Extraction of Vital Sign Signals from Wearable Devices Using Wavelet Hard Threshold Denoising
Tong Wu, Mengyu Miao, Wangbo Wanyan, Le Wang, Chao Wang · 2024
This paper investigates the application of wavelet hard thresholding for denoising wearable vital sign signals. Through an analysis of respiratory signal data from both male and female subjects, the wavelet hard threshold method demonstrates significant advantages in removing multi-frequency noise and improving signal quality. Experimental results show that the hard thresholding technique not only effectively separates noise and significantly enhances the signal-to-noise ratio (SNR) but also retains key signal features, particularly in preserving periodic characteristics and detailed information. Compared to soft thresholding and fixed thresholding methods, hard thresholding offers superior noise suppression while better maintaining signal integrity. The wavelet hard threshold denoising method provides a reliable solution for processing physiological signals in complex environments, with promising applications in wearable health monitoring devices.