Signal acquisition method and device, storage medium, electronic device and program product

Abstract

The invention discloses a signal acquisition method and device, a storage medium, an electronic device and a program product. The method comprises the following steps: acquiring an initial pulse signal of a target biological object; the initial pulse signal is subjected to convolution processing, multiple pieces of mapping data of the initial pulse signal in a Gaussian space are obtained, and the mapping data are used for representing Gaussian components in the initial pulse signal; performing synthesis processing on the plurality of mapping data to obtain a synthesis result; and determining a synthesis result as a blood volume pulsation signal of the target biological object. The technical problem that the blood volume pulse signal extraction accuracy is low is solved.

Description

Technical Field

[0001] The present invention relates to the fields of biology and signal processing, and in particular to a signal acquisition method, device, storage medium, electronic device and program product. Background Art

[0002] Currently, the measurement of blood flow volume pulse signals is usually carried out through Doppler ultrasound technology and contact photoplethysmography (cPPG) technology. However, in order to accurately capture the changes in blood vessel volume caused by the heart pumping blood, the above two methods require the sensor to be tightly attached to the skin surface of the subject. Although the contact measurement method can provide higher accuracy, it is also easily affected by the selection of measurement position, skin contact pressure and the movement of the subject.

[0003] In the related art, an imaging photoplethysmo graphy (iPPG) based on an image sensor is usually used to measure the blood volume pulse signal in a non-contact manner. However, this method has a technical problem of low accuracy in extracting the blood volume pulse signal.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a signal acquisition method, device, storage medium, electronic device and program product to at least solve the technical problem of low accuracy in extracting blood volume pulse signals.

[0006] According to one aspect of an embodiment of the present invention, a signal acquisition method is provided, which may include: acquiring an initial pulse signal of a target biological object; performing convolution processing on the initial pulse signal to obtain multiple mapping data of the initial pulse signal in a Gaussian space, wherein the mapping data is used to represent a Gaussian component in the initial pulse signal; performing synthesis processing on the multiple mapping data to obtain a synthesis result; and determining the synthesis result as a blood volume pulsation signal of the target biological object.

[0007] Optionally, obtaining an initial pulse signal of the target biological object includes: obtaining a chrominance signal of the target biological object; performing bandpass filtering on the chrominance signal; and converting the chrominance signal after the bandpass filtering to obtain the initial pulse signal.

[0008] Optionally, obtaining the chrominance signal of the target biological object includes: using an image sensor to collect video data associated with a target part of the target biological object; and converting the video data to obtain the chrominance signal.

[0009] Optionally, the initial pulse signal is convolved to obtain multiple mapping data of the initial pulse signal in Gaussian space, including: performing Gaussian mixture fitting on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions; and convolving the initial pulse signal using multiple Gaussian kernel functions to obtain multiple mapping data.

[0010] Optionally, the initial pulse signal is convolved with a plurality of Gaussian kernel functions to obtain a plurality of mapping data, including: the initial pulse signal is convolved with a Gaussian kernel function to obtain mapping data.

[0011] Optionally, Gaussian mixture fitting is performed on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions, including: performing Gaussian mixture fitting on the historical blood volume pulse signal to obtain multiple initial Gaussian kernel functions; and expanding the multiple initial Gaussian kernel functions respectively to obtain multiple Gaussian kernel functions.

[0012] Optionally, synthesizing the multiple mapping data to obtain a synthesis result includes: synthesizing the multiple mapping data using a long short-term memory network to obtain a synthesis result.

[0013] According to another aspect of an embodiment of the present invention, a signal acquisition device is also provided, which may include: an acquisition unit, used to acquire an initial pulse signal of a target biological object; a first processing unit, used to perform convolution processing on the initial pulse signal to obtain multiple mapping data of the initial pulse signal in Gaussian space, wherein the mapping data is used to represent the Gaussian component in the initial pulse signal; a second processing unit, used to perform synthesis processing on the multiple mapping data to obtain a synthesis result; and a determination unit, used to determine the synthesis result as the blood volume pulse signal of the target biological object.

[0014] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, which includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the signal acquisition method of an embodiment of the present invention.

[0015] According to another aspect of an embodiment of the present invention, a processor is further provided. The processor is used to run a program, wherein the program executes the signal acquisition method of an embodiment of the present invention when run by the processor.

[0016] According to another aspect of an embodiment of the present invention, a program product is further provided. The program product includes computer instructions. When the computer instructions are executed by a processor, the signal acquisition method of the embodiment of the present invention is implemented.

[0017] In an embodiment of the present invention, an initial pulse signal of a target biological object is acquired; a convolution process is performed on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, wherein the mapping data is used to represent a Gaussian component in the initial pulse signal; a synthesis process is performed on the plurality of mapping data to obtain a synthesis result; and the synthesis result is determined as a blood volume pulse signal of the target biological object. That is, in an embodiment of the present invention, a convolution process is performed on the acquired initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, a synthesis process is performed on the plurality of mapping data, and a final blood volume pulse signal is determined based on the synthesis result obtained after the synthesis process, thereby avoiding the situation where the extraction accuracy of the blood volume pulse signal from the initial pulse signal is low, thereby achieving the technical effect of improving the extraction accuracy of the blood volume pulse signal, and solving the technical problem of the low extraction accuracy of the blood volume pulse signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0019] Figure 1 is a flow chart of a signal acquisition method according to an embodiment of the present invention;

[0020] Figure 2 is a schematic diagram of combining multi-scale Gaussian mixture filtering and LSTM according to an embodiment of the present invention;

[0021] Figure 3 is a schematic diagram of an LSTM structure according to an embodiment of the present invention;

[0022] Figure 4 is a schematic diagram of a test bench for verifying a pulse wave system according to an embodiment of the present invention;

[0023] Figure 5 is a schematic diagram of a reconstructed waveform according to an embodiment of the present invention;

[0024] Figure 6 is a schematic diagram of a first-order derivative waveform according to an embodiment of the present invention;

[0025] Figure 7 is a schematic diagram of a second-order derivative waveform according to an embodiment of the present invention;

[0026] Figure 8 is a schematic diagram of hemodynamic characteristics according to an embodiment of the present invention;

[0027] Fig. 9 is a schematic diagram of a signal acquisition device according to an embodiment of the present invention;

[0028] Fig.10 is a structural block diagram of a computer terminal according to an embodiment of the present invention;

[0029] Fig.11 The present invention is a block diagram of an electronic device for a signal acquisition method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] According to an embodiment of the present invention, an embodiment of a method for acquiring a signal is provided, and the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and, although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0033] In this embodiment, a signal acquisition method is proposed. The method can perform convolution processing on the acquired initial pulse signal to obtain multiple mapping data of the initial pulse signal in Gaussian space, perform synthesis processing on the multiple mapping data, and determine the final blood volume pulse signal based on the synthesis result obtained after the synthesis processing, thereby avoiding the low extraction accuracy of the blood flow volume pulse signal from the initial pulse signal, thereby achieving the technical effect of improving the extraction accuracy of the blood volume pulse signal and solving the technical problem of low extraction accuracy of the blood volume pulse signal.

[0034] Figure 1FIG. 1 is a flow chart of a signal acquisition method according to an embodiment of the present invention. Figure 1 As shown, the method may include the following steps:

[0035] Step S102, acquiring an initial pulse signal of the target biological object.

