Remote physiological signal detection method based on adaptive filter, terminal and medium

Abstract

The invention relates to the technical field of non-contact physiological signals, and discloses a remote physiological signal detection method based on an adaptive filter, a terminal and a medium. The method comprises the following steps: obtaining a to-be-detected face video, and obtaining a noise rPPG signal through noise processing; performing feature extraction on the face video to obtain a multi-scale space-time diagram; fusing the noise rPPG signal and the multi-scale space-time diagram to obtain fused data; performing denoising processing on the fused data by using a denoising module to obtain denoised data; the de-noising module adopts an adaptive filter de-noising module, the interior of the de-noising module is composed of multiple layers of filter structures, and each layer of filter structure comprises a fixed spectrum shaping filter, a scene adaptive shaping filter and a spectrum period feature extraction filter which are connected in sequence; and outputting a predicted rPPG signal according to the denoised data. According to the method, deep fusion and denoising are carried out on the noise and the multi-scale spatial-temporal characteristic data by using the adaptive filter, and high-precision rPPG signal estimation is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of non-contact physiological signal detection, and in particular to a remote physiological signal detection method, terminal and medium based on an adaptive filter. Background Art

[0002] Currently, physiological signal detection technologies based on facial videos mainly include the following different model architectures:

[0003] 1. Non-end-to-end models: These models typically require complex feature preprocessing to extract information from multiple color channels, such as RGB and YUV, and employ feature decoupling strategies to improve the accuracy of rPPG estimation. While these models guarantee accurate signal extraction to a certain extent, they suffer from cumbersome preprocessing, poor real-time performance, and limited noise suppression capabilities.

[0004] 2. End-to-end models: In recent years, end-to-end models (such as the video Transformer architecture) have gradually emerged. These models can adaptively aggregate local and global spatiotemporal features, thereby reducing intermediate preprocessing steps and enhancing the representation of rPPG signals. However, such models are generally computationally intensive and require high hardware performance, making them difficult to operate efficiently in resource-constrained environments.

[0005] 3. Unsupervised models based on contrastive learning: These models leverage the spatial similarity between different regions and the temporal similarity of short-term rPPG signals for feature learning, reducing the reliance on labeled data. However, their model stability and detection accuracy still need to be improved.

[0006] While existing non-contact physiological signal detection technologies have made progress in practical applications, some shortcomings remain. First, the accuracy of signal detection results is often difficult to guarantee due to external interference such as lighting changes, facial occlusions (such as hair, beards, glasses), and head movement. Second, optical differences caused by different races, ages, and skin characteristics limit the generalization ability of models. Furthermore, while some end-to-end models based on deep networks such as 3D CNNs can improve detection accuracy, their high computational complexity limits their application in low-resource scenarios. Summary of the Invention

[0007] To solve the technical problems existing in the prior art, the present invention provides a remote physiological signal detection method, terminal and medium based on an adaptive filter, which uses an adaptive filter to deeply fuse and denoise noise and multi-scale spatiotemporal feature data, thereby achieving high-precision rPPG signal estimation.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] The present invention discloses a remote physiological signal detection method based on an adaptive filter, comprising:

[0010] Obtain the face video to be detected and obtain the noise rPPG signal through noise processing;

[0011] Extracting features from the face video to obtain a multi-scale spatiotemporal graph;

[0012] fusing the noisy rPPG signal and the multi-scale spatiotemporal graph to obtain fused data;

[0013] The fused data is denoised using a denoising module to obtain denoised data; wherein the denoising module adopts an adaptive filter denoising module, which is internally composed of a multi-layer filter structure, and each layer of the filter structure includes a fixed spectrum shaping filter, a scenario-adaptive shaping filter and a spectrum periodicity feature extraction filter connected in sequence; the fixed spectrum shaping filter is used to perform preliminary frequency domain filtering through randomly initialized fixed parameters; the scenario-adaptive shaping filter uses the output of the fixed spectrum shaping filter to generate data-dependent filtering parameters to achieve context-adaptive frequency domain filtering, thereby reconstructing the periodicity and error components in the time series; the spectrum periodicity feature extraction filter is used to perform frequency domain reconstruction on the output of the scenario-adaptive shaping filter to restore the periodicity and error components in the time series;

[0014] Based on the denoised data, the predicted rPPG signal is output.

