Non-contact Video Heart Rate Detection Method Based on Multivariate Empirical Mode Decomposition and Joint Blind Source Separation

Abstract

The invention discloses a non-contact video heart rate detection method based on multivariate empirical mode decomposition and joint blind source separation. 2. Each sub-region selects the green channel reference signal or color difference signal as the input signal; 3. Uses multivariate empirical mode decomposition to process the input signal to obtain the input signal eigenmode component data set; 4. Adopts joint blind source separation Process the eigenmode component data set of the input signal to obtain the source signal matrix, and filter out the pulse signal from it; 5. Use the method of frequency spectrum analysis to extract the heart rate from the pulse signal. The invention can obtain video heart rate detection results robustly and accurately, and has important application prospects in daily medical care.

Description

technical field [0001] The invention belongs to the technical field of biomedical signal processing, and in particular relates to a method for non-contact extraction of human heart rate from video based on multiple empirical mode decomposition combined with joint blind source separation. Background technique [0002] Heart rate is an important physiological parameter of the human body, and its long-term monitoring is of great significance to the early prevention and prognosis diagnosis of cardiovascular diseases, as well as the monitoring of human health. At present, human heart rate monitoring methods are mainly divided into contact monitoring methods and non-contact monitoring methods according to whether they are in contact with human skin or not. The measurement results of the contact method are accurate and reliable, and have a high degree of social acceptance. However, long-term contact with the skin is likely to cause physical discomfort, and it is not suitable for he...

AI Technical Summary

Problems solved by technology

However, the above two methods also have some disadvantages

AliAl-Naji's method only considers the results of a single region of interest, and does not consider the correlation of the corresponding human heart rate between different facial regions, so compared to the results of multi-regional joint analysis of heart rate, the heart rate value is estimated for a single region More prone to outliers and less robust

Although Qi Huan's method takes into account the advantages of joint extraction of heart rate in multiple regions, it uses the RGB raw data corresponding to each region and lacks preprocessing for denoising the input signal, so that the potential source components extracted from different regions are still It is possible to alias common noise source components, so the heart rate extracted by this method is prone to large errors, making the heart rate estimation result inaccurate

Images