CN-121587725-B - Psychological state evaluation parameter determining method, psychological state evaluation parameter determining device, electronic equipment and storage medium

CN121587725BCN 121587725 BCN121587725 BCN 121587725BCN-121587725-B

Abstract

The invention relates to the technical field of non-contact biological signal processing, and provides a psychological state assessment parameter determining method, a device, electronic equipment and a storage medium, wherein the method comprises the steps of inputting millimeter wave radar signals into a psychological state assessment model to obtain psychological state assessment parameters; the psychological state assessment model comprises a physiological characteristic extraction module, a micro-motion characteristic extraction module, a multi-modal characteristic fusion module and a psychological state inference module. According to the invention, the physiological characteristic extraction module is used for reconstructing the electrocardio-activity waveform to obtain fine physiological characteristics, the micro-motion characteristic extraction module is used for extracting the micro-motion characteristic, and the multi-mode characteristic fusion module is used for integrating the characteristics respectively reflecting the internal physiological change and the external behavior expression, so that the fusion characteristics with more comprehensive information and richer dimensionality are formed for the psychological state inference module to carry out comprehensive analysis, and the comprehensiveness and the accuracy of psychological state assessment under the unconstrained daily environment are greatly improved.

Inventors

Ru Yiwei
SUN ZHENAN
LI JUNYANG
XU YANLIN

Assignees

中国科学院自动化研究所

Dates

Publication Date: 20260505
Application Date: 20260128

Claims (9)