[0015] As a further improvement of the above solution, the processing process of the fixed spectrum shaping filter includes:

[0016] Perform Fourier transform on the input fusion data, i.e., the time series T, to obtain the frequency domain representation of the time series

[0017]

[0018] Where, For Fourier transform processing;

[0019] Using the pre-randomly initialized and fixed filter parameters H φ Frequency domain representation Perform element-by-element multiplication to obtain the frequency domain features after preliminary shaping

[0020]

[0021] Where, and All are channel-shared filter parameters;L is the element-wise multiplication in the dimension of sequence length L;

[0022] right Perform inverse Fourier transform to obtain the output T of the fixed spectrum shaping filter plain :

[0023]

[0024] Where, represents the inverse Fourier transform.

[0025] As a further improvement to the above solution, the processing process of the scenario-adaptive shaping filter includes:

[0026] The output of the fixed spectrum shaping filter T plain Perform Fourier transform to obtain the frequency domain representation after Fourier transform

[0027]

[0028] Generating adaptive filter parameters via data-dependent neural networks

[0029]

[0030] Where NN(·) is the neural network processing;

[0031] Representing the frequency domain and adaptive filter parameters Multiply element by element to get the reshaped frequency domain features

[0032]

[0033] The frequency domain features after shaping Perform inverse Fourier transform to obtain the output T of the scene adaptive shaping filter ctx :

[0034]

[0035] As a further improvement of the above solution, the processing process of the spectrum period feature extraction filter includes:

[0036] The output of the scene adaptive shaping filter T ctx Perform a Fourier transform to extract amplitude and phase information:

[0037]

[0038] Where A is the amplitude, Φ is the phase, and ∠ represents the angle;

[0039] Select the first K important frequency components from the amplitude A. The frequencies and phases corresponding to the K important frequency components are and Construct the frequency domain reconstructed signal by weighted synthesis:

[0040]

[0041] Where, T syn Reconstruct the signal in the frequency domain; A k is the amplitude of the kth important frequency component in the amplitude A; t is the time variable; t is the time variable;

[0042] Reconstruct the frequency domain signal T syn is the output of the entire denoising process.

[0043] As a further improvement of the above solution, the noise processing expression is as follows:

[0044]

[0045] Where x0 is the original rPPG signal; x t’ is the noise rPPG signal after the t'th noise processing; t'=1,2,…,N, N is the total number of noise processing times; q(x t’ ||x0) means that given the initial state x0, the state x t’ The conditional probability distribution of ; is the attenuation factor, During the processing, it gradually decreases with the increase of the number of processing times t'; ∈ is noise.

[0046] As a further improvement to the above solution, the specific process of extracting facial features includes:

[0047] Divide the face video into multiple segments, and obtain F frames of video images in each segment;

[0048] Perform face detection on the video image and divide the face into n regions of interest (ROIs);

[0049] Combine n ROIs to obtain a combination of a single ROI and a combination of different ROIs, a total of 2 n -1 combination;

[0050] For each combination, calculate the average pixel value of all ROIs in the combination in each color channel; there are 6 color channels in total, namely RGB and YUV;

[0051] Finally, the dimension is 6×(2 n -1)×F vector as a multi-scale space-time graph.

[0052] As a further improvement to the above solution, the specific process of fusing the noisy rPPG signal and the multi-scale spatiotemporal graph includes:

[0053] The dimension of the noise rPPG signal is 1×F, and the dimension of the noise rPPG signal is expanded to 1×(2 n -1)×F, the expanded noisy rPPG signal and the multi-scale spatiotemporal map are spliced ​​in the channel dimension to obtain the fused data. The dimension of the fused data is 7×(2 n -1)×F.

[0054] As a further improvement of the above solution, the noise rPPG signal is Gaussian noise that obeys a standard normal distribution.

[0055] The present invention also discloses a remote physiological signal detection system based on an adaptive filter, which is applied to the remote physiological signal detection method based on an adaptive filter as described above; the detection system includes: a noise processing module, an MSTmap calculation module, a fusion module and a denoising module.

