A lifting-based 2-D discrete wavelet transform architecture for data compression of bio-potential signals

• Published in: 2015 IEEE 11th International Conference on ASIC (ASICON) pp. 1 - 4
• Main Authors: Yi Ren, Jun Han, Zhiyi Yu, Sizhong Xuan, Xiaoyang Zeng
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2015
• Subjects: Clocks; Computer architecture; Data compression; Discrete wavelet transforms; Energy efficiency; Hardware
• ISBN: 9781479984831; 1479984833
• ISSN: 2162-755X
• DOI: 10.1109/ASICON.2015.7517199

This paper describes an energy efficient architecture for the lifting-based two-dimensional discrete wavelet transform with Z-scanning method, applying to data compression of bio-potential signals. The architecture makes full use of the data relevance between row- and column-wise coefficients to perform the two wises' transforms simultaneously in one single processor with nearly full hardware utilization. Thus the architecture outperforms many existing ones in energy efficiency, latency, area and hardware utilization, that it is very suitable for data compression of bio-potential signals such as electrocardiogram signal in a wireless sensor system where real-time requirements, area and power consumption are key points under consideration. The proposed architecture is synthesized by SMIC 65nm technology for ASIC implementation and consumes only 19 μW power at 1 kHz for a 4-channel node.