Psychological scientific analysis method based on face and limb micro-action recognition

Abstract

The invention discloses a psychological scientific analysis method based on face and limb micro-action recognition, and relates to the technical field of psychological scientific analysis, and the method comprises the following steps: firstly, a laser interference measurement system emits a light beam to a target eyeball area at an initially set laser emission frequency, and obtains an original reflection signal in real time through a photoelectric detector; and basic data is provided for subsequent feature extraction and dynamic adjustment. Initial laser interference measurement and a self-adaptive dynamic modulation strategy are adopted, the microseismic frequency of the eyeball is accurately matched, and the detection sensitivity and stability of high-frequency and micro-amplitude movement are improved. Through key index quantitative analysis, machine learning matching judgment and eyeball physiological region optical feedback, real-time self-adaptive adjustment of laser emission frequency and measurement focus is realized, and the problem that microseism is difficult to capture by fixed-frequency laser is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of psychological science analysis, and particularly to a psychological science analysis method based on the recognition of micro-movements of the face and limbs. Background Art

[0002] Psychological science analysis based on the recognition of micro-movements of the face and limbs refers to inferring an individual's true psychological state, emotional changes, and potential intentions by analyzing micro-expressions and subtle limb movements (such as gestures, body postures, muscle tension, etc.) shown by the individual in an unconscious state. This research field relies on psychology, neuroscience, and artificial intelligence technologies, believing that human micro-movements are often difficult to be controlled voluntarily, so they can reveal true emotions and cognitive activities better than verbal expressions. For example, in a stressful or deceptive situation, an individual may show brief facial twitches, finger tremors, and subconscious defensive gestures, which are all external manifestations of the individual's emotional conflict. Psychological science analysis of micro-movement recognition can be applied to fields such as lie detection, mental health assessment, emotion perception, and social interaction research, helping to more accurately interpret an individual's psychological state and even predict behavioral trends. In recent years, automated micro-movement analysis combining computer vision and deep learning has made psychological research more quantitative and objective, providing a scientific basis and technical support for scenarios such as psychological diagnosis and human-computer interaction.

[0003] Laser interferometry plays a role in high-precision and non-contact monitoring in the recognition of an individual's facial and limb micro-movements. This technology uses a laser beam to irradiate the surface of an individual's face or limb, and detects minute displacements or vibrations by analyzing the phase changes of the reflected light, and even nano-scale movements can be accurately captured. For example, when an individual is in a state of tension, anxiety, or deception, the facial muscles will undergo imperceptible contractions, and there may be slight tremors at the corners of the mouth, eyelids, or jaw, which are difficult to detect by traditional cameras or the naked eye; laser interferometry can accurately track these changes and then analyze the individual's emotional fluctuations. In addition, this technology can also be applied to the recognition of limb micro-movements, such as slight tremors of fingers, toes, and shoulders, and can even detect respiratory patterns and pulse fluctuations caused by psychological activities. Due to its characteristics of being remote, non-invasive, and high-resolution, laser interferometry has great potential in fields such as lie detection, emotion analysis, remote health monitoring, and behavior recognition.

[0004] The prior art has the following deficiencies: Laser Interferometry in the prior art usually uses a laser beam with a fixed frequency (i.e., a stable wavelength) to ensure the high precision and stability of interferometric measurement. This fixed-frequency laser can provide a highly coherent light source, enabling the interferometer to accurately detect minute displacements of the target surface (such as skin, muscle). However, Micro-saccades are extremely tiny but crucial movements for cognitive states (such as attention, deceptive behavior), characterized by high frequency (usually 80 - 120 Hz) and tiny amplitude (micrometer to nanometer scale). Due to their fast movement speed and extremely small amplitude, conventional fixed-frequency lasers may have difficulty capturing these high-speed and subtle eye movements.

[0005] If laser interferometry cannot adaptively match Micro-saccades in micro-motion recognition, it may seriously affect high-precision lie detection. Research shows that the frequency and trajectory of Micro-saccades change in specific patterns when an individual is lying or under high pressure.

[0006] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0007] The object of the present invention is to provide a psychological science analysis method for micro-motion recognition based on facial and limb movements, which adopts an initial laser interferometry and an adaptive dynamic modulation strategy to accurately match the Micro-saccades frequency, and improve the detection sensitivity and stability of high-frequency and micro-amplitude movements. Through quantitative analysis of key indicators, machine learning matching judgment, and optical feedback of the physiological area of the eyeball, real-time adaptive adjustment of the laser emission frequency and the measurement focus is achieved, overcoming the problem that fixed-frequency lasers have difficulty capturing micro-vibrations, so as to solve the problems in the above background art.

