Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
567178 | Signal Processing | 2007 | 12 Pages |
We suggest a novel time domain lossy compression method for digital signals. The method is composed of two main steps: Short-term redundancy reduction (STRR) and long-term redundancy reduction (LTRR). The STRR step uses a variant of the Douglas–Peucker line simplification algorithm. The LTRR step uses a combined block-sorting and curve simplification technique to reduce long-term redundancy. The proposed method is a general-purpose signals compression scheme, with more efficiency for quasi-periodic signals. In that respect, it has been applied to ECG signals selected from the MIT-BIH database and compared to existing recent time domain (CCSP), wavelet-based and TSVD methods. Results show that the proposed algorithm is in a comfortable competitive position with the most performant methods for moderate and near-lossless compression.