An effective low-complexity multi-vital-signs compression technique for embedded-link e-home healthcare

Transplanting the existing e-home healthcare system to source-limited embedded-link device for home-use health monitoring, intelligent medical diagnosis and wireless transmission is attractive. Yet, constrains of portable storage, computing and transmission promote the need of data compression for such applications. Existing compression techniques are mostly desktop-computer-based and computation-consuming, making them unsuitable for mobile device. To tackle such a bottleneck problem, this paper addresses an effective low-complexity multi-vital-signs compression technique based on orthogonal polynomial decomposition (OPD) algorithm using Hermite functions. The technique is proposed and operated on the designated healthcare system with optimized parameters. Validated and tested with cardiovascular disease (CVD) diagnosis based on sphygmogram both experimentally and clinically, the proposed technique achieves comparable good performance with distortion less than 2% and compression ratio up to 6, and preserves significant pathological features of multi-vital-signs for clinical diagnosis. The proposed technique is highly robust even for freaky and pathological signals. In addition, the compressed results reflecting morphological features can be directly adapted to the subsequent medical analysis without further decompression.