[0008] To achieve the above object, the present invention provides the following technical solution: A psychological science analysis method for micro-motion recognition based on facial and limb movements, comprising the following steps:

[0009] First, the laser interferometry system emits a light beam to the target eyeball area at an initially set laser emission frequency, and the original reflection signal is obtained in real time through a photodetector, providing basic data for subsequent feature extraction and dynamic adjustment;

[0010] After obtaining the original reflection signal, the obtained original reflection signal is preprocessed, and key indicators reflecting the mismatch between the laser frequency and the Micro-saccades frequency are extracted from the preprocessed data, and the extracted key indicators are analyzed to quantify the degree of mismatch between the laser emission frequency and the Micro-saccades frequency;

[0011] Input the key metrics after analysis into a pre-trained machine learning model. Through the machine learning model, identify the matching degree between the current laser frequency and the microvibration natural frequency, and determine whether there is a mismatch between the laser frequency and the eye microvibration frequency.

[0012] When the machine learning model determines that there is a mismatch between the current laser frequency and the eye microvibration frequency, based on the spectral information of the eye microvibration through an adaptive regulation algorithm, dynamically modulate the laser emission frequency through the eye movement vibration characteristics captured in real time, and at the same time adjust the laser measurement position, intelligently explore the optical reflection characteristics of different physiological regions of the eye, and dynamically switch the measurement focus according to the real-time feature distribution of the eye movement.

[0013] Preferably, the laser interferometry system emits a light beam to the target eye area at the initially set laser emission frequency, and the specific steps to obtain the original reflection signal in real time through a photodetector are as follows:

[0014] First, initialize the laser interferometry system and set the initial laser emission frequency;

[0015] Secondly, the laser emits a stable and highly coherent light beam at this set frequency and precisely focuses it on the target eye area;

[0016] Next, the photodetector captures the reflection signal on the target surface in real time; then, this reflection signal is transmitted to the data processing module to provide the original data for subsequent filtering, denoising, and time-frequency analysis.

[0017] Finally, these data are stored and organized to form a multi-dimensional feature set, laying a foundation for subsequent feature extraction and dynamic adjustment.

[0018] Preferably, extract the key metrics reflecting the mismatch between the laser frequency and the eye microvibration frequency from the preprocessed data. The extracted key metrics include the amplitude change of the modulation signal calculated on multiple time scales and the dynamic characteristics of the microvibration signal remaining stable within a fixed time. Analyze the amplitude change of the modulation signal calculated on multiple time scales and the dynamic characteristics of the microvibration signal remaining stable within a fixed time under the detection window to generate the multi-scale modulation intensity reference value and the eye microvibration non-steady state ratio reference value respectively, and quantify the mismatch degree between the laser emission frequency and the eye microvibration frequency through the multi-scale modulation intensity reference value and the eye microvibration non-steady state ratio reference value.

[0019] Preferably, first perform multi-scale decomposition on the obtained interference reflection signal, analyze the amplitude change of the modulation signal on different time scales, and introduce the non-linear energy distribution factor and the signal envelope complexity ratio as feature metrics;

[0020] The non-linear energy distribution factor is used to measure the distribution of the interference signal energy on different scales, and the calculation expression is as follows:

[0021]

[0022] , where is the modulation signal intensity at the i-th moment under the scale λ, is the modulation signal intensity at the (i - 1)-th moment under the scale λ, that is, the modulation signal intensity at the previous moment, β is the adaptive non-linear weight factor, and Eλ is the non-linear energy distribution factor;

[0023] The signal envelope complexity ratio is used to characterize the non-linear complexity of the signal envelope at different scales, and the calculation expression is as follows:

[0024]

[0025] , where is the modulation signal intensity at the j-th moment under the scale λ, β is the dynamically adjusted envelope characteristic index, and C μ is the signal envelope complexity ratio;

[0026] After obtaining the non-linear energy distribution factor E λ and the signal envelope complexity ratio C μ , a multi-scale modulation intensity reference value is generated as the final matching evaluation index, and the generation formula is as follows:

[0027]

[0028] , where MSMI is the multi-scale modulation intensity reference value, Γ λ is the scale sensitivity factor, γ is the modulation index control parameter, Λ λ is the non-linear modulation weight, and δ is the envelope complexity control parameter.

