ECG Data Compression by Independent Component Analysis

We propose a technique for compression of ECG signals. In the literature, many techniques were proposed for that end, including the use of Karhunen-Loeve transform (KLT) and wavelets. Here we propose to use the concept of redundancy reduction through independent component analysis (ICA) to find a more efficient code. Results showed that using ICA yields far smaller reconstruction errors when comparing the ICA results to the KLT ones