[0056] The noise processing module is used to obtain the face video to be detected and obtain the noise rPPG signal through noise processing;

[0057] The MSTmap calculation module is used to extract features from face videos and obtain multi-scale spatiotemporal maps;

[0058] The fusion module is used to fuse the noisy rPPG signal and the multi-scale spatiotemporal graph to obtain fused data;

[0059] The denoising module is used to perform denoising on the fused data to obtain denoised data; wherein, the denoising module adopts an adaptive filter denoising module, which is internally composed of a multi-layer filter structure, and each layer of the filter structure includes a fixed spectrum shaping filter, a scenario-adaptive shaping filter and a spectrum periodicity feature extraction filter connected in sequence; the fixed spectrum shaping filter is used to perform preliminary frequency domain filtering through randomly initialized fixed parameters; the scenario-adaptive shaping filter uses the output of the fixed spectrum shaping filter to generate data-dependent filtering parameters to achieve context-adaptive frequency domain filtering, thereby reconstructing the periodicity and error components in the time series; the spectrum periodicity feature extraction filter is used to perform frequency domain reconstruction on the output of the scenario-adaptive shaping filter to restore the periodicity and error components in the time series;

[0060] Prediction module, which is used to output the predicted rPPG signal based on the denoised data

[0061] The present invention also discloses a computer terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the remote physiological signal detection method based on the adaptive filter as described above are implemented.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] 1. The present invention utilizes adaptive filter technology and a multi-layer structure (including a fixed spectrum shaping filter, a scenario-adaptive shaping filter, and a spectrum periodicity feature extraction filter) to achieve hierarchical denoising of noisy rPPG signals and accurate reconstruction of real signals, significantly improving the generalization ability of the model. The adaptive filter fully utilizes the periodicity and spectral characteristics of the rPPG signal during the recovery process, making the final output physiological signal more consistent with the real human body state. Through the collaborative operation of multiple filters, the present invention can effectively cope with external interference such as lighting changes, facial occlusion, and head movement, ensuring the stability and robustness of the detection results and achieving high-precision rPPG signal estimation.

[0064] 2. Compared with traditional deep convolutional networks or end-to-end Transformer models, the adaptive filter uses layer-by-layer denoising and feature fusion to reduce computational complexity while maintaining high-precision signal detection, making it suitable for real-time scenarios with limited resources.

[0065] 3. Extracting effective regional features using a multi-scale spatiotemporal map (MSTmap): During the MSTmap calculation process, a face detection tool is used to segment the effective facial regions in the video, allowing the model to focus on high-information areas, thereby significantly improving the accuracy of signal estimation. MSTmap calculation is not only based on the original RGB channels, but also introduces additional YUV channels, enriching the visual feature information and providing a more comprehensive basis for model training and inference. By permuting and combining multiple regions of interest, the MSTmap can display richer spatial information, allowing the subsequent adaptive filter module to extract more effective features during denoising, further improving detection accuracy.

[0066] 4. The present invention adopts a splicing method to fuse the one-dimensional noisy rPPG signal with the multi-channel MSTmap, maintaining the independence of their respective features. It can not only learn the distribution characteristics of the rPPG signal during the noise restoration process, but also guide the mapping of the MSTmap to the real signal to avoid mutual interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a framework diagram of a remote physiological signal detection system based on an adaptive filter in Example 1 of the present invention.

[0068] Figure 2 This is a flow chart of a remote physiological signal detection method based on an adaptive filter according to embodiment 1 of the present invention.

[0069] Figure 3 Flowchart of the training process and estimation process of the adaptive filter model in Example 1 of the present invention.

[0070] Figure 4 This is a schematic diagram of the structure of a computer terminal in Example 2 of the present invention. DETAILED DESCRIPTION

[0071] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0072] Example 1

[0073] See also Figure 1 This embodiment provides a remote physiological signal detection system based on an adaptive filter, including: a noise processing module, an MSTmap (Multi-scale Stmap) calculation module, a fusion module, a denoising module, a prediction module and a sampling module. The working process of the entire system is divided into two parts: a training process and an estimation process.

[0074] During training:

[0075] A noise processing module is used to add noise to the original rPPG signal to obtain a noisy rPPG signal;

[0076] The MSTmap calculation module is used to process face videos and obtain a multi-scale spatiotemporal map containing visual features;

[0077] The fusion module is used to fuse the noisy rPPG signal with the multi-scale spatiotemporal graph to obtain fused data;

[0078] Denoising module, used to perform denoising on the fused data to obtain denoised data;

[0079] The prediction module is used to output the predicted rPPG signal based on the denoised data.