1. A method for determining a mental state estimation parameter, comprising: Acquiring a millimeter wave radar signal of a target object to be evaluated; inputting the millimeter wave radar signal into a psychological state assessment model to obtain psychological state assessment parameters output by the psychological state assessment model; the psychological state assessment model comprises a physiological characteristic extraction module, a micro-motion characteristic extraction module, a multi-modal characteristic fusion module and a psychological state inference module; The system comprises a physiological characteristic extraction module, a multi-mode characteristic fusion module, a psychological state inference module and a psychological state estimation module, wherein the physiological characteristic extraction module is used for reconstructing an electrocardio-activity waveform based on the millimeter wave radar signal and extracting physiological characteristics from the electrocardio-activity waveform, the micro-motion characteristic extraction module is used for extracting micro-motion characteristics of the millimeter wave radar signal to obtain micro-motion characteristics, the multi-mode characteristic fusion module is used for fusing the physiological characteristics and the micro-motion characteristics to obtain fusion characteristics, the psychological state inference module is used for obtaining the psychological state estimation parameter based on the fusion characteristics, and the electrocardio-activity waveform is a time sequence which is reconstructed from the millimeter wave radar signal and highly simulates a true solid electric activity signal in waveform form and rhythm; the training step of the physiological characteristic extraction module comprises the following steps: Acquiring a millimeter wave radar sample signal and a reference electrocardio activity waveform synchronously acquired with the millimeter wave radar sample signal; Inputting the millimeter wave radar sample signal to an initial physiological characteristic extraction module to obtain an electrocardio-activity waveform to be rebuilt, which is output by the initial physiological characteristic extraction module; Determining diagnosis weights corresponding to different waveform segments in the reference electrocardio-activity waveform according to clinical diagnosis importance of the different waveform segments, calculating weighted morphology differences between the electrocardio-activity waveform to be reconstructed and the reference electrocardio-activity waveform based on the diagnosis weights, and determining morphological fidelity loss; And determining target loss based on the morphological fidelity loss, and carrying out parameter iteration on the initial physiological characteristic extraction module based on the target loss to obtain the physiological characteristic extraction module.
2. The method of claim 1, wherein determining a target loss based on the morphological fidelity loss comprises: Extracting a first heart rate variability parameter and a second heart rate variability parameter from the to-be-reconstructed electrocardio-activity waveform and the reference electrocardio-activity waveform respectively; Determining a consistency loss based on the first heart rate variability parameter and the second heart rate variability parameter; inputting the to-be-reconstructed electrocardio-activity waveform and the reference electrocardio-activity waveform into a generation countermeasure network, and determining a first discrimination output of the reference electrocardio-activity waveform and a second discrimination output of the to-be-reconstructed electrocardio-activity waveform by a discriminator in the generation countermeasure network, wherein the generation countermeasure network comprises a generator and a discriminator, and the generator is the initial physiological characteristic extraction module; determining a arbiter loss for the arbiter based on the first and second arbitration outputs; determining a generator loss for the generator based on the inverse of the second discrimination output; determining an antagonism loss based on the arbiter loss and the generator loss; Determining the target loss based on the morphological fidelity loss, the consistency loss, and the antagonism loss.
3. The method of determining a mental state estimation parameter according to claim 2, wherein the determining the target loss based on the morphological fidelity loss, the consistency loss, and the contrast loss includes: Mapping the electrocardio-activity waveform to be rebuilt and the millimeter wave radar sample signal to a shared feature space respectively to obtain an electrocardio-activity waveform feature to be rebuilt and a millimeter wave radar sample signal feature, and determining content loss based on feature distances of the electrocardio-activity waveform feature to be rebuilt and the millimeter wave radar sample signal feature in the shared feature space; The target loss is determined based on the morphological fidelity loss, the consistency loss, the antagonism loss, and the content loss.
4. A mental state estimation parameter determination method according to any one of claims 1 to 3, wherein the micro-motion feature extraction module is specifically configured to: Filtering the circadian rhythm of the millimeter wave radar signal to obtain a micro-motion residual signal; Inputting the micro-motion residual signal into a time sequence convolution network to obtain an initial time sequence characteristic output by the time sequence convolution network; applying a time sequence attention weight and/or a channel attention weight to the initial time sequence feature to obtain an enhanced time sequence feature; the micro-motion feature is determined based on the enhanced timing feature.
5. A psychological state assessment parameter determining method according to any one of claims 1 to 3, characterized in that the acquiring of millimeter wave radar signals of a target object to be assessed comprises: acquiring a multi-channel millimeter wave radar signal of the target object to be evaluated; respectively determining normalized reflection energy, in-phase and quadrature component diagram circular fitting errors, heartbeat signal periodicity scores, respiratory signal harmonic interference ratios and phase standard deviations of each channel signal in the multichannel millimeter wave radar signals; determining a composite signal quality index based on at least two of the normalized reflected energy, the in-phase-quadrature component map circular fitting error, the heartbeat signal periodicity score, the respiratory signal harmonic to interference ratio, and the phase standard deviation; And selecting a target channel signal from the multi-channel millimeter wave radar signals as the millimeter wave radar signal based on the comprehensive signal quality index.
6. The psychological state assessment parameter determining method according to claim 5, wherein selecting a target channel signal from the multi-channel millimeter wave radar signals as the millimeter wave radar signal based on the integrated signal quality index, comprises: Based on the comprehensive signal quality index, performing spatial filtering processing on the multi-channel millimeter wave radar signal to obtain a spatial filtering signal; performing source signal separation processing on the spatial filtering signals to obtain separated radar signals; Carrying out phase unwrapping processing on the separated radar signals to obtain continuous phase signals, wherein an unwrapping path of the phase unwrapping processing is guided based on a phase quality diagram of the separated radar signals; And carrying out multi-resolution separation on the continuous phase signals to separate out target channel signals related to heart activities, and taking the target channel signals as the millimeter wave radar signals.
7. A psychological state assessment parameter determining apparatus, comprising: The acquisition unit is used for acquiring millimeter wave radar signals of the target object to be evaluated; the psychological state evaluation unit is used for inputting the millimeter wave radar signals into a psychological state evaluation model to obtain psychological state evaluation parameters output by the psychological state evaluation model; the psychological state assessment model comprises a physiological characteristic extraction module, a micro-motion characteristic extraction module, a multi-modal characteristic fusion module and a psychological state inference module; The system comprises a physiological characteristic extraction module, a multi-mode characteristic fusion module, a psychological state inference module and a psychological state estimation module, wherein the physiological characteristic extraction module is used for reconstructing an electrocardio-activity waveform based on the millimeter wave radar signal and extracting physiological characteristics from the electrocardio-activity waveform, the micro-motion characteristic extraction module is used for extracting micro-motion characteristics of the millimeter wave radar signal to obtain micro-motion characteristics, the multi-mode characteristic fusion module is used for fusing the physiological characteristics and the micro-motion characteristics to obtain fusion characteristics, the psychological state inference module is used for obtaining the psychological state estimation parameter based on the fusion characteristics, and the electrocardio-activity waveform is a time sequence which is reconstructed from the millimeter wave radar signal and highly simulates a true solid electric activity signal in waveform form and rhythm; The training device further comprises a training unit, and the training unit is specifically used for: Acquiring a millimeter wave radar sample signal and a reference electrocardio activity waveform synchronously acquired with the millimeter wave radar sample signal; Inputting the millimeter wave radar sample signal to an initial physiological characteristic extraction module to obtain an electrocardio-activity waveform to be rebuilt, which is output by the initial physiological characteristic extraction module; Determining diagnosis weights corresponding to different waveform segments in the reference electrocardio-activity waveform according to clinical diagnosis importance of the different waveform segments, calculating weighted morphology differences between the electrocardio-activity waveform to be reconstructed and the reference electrocardio-activity waveform based on the diagnosis weights, and determining morphological fidelity loss; And determining target loss based on the morphological fidelity loss, and carrying out parameter iteration on the initial physiological characteristic extraction module based on the target loss to obtain the physiological characteristic extraction module.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and running on the processor, characterized in that the processor implements the mental state estimation parameter determination method according to any one of claims 1 to 6 when executing the computer program.
9. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the mental state estimation parameter determination method according to any one of claims 1 to 6.

