A robust signal preprocessing framework for wrist pulse analysis. (January 2016)

• Record Type: Journal Article
• Authors: Wang, Dimin; Zhang, David; Lu, Guangming
• Is Part Of: Biomedical signal processing and control. Volume 23(2016)
• Page Start: 62
• Page End: 75
• Publication Date: 2016-01
• Subjects: Wrist pulse preprocessing -- Wavelet-based decomposition -- Intra-class distance -- Period segmentation; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2015.08.002
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: We provide a robust and effective preprocessing framework for wrist pulse analysis. We introduce a cascade filter based on frequency-dependent analysis for de-noising. The first derivation strategy has the best performance in pulse period segmentation. We propose an adaptive thresholding method to eliminate the outliers. The wrist pulse analysis depends heavily on the pulse preprocessing methods. Abstract: Wrist pulse has been a physical health indicator in Traditional Chinese Medicine (TCM) for a long history. With the development of sensor technology and bioinformatics, quantifying pulse diagnosis by using signal processing technology is attracting increasing attentions in recent years. Since wrist pulse signals collected by the sensors are often corrupted by artifacts in real situations, many approaches on the wrist pulse preprocessing including pulse de-noising and baseline wander removal are introduced for more accurate wrist pulse analysis. However, these scattered methods are incomplete with some limitations when used to preprocess our special pulse data for the clinical applications. This paper presents a robust signal preprocessing framework for wrist pulse analysis. The cascade filter based on frequency-dependent analysis (FDA) is first introduced to remove the high frequency noises and to select the significant pulse intervals. Then the curve fitting method is developed to adjust the direction and the baseline drift with minimum signal distortion. Last, the period segmentation and pulse normalization is applied for the feature extraction. The effectiveness of the proposed pulse preprocessing is validated through experiments on actual pulse records with biochemical markers. In contrast with the traditional methods, the proposed preprocessing framework is effective in extracting more accurate pulse features. And the highest classification rate 91.6% is obtained on diabetes diagnosis. The results demonstrate that our method is superior to the former pulse preprocessing researches and practical for wrist pulse analysis.