[0036] In the technical solution provided in step S102 of the present invention, the initial pulse signal may be an iPPG signal obtained by iPPG of an image sensor, which may be obtained by X t The target biological object may be an object with vital signs such as a human body or an animal. It should be noted that this is only an example and the type of the target biological object is not specifically limited.

[0037] Optionally, in order to facilitate the identification of Gaussian wave components in the iPPG signal to obtain a blood volume pulse signal, a rough iPPG signal (ie, an initial pulse signal) can be first obtained using a traditional method, and then the iPPG signal can be processed to obtain a blood volume pulse signal with higher accuracy.

[0038] For example, a chromatic projection (CHROM) can be used to obtain a preliminary iPPG signal. It should be noted that this is only an example, and other methods can also be used to obtain iPPG signals. The method for obtaining iPPG signals is not specifically limited here.

[0039] Step S104, performing convolution processing on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space, wherein the mapping data is used to represent the Gaussian component in the initial pulse signal.

[0040] In the technical solution provided in the above step S104 of the present invention, the initial pulse signal can be convolved to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space. The above mapping data can be used to represent the Gaussian component in the initial pulse signal, can be a Gaussian mapping, can be used to characterize the Gaussian wave component, can be a Gaussian wave, and can be represented by X. It should be noted that this is only for example, and there is no specific limitation on the form of expression of the mapping data.

[0041] Optionally, a plurality of mapping data may constitute a Gaussian mapping matrix (Gaussian mappings), which may be referred to as a mapping matrix for short, and may be a matrix that retains Gaussian components in a rough iPPG signal and suppresses interference of non-Gaussian shapes.

[0042] Optionally, an initial pulse signal is obtained, and a Gaussian kernel function is used to match the initial pulse signal to obtain potential mapping data in the iPPG signal.

[0043] Optionally, the above-mentioned convolution processing can be Gaussian kernel convolution processing or other similar function kernel processing. It should be noted that there is no specific limitation on the type of convolution processing here. As long as the method of obtaining mapping data by convolution processing on the initial pulse signal should be within the protection scope of the present invention.

[0044] In this embodiment, the initial pulse signal is convoluted to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space, and the mapping data can be used to characterize the Gaussian component in the initial pulse signal. By convoluting the initial pulse signal, the Gaussian component in the iPPG signal is retained, and the interference of the non-Gaussian shape in the iPPG signal is suppressed, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0045] Step S106, performing synthesis processing on the plurality of mapping data to obtain a synthesis result.

[0046] In the technical solution provided in the above step S106 of the present invention, an initial pulse signal is obtained, which may be a rough iPPG signal. By performing convolution processing on the initial pulse signal, a plurality of mapping data are obtained, and the plurality of mapping data may be synthesized to obtain a synthesis result.

[0047] Step S108, determining the blood volume pulse signal of the target biological object based on the synthesis result.

[0048] In the technical solution provided in the above step S108 of the present invention, the above blood volume pulsation (Blood Volume Pulse, referred to as BVP) signal can be a physiological signal representing the effect of heart beat on skin microvascular capacity, can be used to characterize arterial blood volume pulse, can be used to reflect the ability of the heart of the target biological object to pump blood, that is, can be used to reflect how the heart transports blood from the heart to various tissues and organs throughout the body, and the above blood volume pulse signal can also be called a light volume pulse wave signal.

[0049] Optionally, multiple mapping data are synthesized to obtain a synthesis result, which can be a BVP signal. Further, the above-mentioned synthesis result can be determined as the blood volume pulse signal of the target biological object, thereby achieving the purpose of converting the Gaussian mapping of the initial pulse signal into a one-dimensional BVP signal, thereby achieving the purpose of improving the extraction accuracy of the blood volume pulse signal from the initial pulse signal.

[0050] Optionally, an initial pulse signal is obtained, and a Gaussian mapping matrix is ​​obtained by convolution processing of the initial pulse signal, and the Gaussian mapping matrix is ​​transformed to obtain a one-dimensional BVP signal. The one-dimensional BVP signal can be determined as a blood volume pulse signal. Further, by monitoring the BVP signal, the functional state of the heart and the condition of blood circulation can be determined.

[0051] Through the above-mentioned step S102 and step S108 of the present invention, in an embodiment of the present invention, an initial pulse signal of a target biological object is obtained; a convolution process is performed on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, wherein the mapping data is used to represent a Gaussian component in the initial pulse signal; a plurality of mapping data are synthesized to obtain a synthesis result; and based on the synthesis result, a blood volume pulse signal of the target biological object is determined. That is, in an embodiment of the present invention, a convolution process is performed on the acquired initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, a synthesis process is performed on the plurality of mapping data, and a final blood volume pulse signal is determined based on the synthesis result obtained after the synthesis process, thereby avoiding the situation where the extraction accuracy of the blood volume pulse signal from the initial pulse signal is low, thereby achieving the technical effect of improving the extraction accuracy of the blood volume pulse signal, and solving the technical problem of the low extraction accuracy of the blood volume pulse signal.

[0052] The above method of this embodiment is further introduced below.

[0053] As an optional implementation, step S102, obtaining an initial pulse signal of the target biological object, includes: obtaining a chrominance signal of the target biological object; performing bandpass filtering on the chrominance signal; and converting the chrominance signal after the bandpass filtering to obtain an initial pulse signal.

[0054] In this embodiment, the initial pulse signal of the target biological object can be obtained by the following steps: the chrominance signal of the target biological object can be obtained; the chrominance signal can be bandpass filtered; further, the chrominance signal after the bandpass filtering can be converted to obtain the initial pulse signal. Among them, the above-mentioned chrominance signal can be used to characterize the reflection or absorption characteristics of the skin of the target biological object at different wavelengths, can be used to determine the blood oxygen saturation (SpO2) and heart rate of the target biological object, can be generated by the blood in the skin absorbing and reflecting light of different wavelengths, and can be applied to the monitoring of the initial pulse signal and the analysis of the pulse waveform. It should be noted that this is only an example, and there is no specific limitation on the characterization content of the chrominance signal.

[0055] For example, suppose there is a biosensor that can be used to obtain the chromaticity signal of the target biological object. The chromaticity signal may include the reflection or absorption characteristics of the skin color at different wavelengths, which can be used to indirectly obtain the initial pulse signal. Furthermore, the biosensor is used to obtain the chromaticity signal of the target biological object's skin through an optical sensor or an imaging device, and the color changes over time are recorded. The acquired chromaticity signal can be bandpass filtered to extract the signal components related to the heartbeat frequency. Bandpass filtering can remove irrelevant high-frequency and low-frequency noise and retain the frequency of the heartbeat signal. The chromaticity signal after bandpass filtering is converted into an iPPG signal, which can be an optical signal change caused by skin blood flow and reflects the heartbeat information, and can be used to monitor heart rate, pulse waveform and blood oxygen saturation.

[0056] It should be noted that this is only an example, and there is no specific limitation on the method of obtaining the chromaticity signal and the initial pulse signal.

[0057] As an optional implementation, obtaining a chrominance signal of a target biological object includes: using an image sensor to collect video data associated with a target part of the target biological object; and converting the video data to obtain a chrominance signal.