[0029] Preferably, the specific steps for analyzing the stable dynamic characteristics of the microseismic signal within a fixed time under the detection window to generate the reference value of the non-steady state ratio of the ocular microseismic are as follows:

[0030] First, perform non-steady state characteristic analysis on the microseismic signal to evaluate its dynamic change trend over time, and introduce the discrete spectrum offset degree and the instantaneous modulation inflation factor to measure the non-steady state degree of the microseismic signal;

[0031] The discrete spectrum offset degree is used to measure the offset degree of the spectrum center of the microseismic signal from its previous state at a fixed time, and the calculation expression is as follows:

[0032]

[0033] , where DSS is the discrete spectrum offset degree, and f q is the frequency value of the q-th frequency component, and Fref is the center frequency reference value of the previous detection window, P q is the power spectral density of the q-th frequency component;

[0034] The instantaneous modulation expansion factor, which is used to measure the degree of change in the modulation mode of the microseismic signal in a short time, reflects the expansion trend of the non-steady signal, and the calculation expression is as follows:

[0035]

[0036] , where IMEF is the instantaneous modulation expansion factor, M k is the instantaneous modulation amplitude at the k-th moment, M k-1 is the instantaneous modulation amplitude at the previous moment, and ∈ is a small constant to prevent the denominator from being zero;

[0037] After constructing the discrete spectrum shift degree DSS and the instantaneous modulation expansion factor IMEF, generate the reference value of the non-steady ratio of the eye microseismic, comprehensively evaluate the matching degree between the laser frequency and the microseismic signal frequency, and the generation formula is as follows:

[0038]

[0039] , where MSNSR is the reference value of the non-steady ratio of the eye microseismic, ω is the weight coefficient of the discrete spectrum shift degree DSS, is the weight coefficient of the instantaneous modulation expansion factor IMEF, τ is the interaction weight of the discrete spectrum shift degree DSS and the instantaneous modulation expansion factor IMEF, and σ is a small constant to prevent the denominator from being zero.

[0040] Preferably, input the analyzed multi-scale modulation intensity reference value and the reference value of the non-steady ratio of the eye microseismic into a pre-trained machine learning model, generate a frequency mismatch risk coefficient through the machine learning model, and identify the matching degree between the current laser frequency and the microseismic natural frequency through the frequency mismatch risk coefficient, and judge whether there is a mismatch between the laser frequency and the eye microseismic frequency.

[0041] Preferably, compare and analyze the frequency mismatch risk coefficient generated when identifying the matching degree between the current laser frequency and the microseismic natural frequency through a pre-trained machine learning model with a pre-set frequency mismatch risk coefficient reference threshold to judge whether there is a mismatch between the laser frequency and the eye microseismic frequency, and the judgment process is as follows:

[0042] If the frequency mismatch risk coefficient is greater than the pre-set frequency mismatch risk coefficient reference threshold, it is judged that there is a mismatch between the current laser frequency and the eye microseismic frequency; if the frequency mismatch risk coefficient is less than or equal to the pre-set frequency mismatch risk coefficient reference threshold, it is judged that there is no mismatch between the current laser frequency and the eye microseismic frequency.

[0043] Preferably, when the machine learning model determines that there is a mismatch between the current laser frequency and the eye micro-vibration frequency, the specific steps of dynamically modulating the laser emission frequency and adjusting the laser measurement position through the adaptive control algorithm are as follows:

[0044] When it is detected that there is a mismatch between the current laser frequency and the eye micro-vibration frequency, first update the laser emission frequency through the adaptive control algorithm. Let the current laser frequency be f L , and use the real-time captured eye movement vibration spectrum characteristics and the frequency mismatch risk coefficient FMRC, combined with the preset frequency mismatch risk reference threshold to calculate the frequency modulation amount. The calculation formula is as follows:

[0045]

[0046] , where f L ′ is the adjusted laser emission frequency, f L is the current laser emission frequency, θ is the control gain parameter, FMRC ref is the frequency mismatch risk coefficient reference threshold, S E is the eye micro-vibration spectrum characteristic parameter;

[0047] After the laser frequency is modulated, to optimize signal acquisition, the laser measurement area and the focus position will be intelligently adjusted according to the real-time eye movement tracking data and the optical feedback information. Let the current laser measurement position be P L and the current laser focus position be F L , and at the same time introduce the regional optical reflection characteristic index to jointly adjust the measurement position and the focus. The adjustment formula is as follows:

[0048]

[0049] , where P L ′ is the updated laser measurement position, P L is the current laser measurement position, is the position adjustment gain parameter, R O is the real-time regional optical reflection characteristic, is the preset best regional reflection reference value, F L ′ is the updated laser focus position, F L is the current laser focus position, μ is the focus adjustment gain parameter, Δ E is the focus error caused by eye movement.