Description

Psychological state evaluation parameter determining method, psychological state evaluation parameter determining device, electronic equipment and storage medium Technical Field The present invention relates to the field of non-contact biological signal processing technologies, and in particular, to a method and apparatus for determining a psychological state evaluation parameter, an electronic device, and a storage medium. Background Mental health has become a major public health problem of global concern, and timely and accurate mental state assessment is critical for early warning, diagnosis and intervention. Traditional mental state assessment methods, such as subjective scales and clinical interviews, have strong subjectivity, recall bias, low assessment frequency, and other limitations, although having application value. Therefore, a non-contact and objective physiological index monitoring technology becomes an important development direction. The millimeter wave radar technology has great potential in the field of physiological signal monitoring due to the advantages of non-contact detection, clothing penetrability, no influence of illumination, no invasion of visual privacy and the like. The prior art has utilized millimeter wave radars to monitor physiological parameters such as heart rate, respiration rate, etc., and has been studied in an attempt to analyze heart rate variability (HEART RATE Variability, HRV) therefrom. However, when the conventional psychological state assessment technical scheme based on the millimeter wave radar is applied to unconstrained daily environments, the prior art has the defects that firstly, the accuracy of the physiological signals is insufficient, and most of the physiological signals are limited to extracting macroscopic physiological parameters such as average heart rate and respiratory rate, so that the physiological information dimension on which the assessment is based is insufficient. Second, in prior art schemes, these micro-motion signals, which contain important psychological information, are often ignored or filtered out as noise or interference, resulting in the loss of effective information. Finally, the existing evaluation model generally analyzes only single-mode circadian rhythm information, and lacks a framework capable of comprehensively analyzing the characteristics reflecting internal physiological changes and micro-motion characteristics reflecting external behavior, so that the evaluation result is not comprehensive and accurate enough. Disclosure of Invention The invention provides a psychological state assessment parameter determining method, a psychological state assessment parameter determining device, electronic equipment and a storage medium, which are used for solving the defects that in the prior art, the extraction fineness of physiological signals relied on by psychological state assessment is insufficient, micro-motion signals containing important information are generally regarded as noise filtering, so that effective information is lost, in addition, the existing scheme is mostly based on physiological information of a single mode to analyze, and a comprehensive assessment framework capable of cooperatively utilizing internal physiological and external behavioral characteristics is lacking, so that the assessment comprehensiveness and accuracy are limited. The invention provides a psychological state assessment parameter determining method, which comprises the following steps: Acquiring a millimeter wave radar signal of a target object to be evaluated; inputting the millimeter wave radar signal into a psychological state assessment model to obtain psychological state assessment parameters output by the psychological state assessment model; the psychological state assessment model comprises a physiological characteristic extraction module, a micro-motion characteristic extraction module, a multi-modal characteristic fusion module and a psychological state inference module; The physiological characteristic extraction module is used for reconstructing an electrocardio activity waveform based on the millimeter wave radar signal and extracting physiological characteristics from the electrocardio activity waveform, the micro-motion characteristic extraction module is used for extracting micro-motion characteristics of the millimeter wave radar signal to obtain micro-motion characteristics, the multi-mode characteristic fusion module is used for fusing the physiological characteristics and the micro-motion characteristics to obtain fusion characteristics, and the psychological state inference module is used for obtaining the psychological state assessment parameters based on the fusion characteristics. According to the method for determining the psychological state assessment parameters provided by the invention, the training step of the physiological characteristic extraction module comprises the following steps: Acquiring a millimeter