[0058] In this embodiment, the image sensor can also be used to collect video data associated with the target part of the target biological object; by converting the video data, a chromaticity signal can be obtained. The target part can be any part of the target biological object, such as a palm, an arm, etc. It should be noted that this is only an example, and the specific position of the target part is not limited. The image sensor can be a camera, a camera, etc. It should be noted that this is only an example, and the type of image sensor is not specifically limited.

[0059] Optionally, the BVP waveform is usually acquired by detecting and analyzing light scattering signals of skin microvessels, and thus the target site may be a site containing microvessels, such as an arm.

[0060] In this embodiment, an image sensor is used to obtain an initial pulse signal, and the initial pulse signal is processed to achieve the purpose of accurately extracting the bleeding volume pulse signal from the initial pulse signal, thereby completing the accurate measurement of non-contact pulse based on the image sensor (such as a camera).

[0061] For example, a skin video of a target biological object is obtained to obtain video data. The three channels (RGB) in the video data can be spatially averaged to obtain a 3×N matrix Among them, N can be used to represent the length of time, which can be eliminated by designing a projection matrix (for example, the result projection matrix) by utilizing the difference in reflection under different spectra. The skin surface reflection in the interference and obtain two chromaticity signals S1(t) and S2(t):

[0062]

[0063] Among them, P c M can be used to represent the resulting projection matrix. It can be obtained by converting video data (which can be a video signal) using pixel averaging, and can be a timing signal.

[0064] Optionally, S1(t) and S2(t) can be obtained by projecting the time series signal.

[0065] Furthermore, after obtaining the chrominance signal, the iPPG signal (X t ),Right now:

[0066]

[0067] Wherein, std(·) can be used to represent deviation calculation. S′1(t) and S′2(t) can be used to represent the signals of the chrominance signals S1(t) and S2(t) after bandpass filtering, respectively. Α can be a preset parameter value.

[0068] It should be noted that the above-mentioned method of determining the chromaticity signal and the initial pulse signal is only an example and is not specifically limited here.

[0069] As an optional implementation, step S104, performing convolution processing on the initial pulse signal to obtain multiple mapping data of the initial pulse signal in Gaussian space, including: performing Gaussian mixture fitting on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions; performing convolution processing on the initial pulse signal using multiple Gaussian kernel functions respectively to obtain multiple mapping data.

[0070] In this embodiment, the historical blood volume pulse signal can be pre-fitted with a Gaussian mixture to obtain multiple Gaussian kernel functions. After the initial pulse signal is obtained, the initial pulse signal can be convolved with the Gaussian kernel function to obtain multiple mapping data. The Gaussian kernel function can be a Gaussian kernel characterization parameter, which can be used to map the initial pulse signal of the input space to an infinite-dimensional feature space. The Gaussian kernel function can make the nonlinear problem linearly separable in the high-dimensional space. The historical blood volume pulse signal can be pre-measured using a contact pulse meter.

[0071] Optionally, after obtaining a set of Gaussian kernel functions, the Gaussian kernel functions can be used to match Gaussian components in the iPPG signal. The above matching can be understood as using the Gaussian kernel to slide convolve the iPPG signal. In this process, the Gaussian kernel can match components with similar shapes to the components to convert the iPPG signal into a mapping in the Gaussian space, thereby retaining the Gaussian components in the iPPG signal.

[0072] It should be noted that a Gaussian mixture fitting can be performed on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions, and a Gaussian mixture fitting can also be performed on multiple historical blood volume pulse signals of biological objects of the same type as the target biological object to obtain multiple Gaussian kernel functions. It should be noted that this is only an example, and there is no specific limitation on the method of obtaining the Gaussian kernel function.

[0073] For example, multiple Gaussian kernel functions can be obtained in advance through historical blood volume pulse signals, and the iPPG signal is convolved with the pre-designed Gaussian kernel to obtain the mapping of the iPPG signal in the Gaussian space (that is, mapping data), thereby obtaining the synthesis result:

[0074]

[0075] Where X can be used to represent the N×T mapping matrix. T can be used to represent the time length of the iPPG signal. For the mapping matrix X, it retains the Gaussian component in the rough iPPG signal and suppresses the interference of non-Gaussian shapes. x1, x2, x3, ..., x T Can be used to represent multiple mapping data. can be used to represent Gaussian kernel functions respectively. t Can be used to represent the initial pulse signal.

[0076] It should be noted that the above-mentioned Gaussian kernel function can be determined in advance through experimental calculations and does not change during actual use, or it can be obtained by performing Gaussian mixture fitting on multiple historical blood volume pulse signals. There is no specific restriction on the determination time and determination method of the Gaussian kernel function.

[0077] As an optional implementation, the initial pulse signal is convolved with a plurality of Gaussian kernel functions to obtain a plurality of mapping data, including: the initial pulse signal is convolved with the Gaussian kernel function to obtain mapping data.

[0078] In this embodiment, the initial pulse signal can be convolved with a Gaussian kernel function to obtain mapping data. The Gaussian kernel function can be a multi-scale Gaussian kernel, and multiple Gaussian kernels can be used to match Gaussian components in the iPPG signal, that is, the iPPG signal is convolved with a pre-designed Gaussian kernel function to obtain a mapping of the iPPG signal in the Gaussian space, thereby obtaining mapping data.

[0079] As an optional implementation, Gaussian mixture fitting is performed on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions, including: performing Gaussian mixture fitting on the historical blood volume pulse signal to obtain multiple initial Gaussian kernel functions; and expanding the multiple initial Gaussian kernel functions respectively to obtain multiple Gaussian kernel functions.

[0080] In this embodiment, a set of multi-scale Gaussian kernel functions can be designed in advance, and the Gaussian kernel functions can be used to match the potential BVP Gaussian wave components in the acquired iPPG signal to obtain the final BVP signal.

[0081] Optionally, considering that the heart rate of the target biological subject usually varies between 48 and 200 beats per minute, the duration of the BVP signal in one cardiac cycle is about 0.3 to 1.25 seconds. According to the Gaussian mixture model of the BVP signal, the width of the Gaussian kernel can be limited to the range of 0.1 to 1.25 seconds. Assume that a Gaussian kernel function with a time width of 1 second is defined as:

[0082] G0(t)=G(t,Θ0)·[u(t)-u(t-1)]

[0083] Among them, the Gaussian kernel parameter Θ0 = [1, 0.5, 0.15], which can be the initial Gaussian kernel function, can be obtained by fitting the BVP signal collected by the contact PPG device through the experimental test, depends on the shape of the Gaussian wave in the BVP signal, can be obtained by Gaussian mixture fitting of the historical blood volume pulse signal (that is, cPPG data), or can be determined in advance through testing.

[0084] Optionally, in order to generate a Gaussian kernel function with a required time width between 0.1 and 1.25 seconds, G0(t) can be stretched by 0.8 to 10 times to obtain a set of Gaussian kernel functions:

[0085]

[0086] Wherein, N can be used to represent the number of Gaussian kernel functions, and can be used to determine the scale resolution, for example, N = 128. It should be noted that the above numbers are only for illustration and are not specifically limited here.

[0087] As an optional implementation, step S106, synthesizing the multiple mapping data to obtain a synthesis result, includes: synthesizing the multiple mapping data using a long short-term memory network to obtain a synthesis result.

