The set partitioning in hierarchical trees algorithm for data compression in ambulatory electroencephalogram systems

The set partitioning in hierarchical trees (SPIHT) algorithm has achieved notable success in still image coding. In this paper the SPIHT algorithm is applied for the compression of EEG data in single channel and multiple channels. For single channel data, the SPIHT algorithm in one dimension is used with compression ratio ranging from 1.17 to 4.55 for 2 order scale wavelet transform, 1.5 to 8.87 for 3 order scale wavelet transform, and 1.66 to 15.9 for 4 order scale wavelet transform, respectively. For multiple channels data, the two dimensional SPIHT algorithm is used with the compression ratio ranging from 2.72 to 13.08 for 1 order scale wavelet transform, 2.76 to 12.59 for 2 order scale wavelet transform and 2.80 to 12.37 for 3 order scale wavelet transform, respectively. Experimental data is from the people's hospital of Beijing university of normal people brain electrical data, the experiment results of compression binary codes flow and compression CR and PRD parameters are achieved in different wavelet scales, and the experiment results are analyzed and compared. It shows that experiment results of specific compression ratio can be achieved by using the SPIHT Algorithm, at the same time, the scale of Wavelet transform before compression transform affects subsequent algorithm compression, under the experiment data, the bigger scale of Wavelet is transformed, the better result of compression is got. As the same time using the SPIHT algorithm in two dimensions, the result is better than using the SPIHT algorithm in one dimension.