[0080] During the estimation process:

[0081] Sampling module, used for random sampling to generate noise rPPG signal;

[0082] The MSTmap calculation module is used to process the estimated face video to obtain a multi-scale spatiotemporal map containing visual features;

[0083] The fusion module is used to fuse the noisy rPPG signal with the multi-scale spatiotemporal graph to obtain fused data;

[0084] Denoising module, used to perform denoising on the fused data to obtain denoised data;

[0085] The prediction module is used to output the predicted rPPG signal based on the denoised data.

[0086] Depend on Figure 2 and Figure 3 As shown, this embodiment also provides a remote physiological signal detection method based on an adaptive filter. First, during the training process, the collected sample data is input to train the adaptive filter model, and then the trained adaptive filter model is used to perform remote physiological signal estimation. During the estimation process, the input is a face video of the person to be detected, and the output is the corresponding predicted rPPG signal.

[0087] The training process of the adaptive filter model is as follows:

[0088] Step 1.1, obtaining a sample set, the sample including a face video and the original rPPG signal corresponding to the face video;

[0089] In step 1.2, the original rPPG signal is input into the noise processing module to add noise, transforming the ordered original rPPG signal into a disordered noisy rPPG signal, gradually transitioning from an initial state with almost no noise to a state composed mainly of noise. This process provides a rich dataset for model training, including all intermediate states from no noise to high noise, which is crucial for the subsequent denoising process because it relies on these intermediate states to learn how to gradually restore to the original noise-free state. The mathematical expression of noise processing is as follows:

[0090]

[0091] Among them, x0 represents the original rPPG signal; x t’ represents the noise rPPG signal after the t'th noise processing; t'=1,2,...,N, and the noise rPPG signal x after N noise processing is finally obtained N N is the total number of noise processing times, in this embodiment N = 1000; q(x t’ ||x0) means that given the initial state x0, the state x t’ The conditional probability distribution of x0 to x t’ the evolutionary process; is a predefined attenuation factor, Gradually decreases during the treatment process. Controls the proportion of the original rPPG signal retained in each step of noise processing; ∈ represents random noise drawn from the standard normal distribution N(0,I); The proportion of noise in each step of noise processing is determined. As the number of noise processing times t' increases, Gradually increase.

[0092] In step 1.3, the face video corresponding to the original rPPG signal is input into the MSTmap calculation module to calculate the multi-scale spatiotemporal map. The specific method is as follows:

[0093] The input face video is divided into multiple segments with F frames as one segment and 10 frames between each segment, that is, F frames of video images are obtained in each segment. In this embodiment, F=200;

[0094] The facial detection tool OpenFace is used to perform face detection and extract 68 facial feature points. Based on the facial feature points, six ROIs (Region Of Interester) of the face are defined, namely the forehead, left cheek, right cheek, left triangular area, right triangular area and chin.

[0095] Combine n ROIs to obtain a combination of a single ROI and a combination of different ROIs, a total of 2 n -1 combination, in this embodiment n=6.

[0096] For each combination, the average pixel value of all ROIs in the combination in each color channel is calculated. There are 6 color channels in total, namely RGB and YUV.

[0097] Finally, the dimension is 6×(2 n -1)×F is used as a multi-scale space-time graph. In this embodiment, the dimension of the multi-scale space-time graph is 1×63×200.

[0098] In step 1.4, the noisy rPPG signal obtained in step S12 and the multi-scale spatiotemporal map obtained in step S13 are input into the fusion module for fusion to obtain fused data.

[0099] Noise rPPG signal x after noise processing N The dimension is 1×300. First, the noise rPPG signal x N The dimension is expanded to 1×63×300, and the expanded noisy rPPG signal is connected with the multi-scale spatiotemporal graph using the concat function in the channel dimension to obtain the fused data y N , fused data y NThe dimensions are 7×63×200.

[0100] In step 1.5, the fused data is input into the denoising module for denoising to obtain the denoised data, thereby obtaining a stable and pure (i.e., noise-free) rPPG feature vector.