[0088] In this embodiment, the process of converting the Gaussian mapping matrix into a blood volume pulse signal can be regarded as a regression analysis task. Considering the time series nature of the BVP signal and its inherent causal relationship, this embodiment selects a neural network model based on a long short-term memory (LSTM) unit to perform this task. The LSTM network can effectively capture long-term dependencies. It should be noted that there is no specific limitation on obtaining the blood volume pulse signal here. As long as the method of obtaining the blood volume pulse signal is based on the synthesis result, it should be within the scope of protection of the present invention.

[0089] For example, the network framework used by the neural network model based on long short-term memory units can include a two-layer LSTM structure and a linear projection layer for output conversion. Under this framework, the regression conversion process from the Gaussian mapping matrix to the one-dimensional BVP signal can be formally defined as:

[0090]

[0091] Among them, Y t Can be used to indicate BVP signal. LG(X t ) can be used to represent the LSTM-based BVP generator. t ,i t , C t , o t ,h t They can be used to represent the forget gate, input gate, cell state, output gate and hidden state of LSTM, which together determine the time dependency of LSTM in solving timing signals. f ,W i ,W c ,W o Can be used to represent the weight matrix of LSTM. f ,U i ,U c ,U o Can be used to represent hidden nodes h t-1 The weight matrix of b. f ,b i ,b c ,b oCan be used to represent the bias vector. σ(·) and tanh(·) can be used to represent activation functions (e.g., sigmoid activation function and tanh activation function), respectively. ⊙ can be used to represent a dot product operation. The matrix P can be a linear transformation matrix.

[0092] Optionally, through the above steps, the extraction of the blood flow volume initial pulse signal from the initial pulse signal can be completed.

[0093] It should be noted that, in addition to the long short-term memory network (LSTM), other neural network frameworks can also be used to complete the regression of the Gaussian mapping matrix to the pulse wave signal. There is no specific restriction on the type of neural network model used here.

[0094] At present, in order to improve the extraction accuracy of BVP waveform, a signal dictionary joint reconstruction method is proposed. This method can extract BVP signal from face video. Although this method improves the extraction effect of BVP waveform to a certain extent, it does not fully consider the physiological mechanism of BVP formation, resulting in the extracted BVP waveform is still not accurate enough. There is still a technical problem of low accuracy of blood volume pulse signal extraction.

[0095] Therefore, considering that the BVP signal is essentially an initial waveform generated by the heart pumping blood, and a complex physiological waveform composed of the reflected waves generated when it encounters impedance mismatch points in various parts of the human circulatory system, in order to achieve high-precision recovery of the initial pulse signal of arterial blood volume, in this embodiment, an optical domain non-contact hemodynamic monitoring method is proposed. The method combines multi-scale Gaussian mixed filtering and LSTM, regards the BVP signal as a superposition of a series of basic waveforms, and gives a corresponding mathematical model expression. According to the above mathematical model expression, a specific BVP waveform template can be designed. These mathematical models can represent the typical characteristics of the BVP signal under ideal conditions. By comparing and fitting these templates with the actual collected, noisy iPPG signals, a more accurate BVP signal waveform can be effectively identified and extracted from the complex original data (that is, the initial pulse signal). This method not only improves the signal quality, but also provides a more reliable data source for subsequent hemodynamic analysis, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0096] In an embodiment of the present invention, a convolution process is performed on the acquired initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space, and a synthesis process is performed on the plurality of mapping data. Based on the synthesis result obtained after the synthesis process, a final blood volume pulse signal is obtained, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0097] The technical solution of the embodiment of the present invention is illustrated below in conjunction with preferred implementation modes.

[0098] The measurement of BVP signals is usually carried out through Doppler ultrasound technology and contact photoplethysmography (cPPG) technology. However, in order to accurately capture the changes in blood vessel volume caused by the heart pumping blood, the above two methods require the sensor to be tightly attached to the skin surface of the subject. Although the contact measurement method can provide higher accuracy, it is also easily affected by the selection of measurement position, skin contact pressure and the movement of the subject, especially when the subject is in an active state. These factors will cause significant motion artifacts, which in turn affect the accuracy of the measurement. In addition, in the case of long-term continuous monitoring, contact devices may cause discomfort to the user and may even cause skin allergies and other problems.

[0099] In order to avoid the above problems, non-contact BVP signal measurement technology has gradually become a research hotspot. Compared with the contact method, non-contact measurement not only avoids the discomfort caused by physical contact, but also reduces the impact of motion artifacts. In the related technology, imaging photoplethysmography based on image sensors is usually used to complete the measurement of non-contact BVP signals. In the above method, the iPPG system works by capturing the subtle changes in human skin color caused by BVP signals. This change is due to the difference in the absorption of different wavelengths of light by hemoglobin in the blood. With each beat of the heart, the blood volume under the skin will change periodically, resulting in almost invisible fluctuations in skin color. The iPPG system infers the BVP signal by capturing these fluctuations. Although the iPPG system is simple in principle and easy to implement, the propagation of light in free space is affected by multiple factors such as air scattering, surface reflection, and changes in ambient light, which results in a low signal-to-noise ratio for the iPPG system. Therefore, most existing iPPG systems can only extract a relatively rough waveform of the initial pulse signal for estimating heart rate and heart rate variability, but are insufficient in extracting detailed BVP waveform features, and there is a technical problem of low accuracy in extracting blood volume pulse signals.

[0100] Optionally, a complete BVP waveform may include components such as the main wave, the dicrotic wave, and the descending isthmus. These characteristic waveforms are formed by the contraction and relaxation of the heart and the interaction between the blood and the arterial wall. The morphology of the BVP waveform is of great significance for evaluating the health of the cardiovascular system and can provide important hemodynamic indicators such as pulse width, reflectivity index, and arterial stiffness index. These indicators are important for early detection of heart disease, monitoring changes in cardiac function, and evaluating treatment effects. Therefore, it is of great significance to provide a technology that can accurately extract the BVP waveform from the iPPG signal.

[0101] In the related art, in order to improve the extraction accuracy of BVP waveform, a signal dictionary joint reconstruction method is proposed, which can extract BVP signal from face video; a method using singular spectrum analysis (SSA) to extract pulse wave signals with more details from original iPPG signals is also proposed. Although the above methods have improved the extraction effect of BVP waveform to a certain extent, they have not fully considered the physiological mechanism of BVP formation, resulting in the extracted BVP waveform is still not accurate enough, and there is still a technical problem of low accuracy in blood volume pulse signal extraction.

[0102] As can be seen from the above content, the accuracy of pulse waves measured by most remote pulse wave measuring instruments is limited. Although the existing advanced methods can accurately restore the pulse rate to estimate the heart rate, they can rarely extract pulse waveforms containing details, and the requirements for the measurement environment are relatively high, making it difficult to apply in actual monitoring scenarios. Therefore, considering that the BVP signal is essentially an initial waveform generated by the heart pumping blood, and a complex physiological waveform composed of the reflected waves generated when it encounters impedance mismatch points in various parts of the human circulatory system, in order to achieve high-precision arterial blood volume initial pulse signal recovery, in this embodiment, an optical domain non-contact hemodynamic monitoring method is proposed. This method combines multi-scale Gaussian mixture filtering and LSTM, regards the BVP signal as a superposition of a series of basic waveforms, and gives a corresponding mathematical model expression. According to the above mathematical model expression, a specific BVP waveform template can be designed. These mathematical models can represent the typical characteristics of the BVP signal under ideal conditions. By comparing and fitting these templates with the actual collected noisy iPPG signals, a more accurate BVP signal waveform can be effectively identified and extracted from the complex raw data (i.e., the initial pulse signal). This method not only improves the signal quality, but also provides a more reliable data source for subsequent hemodynamic analysis, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0103] Optionally, this embodiment restores the real pulse waveform from the algorithm level without hardware upgrade of the imaging pulse measurement system, and is sufficient for application in actual monitoring scenarios.

