Dictionary based coding for ECG data compression

ECG data compression is usually performed in two steps: (a) modelling of the signal and (b) coding of the data that result from the model in (a). A dictionary based coding scheme is proposed that can efficiently handle the arbitrary valued data that result from the modelling phase. The majority of the compression schemes described in the literature employ Huffman and more rarely arithmetic encoders, adopted from text compression, that were not designed to handle data exceeding the range covered by an 8-bit quantity. The coding scheme proposed can handle 16-bit symbols and furthermore is faster and exhibits better compression ratio than 0-order Huffman encoders.<<ETX>>