[0101] The denoising module uses an adaptive filter denoising module (MFD), which consists of a c-layer filter structure. Each filter layer includes a fixed spectrum shaping filter, a context-adaptive shaping filter, and a spectrum period feature extraction filter connected in sequence. This adaptive extraction of features through multiple filter modules fuses multiple layers of embedded features to better understand data relationships. In this embodiment, c = 6.

[0102] The fixed spectrum shaping filter uses pre-randomly initialized and fixed parameters to filter the frequency domain representation of the input signal element by element, directly suppressing noise and preserving key spectral features.

[0103] The context-adaptive shaping filter generates data-dependent filtering parameters according to the input signal, and realizes context-adaptive spectrum shaping by adaptively adjusting the frequency domain filtering process, thereby reconstructing the periodicity and error components in the time series.

[0104] The spectrum period feature extraction filter extracts the amplitude and phase information of the input signal through Fourier transform, selects the first K important frequency components and performs weighted synthesis, thereby realizing frequency domain reconstruction and restoring the periodicity and error components in the time series.

[0105] The mathematical expression of the denoising module during training is as follows:

[0106] z=MFD(y N )

[0107] Among them, y N is the fused data, z is the denoised data; MFD(·) is the processing function of the adaptive filter denoising module.

[0108] MFD(y N )=Fourier(Contextual(Plain(y N )))

[0109] Among them, Plain(·) is the processing function of the fixed spectrum shaping filter, and the processing method is:

[0110] First, perform Fourier transform on the input fusion data, i.e., the time series T, to obtain the frequency domain representation of the time series

[0111]

[0112] Where, For Fourier transform processing;

[0113] Then, the filter parameters H are randomly initialized and fixed in advance. φ Frequency domain representation Perform element-by-element multiplication to obtain the frequency domain features after preliminary shaping

[0114]

[0115] Where, and All are channel-shared filter parameters; L is the element-wise multiplication in the dimension of sequence length L;

[0116] The last pair Perform inverse Fourier transform to obtain the output T of the fixed spectrum shaping filter plain :

[0117]

[0118] Where, represents the inverse Fourier transform.

[0119] Contextual(·) is the processing function of the context-adaptive shaping filter, and the processing method is:

[0120] First, the output T of the fixed spectrum shaping filter is plain Perform Fourier transform to obtain the frequency domain representation after Fourier transform

[0121]

[0122] Then, adaptive filter parameters are generated through data-dependent neural networks

[0123]

[0124] Where NN(·) is the neural network processing;

[0125] Next, the frequency domain representation and adaptive filter parameters Multiply element by element to get the reshaped frequency domain features

[0126]

[0127] Finally, the frequency domain features after shaping Perform inverse Fourier transform to obtain the output T of the scene adaptive shaping filter ctx :

[0128]

[0129] Fourier(·) is the processing function of the spectrum period feature extraction filter, and the processing method is:

[0130] First, the output of the scene adaptive shaping filter T ctx Perform a Fourier transform to extract amplitude and phase information:

[0131]

[0132] Where A is the amplitude, Φ is the phase, and ∠ represents the angle;

[0133] Then select the first K important frequency components from the amplitude A. The frequencies and phases corresponding to the K important frequency components are and Construct the frequency domain reconstructed signal by weighted synthesis:

[0134]

[0135] Where, T syn Reconstruct the signal in the frequency domain; A k is the amplitude of the kth important frequency component in amplitude A; t is the time variable;

[0136] Finally, the frequency domain reconstructed signal T syn As the output of the entire denoising process.

[0137] In step 1.6, the denoised feature vector z obtained in step 1.5 is input into the prediction module to obtain the predicted rPPG signal.

[0138] The model is trained by minimizing the difference between the predicted rPPG signal and the original rPPG signal to obtain a trained model.

[0139] First, the denoised feature vector z (dimension is 7×63×200) is average pooled in the second dimension to obtain the intermediate feature (dimension is 7×1×200). Then, a Linear layer is passed to reduce the channel dimension to finally obtain the predicted rPPG signal (dimension is 1×1×200).

[0140] The trained adaptive filter model is used to estimate the remote physiological signal. The estimation process is as follows:

[0141] Step 2.1, use the sampling module to obtain the normal distribution An initial noise rPPG signal x' is sampled from the Gaussian noise N , and then use the randomly generated noise rPPG signal x' N As the starting point of the denoising process. The mathematical expression of sampling is as follows:

[0142]

[0143] Among them, the randomly generated noise rPPG signal x' N The dimension is 1×200.

