ECG data compression for ambulatory device

Even if highly integrated CMOS technology makes portable devices with high-capacity RAM memory feasible, storage of 24-h 2-lead ECG data requires considerable real-time compression of data. The authors present a method (turning point with extremes) for compression of data and compare it with several of the most frequency used methods (the scan-along polygonal approximation, the sample skipping method and turning-point recursive improvement). Performance is evaluated by computing the percentile root-mean-square difference, which is based on the amplitude of the difference between the original and reconstructed samples, and by evaluating computation time. All the algorithms demonstrated good fidelity, even for high compression ratio. However, turning point with extremes is nearly four times faster than the other methods