[0104] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments; the size of the numbers in the embodiments is only for illustration and is not specifically limited here. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention and are not specifically limited here.

[0105] In this embodiment, considering that the BVP wave signal is a composite wave composed of a main wave directly generated by the heart and several reflected waves in the arterial vascular system, from a physiological point of view, in one cardiac cycle of the heart, the BVP signal is composed of the main wave generated by the heart pumping blood and the reflected waves formed by various impedance mismatch points encountered during the blood transmission along the artery, and the morphology of these pulse waves is close to Gaussian distribution. Therefore, based on the above physiological principles, this embodiment regards the BVP signal formed in the artery as a superposition of multiple Gaussian waves, thereby proposing a method combining multi-scale Gaussian mixture filtering and LSTM, and using this method to realize the recovery of the real BVP signal from the rough iPPG signal.

[0106] Alternatively, according to the test, the BVP waveform is derived from cPPG sensors worn by different individuals. The fitting coefficient of determination R^2 of all BVP waveforms exceeds 0.99, indicating that the Gaussian wave function can well describe the actually observed BVP waveform. Based on these fitting results, it is reasonable to assume that the signal obtained by non-contact means such as iPPG technology is actually composed of effective Gaussian waves superimposed with other interference components. Therefore, by adopting Gaussian filtering technology, a more realistic BVP signal can be effectively refined and extracted from the noisy iPPG signal.

[0107] As an optional embodiment, in order to facilitate the identification of Gaussian wave components in the iPPG signal, a rough iPPG signal can be first obtained using a traditional method. If this rough iPPG signal can be closer to the reference BVP signal, it is likely to significantly improve the accuracy of subsequent Gaussian wave component extraction.

[0108] For example, a CHROM method may be used to obtain a preliminary iPPG signal. It should be noted that this is only an example, and other methods may also be used to obtain an iPPG signal. There is no specific limitation on the method for obtaining an iPPG signal.

[0109] For example, to obtain a skin video of a target biological object, the palm of the target biological object can be obtained by performing spatial pixel averaging on the RGB channels of the video to obtain a 3×N matrix Among them, N can be used to represent the length of time. The CHROM method can eliminate the difference in reflection under different spectra by designing the projection matrix. The skin surface reflection in the interference and obtain two chromaticity signals S1(t) and S2(t):

[0110]

[0111] Among them, P c M can be used to represent the resulting projection matrix used by CHROM.

[0112] By tuning α, the iPPG signal (X t ),Right now:

[0113]

[0114] Among them, std(·) can be used to represent deviation calculation, and S′1(t) and S′2(t) can be used to represent the signals of the chrominance signals S1(t) and S2(t) after bandpass filtering, respectively.

[0115] In this embodiment, a set of multi-scale Gaussian kernel functions can be designed in advance. The Gaussian kernel functions can be used to match the potential BVP Gaussian wave components (i.e., characteristic Gaussian wave components) in the acquired iPPG signal, and further the final BVP signal can be obtained.

[0116] Alternatively, considering that the human heart rate usually varies between 48 and 200 beats per minute, the duration of the BVP signal in one cardiac cycle is about 0.3 to 1.25 seconds. According to the Gaussian mixture model of the BVP signal, the width of the Gaussian kernel can be limited to the range of 0.1 to 1.25 seconds. Assume that a Gaussian kernel function with a time width of 1 second is defined as:

[0117] G0(t)=G(t,Θ0)·[u(t)-u(t-1)]

[0118] Among them, Gaussian kernel parameter Θ0 = [1, 0.5, 0.15], which is obtained through the BVP signal fitting experiment test collected by the contact PPG device, and depends on the shape of the Gaussian wave in the BVP signal. It is obtained by Gaussian mixture fitting of cPPG data, which can be determined in advance through testing.

[0119] Optionally, in order to generate a Gaussian kernel with the required time width between 0.1 and 1.25 seconds, G0(t) can be stretched from 0.8 to 10 times to obtain a set of Gaussian kernel functions:

[0120]

[0121] Wherein, N can be used to represent the number of Gaussian kernel functions, and can be used to determine the scale resolution, for example, N = 128. It should be noted that the above numbers are only for illustration and are not specifically limited here.

[0122] In this embodiment, after obtaining a set of Gaussian kernel functions, the Gaussian kernel functions can be used to match the Gaussian components in the iPPG signal. The above matching can be understood as using the Gaussian kernel to slide the convolution of the iPPG signal. In this process, the Gaussian kernel can match the components with similar shapes to the components, so as to convert the iPPG signal into a mapping in the Gaussian space, thereby retaining the Gaussian components in the iPPG signal.

[0123] Optionally, the iPPG signal is convolved with a pre-designed Gaussian kernel to obtain its mapping in Gaussian space (ie, mapping data):

[0124]

[0125] Wherein, X can be used to represent the N×T mapping matrix, and T can be used to represent the time length of the iPPG signal. For the mapping matrix X, it retains the Gaussian component in the rough iPPG signal and suppresses the interference of non-Gaussian shapes.

[0126] In this embodiment, a plurality of mapping data may be synthesized to obtain a synthesis result, which may be determined as a blood volume pulse signal of the target biological object.

[0127] Optionally, a synthesis result is obtained by synthesizing the above-mentioned multiple Gaussian mappings, and the synthesis result may be an accurate BVP signal.

[0128] Optionally, the process of converting the Gaussian mapping matrix into a photovolume pulse wave signal can be regarded as a regression analysis task. Considering the time series nature of the BVP signal and its inherent causal relationship, this embodiment selects a neural network model based on long short-term memory units to perform this task. The LSTM network is known for its powerful ability to process time series data and can effectively capture long-term dependencies.

[0129] For example, the network framework used by the above neural network can include a two-layer LSTM structure and a linear projection layer for output conversion. Under this framework, the regression conversion process from Gaussian mapping to one-dimensional BVP signal can be formally defined as:

[0130]

[0131] Among them, Y t Can be used to indicate BVP signal. LG(X t ) can be used to represent the LSTM-based BVP generator. t ,i t , C t , o t ,h t They can be used to represent the forget gate, input gate, cell state, output gate and hidden state of LSTM, which together determine the time dependency of LSTM in solving timing signals. f ,W i ,W c ,W o Can be used to represent the weight matrix of LSTM. f ,U i ,U c ,U o Can be used to represent hidden nodes h t-1 The weight matrix of b. f ,b i ,b c ,b o Can be used to represent the bias vector. σ(·) and tanh(·) can be used to represent the sigmoid and tanh activation functions respectively. ⊙ represents the dot product operation. The matrix P can be a linear transformation matrix.