[0144] Step 2.2: Following the method in step 1.3, input the input face video into the MSTmap calculation module to calculate the multi-scale spatiotemporal map.

[0145] Step 2.3: Follow the steps in step 1.4 to convert the randomly generated noise rPPG signal x' N And the multi-scale spatiotemporal graph is input into the fusion module for fusion to obtain the fused data y' N .

[0146] Step 2.4, the fused data y' N Input into the denoising module for denoising to obtain denoised data.

[0147] The denoising method during the estimation process is different from that during the training process. The feature vector output by the denoising module during the estimation process is used as input data again and input into the denoising module for denoising. The final output is obtained after K iterations of the denoising process. The expression is as follows:

[0148] z k =MFD(z k+1 )+z k+1

[0149] MFD(z k+1 )=Fourier(Contextual(Plain(z k+1 ))×c

[0150] Where k is the number of iterations of the denoising process, k = 1, 2, ..., K; z k is the eigenvector of the kth denoising process, z0=y' N , and finally obtain the feature vector z after K denoising processes K , that is, the denoised data, z K The dimensions are 7×63×200.

[0151] z1=y 800 =y 1000 +MFD(y1000 ), k=1

[0152] z2=y 600 =y 800 +MFD(y 800 ), k=2

[0153] z3=y 400 =y 600 +MFD(y 600 ), k=3

[0154] z4=y 200 =y 400 +MFD(y 400 ), k=4

[0155] Step 2.5: Follow the steps in step 1.6 to convert the feature vector z K Input into the prediction module and finally obtain the predicted rPPG signal.

[0156] The present invention uses a multi-scale spatiotemporal map, i.e., MSTmap, instead of directly inputting the video, which has the following advantages:

[0157] During the MSTmap calculation process, facial recognition tools are used to precisely locate faces, effectively improving the model's recognition accuracy. Directly feeding the video into the model can be difficult for the model to accurately distinguish valid areas within the frame, leading to errors such as mistaking background for faces. However, the use of facial recognition tools allows the model to focus on valid facial areas, further improving recognition accuracy.

[0158] In addition to using facial recognition tools, MSTmap innovatively introduces YUV channels in addition to the existing RGB channels. These three additional channels have been proven effective by previous research and can provide the model with richer feature information, thereby enabling higher-quality training and inference.

[0159] Furthermore, MSTmap cleverly arranges and combines facial ROIs, generating 63 different combinations from the original six regions. These rich combinations enable the multi-filter module in the MFD layer of the Denoiser to extract more valuable information, resulting in more accurate calculation results.

[0160] This invention fuses the rPPG signal with the color channel dimension of the MSTmap. During the channel fusion process, the one-dimensional noisy rPPG signal is combined with the six-dimensional color channel of the MSTmap. Because the MSTmap itself occupies a dominant position in terms of dimension, it is not affected by the interference of the noisy rPPG signal and thus affects the model's prediction performance. Furthermore, the noisy rPPG signal is fused with the MSTmap via a concat method. This process does not modify the numerical values ​​of the noisy rPPG signal and the MSTmap; it simply concatenates their dimensions, and the features they represent are independent of each other. This unique fusion method enables the model to learn the data distribution of the rPPG signal from the noisy rPPG signal and, using this data distribution, provides powerful guidance for mapping the MSTmap to the true rPPG signal.

[0161] Example 2

[0162] This embodiment provides a computer terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the remote physiological signal detection method based on the adaptive filter as described in Example 1 are implemented.

[0163] like Figure 4 As shown, the computer terminal provided in this embodiment includes: at least one processor 101, and a memory 102 connected to the at least one processor 101. The specific connection medium between the processor 101 and the memory 102 is not limited in this embodiment. Figure 4 In the example, the processor 101 and the memory 102 are connected via the bus 100. Figure 4 The bus 100 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 The bus is represented by only one thick line, but it does not mean that there is only one bus or one type of bus. Alternatively, the processor 101 can also be called a controller, and there is no limitation on the name.

[0164] In this embodiment, the memory 102 stores instructions that can be executed by at least one processor 101 , and the at least one processor 101 can perform the aforementioned method by executing the instructions stored in the memory 102 .

