Adaptive Band Limit Estimation based PPG data compression for portable home monitors. (February 2019)

• Record Type: Journal Article
• Authors: Sadhukhan, Deboleena; Pal, Saurabh; Mitra, Madhuchhanda
• Is Part Of: Measurement. Volume 134(2019)
• Page Start: 153
• Page End: 165
• Publication Date: 2019-02
• Subjects: Band-limit -- Compression -- Discrete Fourier transform -- Encoding -- Heart rate -- Photoplethysmography -- Spectral analysis -- Quantization; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2018.10.061
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: PPG data compression algorithm is proposed for portable home monitoring devices. The signal redundancy in frequency domain is fully utilized. The useful band limit is estimated adaptively based on signal characteristics. The corresponding DFT coefficients are adaptively quantized and optimally encoded. Provides optimal compression gain and significant noise reduction ability. Abstract: Continuous long term monitoring of Photoplethysmogram (PPG) in home monitoring applications necessitates the need for data compression to reduce the storage requirement. Limited PPG data compression techniques are available which have fully exploited the signal redundancies. This paper presents a PPG data compression technique based on adaptive estimation of the useful band-limit from the individual signal characteristic. Each data block is then represented using the corresponding discrete Fourier transform coefficients, which are encoded by adaptive quantization and optimal bit allocation to further enhance the compression performance. The algorithm achieves an overall compression ratio of 35.95 with signal distortion maintained below 4%, as tested with 16 bit quantized PPG data sampled at 125 Hz. The algorithm is also implemented and validated on a microcontroller based Arduino board. High compression performance, implementation simplicity and noise robustness justifies the potentiality of the algorithm for its use with portable home monitoring devices.