Title :
An investigation of adaptive beat estimation for improved data compression
Author :
Hamilton, Patrick S.
Author_Institution :
Lafayette Coll., Easton, PA, USA
Abstract :
The authors have investigated the use of adaptive beat estimation for ECG data compression. On selected MIT/BIH database records, they estimated QRS complexes from previous QRS complexes with simple averaging, LMS adaptive estimation, and averaging based on the mean axis direction of the QRS complex. The authors subtracted beat estimates from actual beats to produce residual beats. They calculated entropy of the first differenced residual data. Differenced residual signal entropy was reduced by adaptive estimation in thirteen of fourteen records, with a maximum reduction of 0.53%. Averaging based on mean QRS axis reduced the differenced residual signal entropy in 11 of 13 records, with a maximum entropy reduction of 4.1%
Keywords :
electrocardiography; ECG data compression; MIT/BIH database records; QRS complexes; adaptive beat estimation; differenced residual signal entropy; improved data compression; least mean squares adaptive estimation; mean QRS axis; residual beats; simple averaging; Adaptive algorithm; Adaptive estimation; Code standards; Data compression; Databases; Educational institutions; Electrocardiography; Entropy; Equations; Least squares approximation;
Conference_Titel :
Engineering in Medicine and Biology Society, 1994. Engineering Advances: New Opportunities for Biomedical Engineers. Proceedings of the 16th Annual International Conference of the IEEE
Conference_Location :
Baltimore, MD
Print_ISBN :
0-7803-2050-6
DOI :
10.1109/IEMBS.1994.415426