[0165] Among them, the processor 101 is the control center of the device, which can use various interfaces and lines to connect the various parts of the entire control device, and monitor the device as a whole by running or executing instructions stored in the memory 102 and calling data stored in the memory 102, the various functions of the device and processing data.

[0166] In one possible design, processor 101 may include one or more processing units. Processor 101 may integrate an application processor and a modem processor. The application processor primarily processes the operating system, user interface, and application programs, while the modem processor primarily handles wireless communications. It is understood that the modem processor may not be integrated into processor 101. In some embodiments, processor 101 and memory 102 may be implemented on the same chip. In some embodiments, they may also be implemented on separate chips.

[0167] The processor 101 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component, and can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the remote physiological signal detection method based on the adaptive filter disclosed in Example 1 can be directly implemented as a hardware processor, or can be implemented by a combination of hardware and software modules in the processor 101.

[0168] The memory 102 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 102 may include at least one type of storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory, a random access memory (RAM), a static random access memory (SRAM), a programmable read-only memory (PROM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic memory, a magnetic disk, an optical disk, etc. The memory 102 is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 102 in this embodiment can also be a circuit or any other device that can implement a storage function, for storing program instructions and / or data.

[0169] By designing and programming the processor 101, the code corresponding to the security verification method described in the above embodiment can be fixed into the chip, so that the chip can execute the security verification method when it is running. Figure 2 The steps of the remote physiological signal detection method based on the adaptive filter are shown. How to design and program the processor 101 is a technology well known to those skilled in the art and will not be described in detail here.

[0170] Example 3

[0171] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the steps of the remote physiological signal detection method based on the adaptive filter as described in Example 1 are implemented.

[0172] The computer-readable storage medium may include flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the storage medium may be an internal storage unit of a computer device, such as the hard disk or memory of the computer device. In other embodiments, the storage medium may also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a smart media card (SMC), a secure digital (SD) card, a flash memory card, etc. Of course, the storage medium may also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the memory is generally used to store the operating system and various application software installed on the computer device. In addition, the memory may also be used to temporarily store various types of data that have been output or are about to be output.

[0173] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A remote physiological signal detection method based on an adaptive filter, characterized in that: include: Obtain the face video to be detected and obtain the noise rPPG signal through noise processing; Extracting features from the face video to obtain a multi-scale spatiotemporal graph; fusing the noisy rPPG signal and the multi-scale spatiotemporal graph to obtain fused data; The fused data is denoised using a denoising module to obtain denoised data; wherein the denoising module adopts an adaptive filter denoising module, which is internally composed of a multi-layer filter structure, and each layer of the filter structure includes a fixed spectrum shaping filter, a scenario-adaptive shaping filter and a spectrum periodicity feature extraction filter connected in sequence; the fixed spectrum shaping filter is used to perform preliminary frequency domain filtering through randomly initialized fixed parameters; the scenario-adaptive shaping filter uses the output of the fixed spectrum shaping filter to generate data-dependent filtering parameters to achieve context-adaptive frequency domain filtering, thereby reconstructing the periodicity and error components in the time series; the spectrum periodicity feature extraction filter is used to perform frequency domain reconstruction on the output of the scenario-adaptive shaping filter to restore the periodicity and error components in the time series; Based on the denoised data, the predicted rPPG signal is output.

2. The remote physiological signal detection method based on adaptive filter according to claim 1, characterized in that: The processing process of the fixed spectrum shaping filter includes: Perform Fourier transform on the input fusion data, i.e., the time series T, to obtain the frequency domain representation of the time series Where, For Fourier transform processing; Using the pre-randomly initialized and fixed filter parameters H φ Frequency domain representation Perform element-by-element multiplication to obtain the frequency domain features after preliminary shaping Where, and All are channel-shared filter parameters; L is the element-wise multiplication in the dimension of sequence length L; right Perform inverse Fourier transform to obtain the output T of the fixed spectrum shaping filter plain : Where, represents the inverse Fourier transform.

