ECG Data Compression using Adaptive Beat Subtraction Method

This paper proposes an adaptive beat subtraction method for ECG compression scheme, which can be maximized the compression ratio (CR) and minimized the percent root-mean-square difference (PRD). The compression process of this scheme is based on the wavelet domain, combining with the various preprocessing process such as: beat detection, beat normalization and beat difference. The advantage of this scheme is applied the adaptive average beat subtraction before introduce to the scalar quantization process and Huffman coding process. Since, ECG signal is generally composed of a number of beats, repeated at fairly regular intervals. The Huffman code should be able to increase the compression ratio by using adaptive average beat subtraction method. This method is operated the subtraction between an adaptive average beat with individual normalized beat intervals. The experimental results show the compression performance of the proposed method refer to the MIT/BIH arrhythmia database and the record 114, 222, and 234 have been employed as input data. The obtained of compression ratio is approximately 4 to 15, and the percent root-mean-square difference is 0.05% to 1.5%.