[0132] Optionally, through the above steps, the extraction of the blood flow volume initial pulse signal from the initial pulse signal can be completed.

[0133] Figure 2 is a schematic diagram of combining multi-scale Gaussian mixture filtering and LSTM according to an embodiment of the present invention, such as Figure 2 As shown, assuming that the target biological object is a human, a skin video of the human hand can be obtained, and the obtained skin video can be converted using the video to iPPG signal module 201 to obtain an original rough iPPG signal. The original rough iPPG signal can be processed using the multi-scale Gaussian convolution kernel 202 to obtain mapping data, which can be a Gaussian mapping. Multiple Gaussian mappings can be synthesized to obtain a synthesis result. Further, the mapping matrix can be processed using the BVP waveform generator 803 to obtain a blood volume pulse signal.

[0134] Optionally, the multi-scale Gaussian kernel 202 may include multiple Gaussian kernels (which may be multi-scale Gaussian kernels), and the multiple Gaussian kernels are used to match the Gaussian components in the iPPG signal, that is, the iPPG signal is convolved with a pre-designed Gaussian kernel function to obtain a mapping of the iPPG signal in the Gaussian space.

[0135] Optionally, the BVP waveform generator 203 can also be called a BVP generator, which can include an LSTM structure (Cells) and a linear layer (Linear layer), and the linear layer can be a linear projection layer, that is, the BVP waveform generator can be a network framework including two layers of LSTM structures and a linear projection layer for output conversion. Further, under this framework, the mapping data can be regressed and converted into a one-dimensional BVP signal, which can be converted by the following formula:

[0136]

[0137] In this embodiment, a Gaussian mixture model of the pulse wave signal is proposed according to the physiological formation mechanism of the pulse wave; by using a specific multi-scale Gaussian kernel function, the Gaussian wave component in the pulse wave is matched; by combining multi-scale Gaussian kernel convolution and long short-term memory network, the real pulse wave signal is restored from the original iPPG signal. That is, this method forms a methodology based on the pulse wave formation mechanism, which can effectively improve the pulse wave measurement accuracy. Compared with the existing technology, the error of the extracted waveform can be reduced from 0.18 to 0.07, and the error is reduced by one order of magnitude. Compared with most iPPG technologies, the recovered pulse wave can extract many hemodynamic indicators, and is based on the algorithm implementation, does not require the use of additional precision instruments, and the system is simple and easy to deploy.

[0138] Figure 3 is a schematic diagram of an LSTM structure according to an embodiment of the present invention, such as Figure 3 As shown, Y t Can be used to indicate BVP signal. LG(X t ) can be used to represent the LSTM-based BVP generator. t ,i t , C t , o t ,h t They can be used to represent the forget gate, input gate, cell state, output gate and hidden state of LSTM, which together determine the time dependency of LSTM in solving timing signals. f ,W i ,W c ,W o Can be used to represent the weight matrix of LSTM. f ,U i ,Uc ,U o Can be used to represent hidden nodes h t-1 The weight matrix of b. f ,b i ,b c ,b o Can be used to represent the bias vector. σ(·) and tanh(·) can be used to represent the sigmoid and tanh activation functions, respectively. Represents a dot product operation. The matrix P can be a linear transformation matrix.

[0139] Figure 4 is a schematic diagram of a test bench for verifying a pulse wave system according to an embodiment of the present invention. Figure 4 As shown, the experimental platform may include a camera, a PPG sensor and a computer terminal.

[0140] To further illustrate the feasibility of the embodiments of the present invention, further experiments were conducted on the above method. Furthermore, the palm can be selected as the region of interest (ROI), and the PPG sensor can be used to collect cPPG signals at the fingers as reference signals. The data source of the verification process may involve 18 individuals, including 12 males and 6 females, and the ages of the above individuals are between 20 and 35 years old and have no history of cardiovascular disease. A video of the palm of the above individual can be recorded using a consumer-grade camera, and the spatial resolution of the video can be 1280×720 pixels, and the acquisition frame rate is 60 frames per second. At the same time, a reference BVP signal is obtained using a finger cPPG sensor.

[0141] In order to verify the effectiveness of the above embodiment, a comparative experiment is conducted here to compare the reconstructed pulse waveform with the original imaging photoplethysmography (iPPG) waveform and the waveform obtained only by long short-term memory (LSTM) filtering. Figure 5 is a schematic diagram of a reconstructed waveform according to an embodiment of the present invention, such as Figure 5 As shown, the method proposed in the embodiment of the present invention successfully restores the BVP signal. It is worth noting that the double peak feature is well reproduced in the reconstructed waveform, while this feature is not obvious in the original iPPG signal.

[0142] Furthermore, the results of the method of the embodiment of the present invention were compared with those of several other comparative methods. Figure 6 is a schematic diagram of a first-order derivative waveform according to an embodiment of the present invention, Figure 7 is a schematic diagram of a second-order derivative waveform according to an embodiment of the present invention, combined with Figure 6 and Figure 7 It can be seen that the method proposed in the embodiment of the present invention significantly improves the quality of the restored BVP waveform.

[0143] In addition, the hemodynamic feature extraction of the recovered pulse wave was further compared. Figure 8 is a schematic diagram of the hemodynamic characteristics according to an embodiment of the present invention, such as Figure 8 As shown, it can be divided into four different groups: time (or time domain) features, which may include pulse interval (PI), peak time (CT), diastolic time (DT), main and secondary peak interval (△T), double notch time (Tn), time span from double notch to the end of the waveform (A2T) and first-order derivative peak time (D1PT); derivative (or inverse) based features, which may include points a, b, c, d and e, each point corresponding to a specific detection on the second-order derivative of the waveform; regional features (also known as area features), which may include the pulse area (PA), the area between the beginning of the wave and the inflection point (A1), and the area from the inflection point to the end of the wave (A2), where PA = A1 + A2; amplitude features, which may include systolic amplitude (SA) and diastolic amplitude (DA).

[0144] In order to facilitate the comparison of results, the Normalized Root Mean Square Error (NRMSE) is also used as an error evaluation indicator. Table 1 is a hemodynamic feature error table. As shown in Table 1, the hemodynamic feature error contained in the BVP signal extracted by the embodiment of the present invention is lower than that of the prior art. For example, for the feature Tn, the NRMSE value of the embodiment of the present invention is 0.26, which is significantly lower than 0.79 obtained by the CHROM method and 0.72 reported by the LSTM-based method. In addition, features related to waveform area, such as SA, DA, and PA, also show lower errors. Although the complexity of the BVP waveform makes its estimation quite challenging, the experimental results show that our reconstructed BVP waveform has a very small error. Specifically, the NRMSE values ​​of SA and DA are 0.19 and 0.35, respectively. Our method can also better preserve the waveform features related to derivatives. The NRMSE values ​​of derivative-based features (such as b / a, c / a, e / a, and d / a) are 1.04, 1.09, 0.79, and 0.86, respectively, indicating that these features are also reproduced relatively accurately.

