Redundancy Reduction and Adaptive Bit Length Encoding-Based Purely Lossless ECG Compression

In the proposed work, the redundancy reduction-based two-dimensional (2D) electrocardiogram (ECG) data compression technique is performed wherein the inter-beat and intra-beat ECG samples were exploited to reduce their amplitudes. The resultant samples showed significant reduction in amplitudes out of which few variations with relatively high magnitudes were observed. Implementation of fixed length coding schemes degrades the compression efficiency. Whereas the existing variable-length encoding schemes failed to provide variable codelengths to the sample amplitudes. In the proposed work, a new Adaptive Bit Length Encoding technique was designed that adaptively selected the number of binary bits required to encode the elements in a block, thus reducing the data size efficiently. The proposed approach ensured lossless ECG compression with guaranteed signal reconstruction. The experiments were performed on all 48 records of the standard MIT-BIH arrhythmia database of 5 min duration. The 100% reconstruction of the compressed signal resulted in zero Percentage Residual Difference (PRD) and Wavelet Energy-based Diagnostic Distortion (WEDD) while the Compression Ratio (CR) comes out to be 3.14 when measured on all records of arrhythmia database which depicted the proficiency of the proposed technique as compared to other state of the art techniques.