Coding Categories Based Electrocardiogram (ECG) Lossy Compression Scheme for IoT Systems

Abstract

Nowadays, IoT is widely used for intelligent and distant monitoring. The IoT performances are mainly based on the wireless communication networks. This is the key stone of several applications in the medical applications like e health monitoring, vision and medical imaging. Several operations slow down such a communication systems. The most important one is the compression and decompression blocks. The paper presents a new ECG signal compressor/ decompressor. Low complexity and high accuracy are the principal characteristics of the introduced scheme. The proposed scheme is coding categories based. Low coding category and high coding category and a new frame format are defined. The new frame composition allows reaching high compression ratios. Tests are done using the physionet MIT-BIH and the PTB diagnostic databases. Over than 250 signals, with different cardiac pathologies were used for the tests. We reach a maximum compression ratio (CR) of 40 with a PRD of 0,5%. The introduced compressor outperforms the earlier techniques in the state of the art.