[0145] Table 1 Hemodynamic characteristic error table

[0146] Compare <![CDATA[PI NRMSE ]]> <![CDATA[CT NRMSE ]]> <![CDATA[DT NRMSE ]]> <![CDATA[Tn NRMSE ]]> <![CDATA[A2T NRMSE ]]> <![CDATA[△T NRMSE ]]> <![CDATA[DIPT NRMSE ]]> CHROM Compare cPPG 0.20 0.42 0.29 0.79 0.76 0.80 0.43 LSTM Comparison cPPG 0.22 0.46 0.31 0.72 0.58 0.71 0.46 This application compares cPPG 0.16 0.25 0.22 0.26 0.33 0.23 0.41 method <![CDATA[SA NRMSE ]]> <![CDATA[DA NRMSE ]]> <![CDATA[RI NRMSE ]]> <![CDATA[PA NRMSE ]]> <![CDATA[A1 NRMSE ]]> <![CDATA[A2 NRMSE ]]> <![CDATA[IPA NRMSE ]]> CHROM Compare cPPG 0.48 0.97 1.24 0.70 0.85 3.43 69.61 LSTM Comparison cPPG 0.34 0.83 1.25 0.52 0.81 1.46 33.90 This application compares cPPG 0.19 0.35 0.42 0.16 0.26 0.45 6.99 method b / a c / a e / a d / a CHROM Compare cPPG 3.82 2.04 2.28 2.40 LSTM Comparison cPPG 1.82 1.00 1.05 0.99 This application compares cPPG 1.04 1.09 0.79 0.86

[0147] In an embodiment of the present invention, a convolution process is performed on the acquired initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space, and a synthesis process is performed on the plurality of mapping data. Based on the synthesis result obtained after the synthesis process, a final blood volume pulse signal is obtained, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0148] According to an embodiment of the present invention, a signal acquisition device is also provided. It should be noted that the signal acquisition device of this embodiment can be used to execute the signal acquisition method of the embodiment of the present invention.

[0149] Fig. 9 is a schematic diagram of a signal acquisition device according to an embodiment of the present invention. Fig. 9 As shown, the signal acquisition device 90 may include: an acquisition unit 902 , a first processing unit 904 , a second processing unit 906 and a determination unit 908 .

[0150] The acquisition unit 902 is used to acquire an initial pulse signal of the target biological object.

[0151] The first processing unit 904 is used to perform convolution processing on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space, wherein the mapping data is used to represent the Gaussian component in the initial pulse signal.

[0152] The second processing unit 906 is used to perform synthesis processing on the multiple mapping data to obtain a synthesis result.

[0153] The determination unit 908 is used to determine the synthesis result as the blood volume pulse signal of the target biological object.

[0154] The signal acquisition device of this embodiment acquires the initial pulse signal of the target biological object through the acquisition unit; performs convolution processing on the initial pulse signal through the first processing unit to obtain multiple mapping data of the initial pulse signal in the Gaussian space, wherein the mapping data is used to represent the Gaussian component in the initial pulse signal; performs synthesis processing on the multiple mapping data through the second processing unit to obtain a synthesis result; and determines the synthesis result as the blood volume pulse signal of the target biological object through the determination unit, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0155] The embodiment of the present invention can provide a computer terminal, which can be any computer terminal device in a computer terminal group. Optionally, in this embodiment, the computer terminal can also be replaced by a terminal device such as a mobile terminal.

[0156] Optionally, in this embodiment, the computer terminal may be located in at least one network device among a plurality of network devices of the computer network.

[0157] In this embodiment, the above-mentioned computer terminal can execute the program code of the following steps in the signal acquisition method: obtaining the initial pulse signal of the target biological object; performing convolution processing on the initial pulse signal to obtain multiple mapping data of the initial pulse signal in the Gaussian space, wherein the mapping data is used to represent the Gaussian component in the initial pulse signal; performing synthesis processing on the multiple mapping data to obtain a synthesis result; and determining the synthesis result as the blood volume pulsation signal of the target biological object.

[0158] Optionally, Fig.10 is a structural block diagram of a computer terminal according to an embodiment of the present invention. Fig.10 As shown, the computer terminal 1008 may include: one or more (only one is shown in the figure) processors 1002 , a memory 1004 and a transmission device 1006 .

[0159] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the signal acquisition method and device in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned signal acquisition method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal 1008 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0160] The processor can call the information and application stored in the memory through the transmission device to execute the following steps: obtain the initial pulse signal of the target biological object; perform convolution processing on the initial pulse signal to obtain multiple mapping data of the initial pulse signal in the Gaussian space, wherein the mapping data is used to represent the Gaussian component in the initial pulse signal; perform synthesis processing on the multiple mapping data to obtain a synthesis result; and determine the synthesis result as the blood volume pulsation signal of the target biological object.

[0161] It can be understood by those skilled in the art that Fig.10 The structure shown is for illustration only, and the computer terminal 1008 may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, abbreviated as MID), a PAD, or other terminal devices. Fig.10 The structure of the computer terminal 1008 is not limited. For example, the computer terminal 1008 may also include Fig.10 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Fig.10 Different configurations are shown.

[0162] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0163] According to an embodiment of the present invention, a computer-readable storage medium is further provided. The computer-readable storage medium includes a stored program, wherein the program executes the signal acquisition method in the embodiment.

[0164] Optionally, in this embodiment, the computer-readable storage medium may be located in any one of the computer terminals in a computer terminal group in a computer network, or in any one of the mobile terminals in a mobile terminal group.

[0165] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: obtaining an initial pulse signal of a target biological object; performing convolution processing on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, wherein the mapping data is used to represent a Gaussian component in the initial pulse signal; performing synthesis processing on the plurality of mapping data to obtain a synthesis result; and determining the synthesis result as a blood volume pulsation signal of the target biological object.

[0166] Optionally, the computer-readable storage medium may also execute program codes of the following steps: acquiring a chrominance signal of a target biological object; performing bandpass filtering on the chrominance signal; and converting the chrominance signal after the bandpass filtering to obtain an initial pulse signal.

[0167] Optionally, the computer-readable storage medium may also execute program codes for the following steps: using an image sensor to collect video data associated with a target part of a target biological object; and converting the video data to obtain a chrominance signal.

[0168] Optionally, the computer-readable storage medium may also execute program code for the following steps: performing Gaussian mixture fitting on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions; and performing convolution processing on the initial pulse signal using the multiple Gaussian kernel functions to obtain multiple mapping data.

[0169] Optionally, the computer-readable storage medium may also execute program code of the following steps: performing convolution processing on the initial pulse signal and the Gaussian kernel function to obtain mapping data.

[0170] Optionally, the computer-readable storage medium may also execute program codes of the following steps: performing Gaussian mixture fitting on the historical blood volume pulse signal to obtain multiple initial Gaussian kernel functions; and respectively expanding the multiple initial Gaussian kernel functions to obtain multiple Gaussian kernel functions.

[0171] Optionally, the computer-readable storage medium may also execute program code of the following steps: synthesizing multiple mapping data using a long short-term memory network to obtain a synthesis result.

[0172] In this embodiment, the acquired initial pulse signal is convolved to obtain multiple mapping data of the initial pulse signal in Gaussian space, and the multiple mapping data are synthesized to obtain the final blood volume pulse signal based on the synthesis result obtained after the synthesis, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0173] According to an embodiment of the present invention, a processor is further provided. The processor is used to run a program. When the program is run by the processor, the signal acquisition method in the embodiment is executed.

[0174] Optionally, in this embodiment, the computer terminal may be located in at least one network device among a plurality of network devices of the computer network.

