Advancements in VLSI implementation of lossless ECG signal compression schemes

This paper analyzes various prediction methods of ECG signal compression and its hardware implementation. The comparative study of various prediction techniques of ECG compression both for single channel and multichannel ECG signals demonstrate significant compression ratios while maintaining acceptable reconstruction quality and low computational overhead. The paper examines recent progress in VLSI architectures customized for integrating prediction-based lossless ECG compression techniques. Various techniques like adaptive linear prediction, Golomb rice coding and Huffman coding are analyzed for ECG signals from different database. After the comparative study, it shows that the lossless ECG signal compression technique with Golomb rice coding has improved the compression ratio. For ECG signals from MIT-BIH database, with Golomb rice coding compression ratio has become 2.89. For multi-channel ECG signals from PTB database, Compression ratio has become 4.07 using Golomb rice coding. For PTB database ECG signals, the gate count is 476K and for single channel ECG signals its 3.1K by using Golomb rice coding.