3. The remote physiological signal detection method based on adaptive filter according to claim 2, characterized in that: The processing of the scenario-adaptive shaping filter includes: The output of the fixed spectrum shaping filter T plain Perform Fourier transform to obtain the frequency domain representation after Fourier transform Generating adaptive filter parameters via data-dependent neural networks Where NN(·) is the neural network processing; Representing the frequency domain and adaptive filter parameters Multiply element by element to get the reshaped frequency domain features The frequency domain features after shaping Perform inverse Fourier transform to obtain the output T of the scene adaptive shaping filter ctx :

4. The remote physiological signal detection method based on adaptive filter according to claim 3, characterized in that: The processing process of the spectrum period feature extraction filter includes: The output of the scene adaptive shaping filter T ctx Perform a Fourier transform to extract amplitude and phase information: Where A is the amplitude, Φ is the phase, and ∠ represents the angle; Select the first K important frequency components from the amplitude A. The frequencies and phases corresponding to the K important frequency components are and Construct the frequency domain reconstructed signal by weighted synthesis: Where, T syn Reconstruct the signal in the frequency domain; A k is the amplitude of the kth important frequency component in amplitude A; t is the time variable; Reconstruct the frequency domain signal T syn As the output of the entire denoising process.

5. The remote physiological signal detection method based on adaptive filter according to claim 1, characterized in that: The noise processing expression is as follows: Where x0 is the original rPPG signal; x t’ is the noise rPPG signal after the t'th noise processing; t'=1,2,…,N, N is the total number of noise processing times; q(x t’ |x0) means that given the initial state x0, the state x t’ The conditional probability distribution of ; is the attenuation factor, During the processing, it gradually decreases with the increase of the number of processing times t'; ∈ is noise.

6. The remote physiological signal detection method based on adaptive filter according to claim 1, characterized in that: The specific process of feature extraction for facial features includes: Divide the face video into multiple segments, and obtain F frames of video images in each segment; Perform face detection on the video image and divide the face into n regions of interest (ROIs); Combine n ROIs to obtain a combination of a single ROI and a combination of different ROIs, a total of 2 n -1 combination; For each combination, calculate the average pixel value of all ROIs in the combination in each color channel; there are 6 color channels in total, namely RGB and YUV; Finally, the dimension is 6×(2 n -1)×F vector as a multi-scale space-time graph.

7. The remote physiological signal detection method based on adaptive filter according to claim 6, characterized in that: The specific process of fusing the noisy rPPG signal and the multi-scale spatiotemporal graph includes: The dimension of the noise rPPG signal is 1×F, and the dimension of the noise rPPG signal is expanded to 1×(2 n -1)×F, the expanded noisy rPPG signal and the multi-scale spatiotemporal map are spliced ​​in the channel dimension to obtain the fused data. The dimension of the fused data is 7×(2 n -1)×F.

8. The remote physiological signal detection method based on adaptive filter according to claim 1, characterized in that: The noise rPPG signal is Gaussian noise that obeys a standard normal distribution.

9. A remote physiological signal detection system based on an adaptive filter, characterized in that: The method for remote physiological signal detection based on an adaptive filter according to any one of claims 1 to 8 is applied; the detection system comprises: The noise processing module is used to obtain the face video to be detected and obtain the noise rPPG signal through noise processing; MSTmap calculation module, used to extract features from face videos and obtain multi-scale spatiotemporal maps; a fusion module, configured to fuse the noisy rPPG signal and the multi-scale spatiotemporal graph to obtain fused data; A denoising module is configured to denoise the fused data to obtain denoised data; wherein the denoising module adopts an adaptive filter denoising module, which is internally composed of a multi-layer filter structure, and each layer of the filter structure includes a fixed spectrum shaping filter, a scenario-adaptive shaping filter and a spectrum periodicity feature extraction filter connected in sequence; the fixed spectrum shaping filter is configured to perform preliminary frequency domain filtering using randomly initialized fixed parameters; the scenario-adaptive shaping filter uses the output of the fixed spectrum shaping filter to generate data-dependent filtering parameters to achieve context-adaptive frequency domain filtering, thereby reconstructing the periodicity and error components in the time series; the spectrum periodicity feature extraction filter is configured to perform frequency domain reconstruction on the output of the scenario-adaptive shaping filter to restore the periodicity and error components in the time series; The prediction module is used to output the predicted rPPG signal based on the denoised data.

10. A computer terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the remote physiological signal detection method based on the adaptive filter are implemented as claimed in any one of claims 1 to 8.

Images