduced by the ECG data compression procedure can be considered clinically negligible
Keywords :
data compression; electrocardiography; 360 Hz; ECG data compression; MIT/BIH arrhythmia database; algorithm; arithmetic coding; compression rate; diagnostic sensitivity; fan method; hybrid data compression arrangement; lossless technique; lossy techniques; sampling frequency; turning point; Arithmetic; Data compression; Databases; Distortion; Electrocardiography; Entropy coding; Image coding; Performance evaluation; Sampling methods; Turning;
Conference_Titel :
Computers in Cardiology 1991, Proceedings.
Conference_Location :
Venice
Print_ISBN :
0-8186-2485-X
DOI :
10.1109/CIC.1991.168981