[0175] In this embodiment, the computer terminal can execute the program code of the following steps in the multi-language translation method: obtaining an initial pulse signal of a target biological object; performing convolution processing on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, wherein the mapping data is used to represent a Gaussian component in the initial pulse signal; performing synthesis processing on the plurality of mapping data to obtain a synthesis result; and determining the synthesis result as a blood volume pulsation signal of the target biological object.

[0176] Wherein, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the multilingual translation method and device in the embodiment of the present invention, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned multilingual translation method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0177] The processor can call the information and application stored in the memory through the transmission device to execute the following steps: obtain the initial pulse signal of the target biological object; perform convolution processing on the initial pulse signal to obtain multiple mapping data of the initial pulse signal in the Gaussian space, wherein the mapping data is used to represent the Gaussian component in the initial pulse signal; perform synthesis processing on the multiple mapping data to obtain a synthesis result; and determine the synthesis result as the blood volume pulsation signal of the target biological object.

[0178] Optionally, the processor may also execute program codes of the following steps: acquiring a chrominance signal of a target biological object; performing bandpass filtering on the chrominance signal; and converting the chrominance signal after the bandpass filtering to obtain an initial pulse signal.

[0179] Optionally, the processor may also execute program codes of the following steps: using an image sensor to collect video data associated with a target part of a target biological object; and converting the video data to obtain a chrominance signal.

[0180] Optionally, the processor may also execute the following program code: performing Gaussian mixture fitting on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions; performing convolution processing on the initial pulse signal using the multiple Gaussian kernel functions to obtain multiple mapping data.

[0181] Optionally, the processor may also execute the program code of the following steps: performing convolution processing on the initial pulse signal and the Gaussian kernel function to obtain mapping data.

[0182] Optionally, the processor may also execute program codes of the following steps: performing Gaussian mixture fitting on the historical blood volume pulse signal to obtain multiple initial Gaussian kernel functions; and respectively expanding the multiple initial Gaussian kernel functions to obtain multiple Gaussian kernel functions.

[0183] Optionally, the processor may also execute the program code of the following steps: converting the synthesis result using a long short-term memory network to obtain a blood volume pulse signal.

[0184] By adopting the embodiment of the present invention, the acquired initial pulse signal is convolved to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space, and the plurality of mapping data are synthesized to obtain a final blood volume pulse signal based on the synthesis result obtained after the synthesis, thereby achieving the technical effect of improving the accuracy of blood volume pulse signal extraction and solving the technical problem of low accuracy of blood volume pulse signal extraction.

[0185] According to an embodiment of the present invention, a computer program product is further provided. The computer program product includes computer instructions. When the computer instructions are executed by a processor, the signal acquisition method in the embodiment is implemented.

[0186] An embodiment of the present invention may provide an electronic device, which may include a memory and a processor.

[0187] Fig.11 1 is a block diagram of an electronic device for a signal acquisition method according to an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0188] like Fig.11 As shown, the device 1100 includes a computing unit 1101, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1102 or a computer program loaded from a storage unit 1108 to a random access memory (RAM) 1103. In the RAM 1103, various programs and data required for the operation of the device 1100 can also be stored. The computing unit 1101, the ROM 1102, and the RAM 1103 are connected to each other via a bus 1104. An input / output (I / O) interface 1105 is also connected to the bus 1104.

[0189] A number of components in the device 1100 are connected to the I / O interface 1105, including: an input unit 1106, such as a keyboard, a mouse, etc.; an output unit 1104, such as various types of displays, speakers, etc.; a storage unit 1108, such as a disk, an optical disk, etc.; and a communication unit 1109, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1109 allows the device 1100 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0190] The computing unit 1101 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 1101 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 1101 performs the various methods and processes described above, such as a verification method for data. For example, in some embodiments, the verification method for data may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 1108. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 1100 via the ROM 1102 and / or the communication unit 1109. When the computer program is loaded into the RAM 1103 and executed by the computing unit 1101, one or more steps of the verification method for data described above may be executed. Alternatively, in other embodiments, the computing unit 1101 may be configured to execute the data verification method in any other appropriate manner (for example, by means of firmware).

[0191] According to an embodiment of the present invention, a method for acquiring a signal is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0192] Various embodiments of the systems and techniques described above herein may be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor that may be a dedicated or general purpose programmable processor that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0193] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0194] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0195] To provide interaction with a user, the systems and techniques described herein may be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or an LCD (liquid crystal display, monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a path ball) through which the user can provide input to the computer. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0196] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a Local Area Network (LAN), a Wide Area Network (WAN), and the Internet.

[0197] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0198] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0199] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0200] In the several embodiments provided by the present invention, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic, for example, the division of units can be a logical function division, and there can be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0201] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0202] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0203] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk or optical disk, etc., various media that can store program codes.

[0204] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for acquiring a signal, characterized in that: include: Acquiring an initial pulse signal of a target biological object; Performing convolution processing on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, wherein the mapping data is used to represent a Gaussian component in the initial pulse signal; Performing synthesis processing on the plurality of mapping data to obtain a synthesis result; The synthesis result is determined as the blood volume pulse signal of the target biological object.

2. The method according to claim 1, characterized in that The step of obtaining an initial pulse signal of a target biological object comprises: Acquiring a chromaticity signal of the target biological object; Performing bandpass filtering on the chrominance signal; The chrominance signal after the bandpass filtering is converted to obtain the initial pulse signal.

3. The method according to claim 2, characterized in that The step of acquiring the chromaticity signal of the target biological object comprises: using an image sensor to collect video data associated with a target portion of the target biological object; The video data is converted to obtain the chrominance signal.

4. The method according to claim 1, characterized in that: The convolution processing is performed on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in the Gaussian space, including: Performing Gaussian mixture fitting on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions; The initial pulse signal is convolved using a plurality of the Gaussian kernel functions to obtain a plurality of the mapping data.

5. The method according to claim 4, characterized in that The convolution processing of the initial pulse signal using the multiple Gaussian kernel functions to obtain the multiple mapping data includes: The initial pulse signal is convolved with the Gaussian kernel function to obtain the mapping data.

6. The method according to claim 4, characterized in that The Gaussian mixture fitting is performed on the historical blood volume pulse signal of the target biological object to obtain multiple Gaussian kernel functions, including: Performing Gaussian mixture fitting on the historical blood volume pulse signal to obtain multiple initial Gaussian kernel functions; The multiple initial Gaussian kernel functions are respectively expanded to obtain multiple Gaussian kernel functions.

7. The method according to claim 1, characterized in that The synthesizing process is performed on the plurality of mapping data to obtain a synthesized result, comprising: The plurality of mapping data are synthesized using a long short-term memory network to obtain the synthesis result.

8. A signal acquisition device, characterized in that: include: An acquisition unit, used for acquiring an initial pulse signal of a target biological object; a first processing unit, configured to perform convolution processing on the initial pulse signal to obtain a plurality of mapping data of the initial pulse signal in a Gaussian space, wherein the mapping data is used to represent a Gaussian component in the initial pulse signal; A second processing unit, configured to perform synthesis processing on the plurality of mapping data to obtain a synthesis result; The determination unit is used to determine the synthesis result as the blood volume pulse signal of the target biological object.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 7 when executed by a processor.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 7.

11. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method as claimed in any one of claims 1 to 7.

Images