A 12.9-µW 2-D ECG compression core based on NLSPIHT
Gaowei Xu, Lingyun Zeng, Yao Zou, Jun Shu Han, Xiaoyang Zeng · 2014
This paper proposed a 2-D electrocardiogram (ECG) compression core based on no list set partitioning in hierarchical trees (NLSPIHT) algorithm. The hardware solution presented here exploits the fact that ECG signals generally reveal both spatial and temporal redundancy. In order to reduce hardware cost, tiling approach is adopted to implement efficient low-cost circuits. Evaluated in SMIC 65-nm process, the 7 kHz compressor with 64×64 tile draws only 12.9-μW, which is more favorably in portable battery-driven biomedical devices. In addition, the compression performance is maintained compared to similar 2-D compression algorithm in software.