[0050] In the above technical solution, the technical effects and advantages provided by the present invention:

[0051] The present invention adopts an initial laser interference measurement and an adaptive dynamic modulation strategy, which can match the frequency characteristics of eye micro-vibrations in real time and accurately, and significantly improve the detection sensitivity and stability of the laser interference system for such high-frequency and micro-amplitude movements as eye micro-vibrations. By introducing the quantitative analysis of key indicators after preprocessing and the intelligent matching judgment of machine learning, and combining the eye movement vibration characteristics and the optical feedback information of the eye physiological region, the present invention realizes the efficient, real-time and adaptive adjustment of the laser emission frequency and the measurement focus, effectively overcoming the problem that the fixed-frequency laser in the prior art is difficult to capture eye micro-vibrations. Therefore, the present invention can greatly improve the accuracy of lie detection and has important application value and social significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0053] Figure 1 It is a method flow chart of the psychological science analysis method for micro-action recognition based on the face and limbs of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that this disclosure will be more complete and comprehensive, and will fully convey the concept of the example embodiments to those skilled in the art.

[0055] The present invention provides a psychological science analysis method for micro-action recognition based on the face and limbs as shown in Figure 1 the following, including the following steps:

[0056] First, the laser interference measurement system emits a light beam to the target eye area at an initially set laser emission frequency, and the original reflection signal is obtained in real time through a photodetector, providing basic data for subsequent feature extraction and dynamic adjustment;

[0057] The initial frequency is usually set according to experience or prior information (such as the common human physiological vibration range). When the laser irradiates the eye surface, part of the light is reflected back and collected by a photodetector. The key function of this step is to obtain the original interference signal, providing basic data for subsequent feature extraction and dynamic adjustment. If the initial frequency is reasonably selected, it can capture the eye micro-vibration signal in general cases to a large extent; if there is a large mismatch, it can also provide a baseline reference for the next adjustment.

[0058] The laser interferometry system emits a light beam towards the target eye region at an initially set laser emission frequency, and a raw reflection signal is acquired in real time by a photodetector. The specific steps are as follows:

[0059] First, the laser interferometry system is initialized and the initial laser emission frequency is set. Second, the laser emits a stable and highly coherent light beam at this set frequency, which is precisely focused on the target eye region. Then, the photodetector captures the reflection signal on the target surface in real time. Next, this reflection signal is transmitted to the data processing module to provide raw data for subsequent filtering, denoising, and time-frequency analysis. Finally, these data are stored and organized to form a multi-dimensional feature set, laying a foundation for subsequent feature extraction and dynamic adjustment.

[0060] After the raw reflection signal is acquired, the acquired raw reflection signal is preprocessed, and key indicators reflecting the mismatch between the laser frequency and the eye micro-vibration frequency are extracted from the preprocessed data. The extracted key indicators are analyzed to quantify the degree of mismatch between the laser emission frequency and the eye micro-vibration frequency.

[0061] First, a filtering algorithm (such as low-pass filtering, high-pass filtering, or adaptive filtering) is applied to remove background noise and high-frequency interference. Then, a denoising algorithm (such as wavelet denoising, Kalman filtering) is used to retain as much detail of the eye micro-vibration information as possible. Subsequently, time-frequency analysis (Time-Frequency Analysis), such as short-time Fourier transform (STFT) or continuous wavelet transform (CWT), is performed to map the signal from the time domain to the time-frequency domain to more clearly observe the energy distribution of the micro-vibration over time. Through this preprocessing step, the signal-to-noise ratio of the data can be significantly improved, helping to accurately extract key features related to eye micro-vibration in subsequent steps.

[0062] Key indicators reflecting the mismatch between the laser frequency and the eye micro-vibration frequency are extracted from the preprocessed data. The extracted key indicators include the amplitude change of the modulation signal calculated on multiple time scales and the dynamic characteristics of the micro-vibration signal remaining stable within a fixed time. The amplitude change of the modulation signal calculated on multiple time scales and the dynamic characteristics of the micro-vibration signal remaining stable within a fixed time are analyzed under the detection window to generate a multi-scale modulation intensity reference value and an eye micro-vibration non-steady state ratio reference value respectively. The degree of mismatch between the laser emission frequency and the eye micro-vibration frequency is quantified through the multi-scale modulation intensity reference value and the eye micro-vibration non-steady state ratio reference value.

[0063] The distribution of the modulation signal intensity exhibits extremely violent oscillations at different scales, indicating a mismatch between the current laser frequency and the eyeball micro-vibration frequency. The reason is that laser interferometric measurement relies on the resonance matching between the laser frequency and the target micro-vibration frequency to form a stable interference signal. When there is a mismatch between the two, the system cannot effectively capture the inherent vibration characteristics of the eyeball micro-vibration, resulting in the non-stationarity of the energy distribution of the interference signal at different time scales, that is, the modulation intensity at different scales will exhibit extremely violent oscillations. Specifically, when the laser frequency deviates from the inherent frequency of the micro-vibration, the main energy components in the interference signal will continuously shift between multiple frequency bands, causing the modulation intensity to fluctuate significantly in a short period of time, and may even trigger frequency hopping or non-linear aliasing effects. In addition, due to the time-varying characteristics of the eyeball micro-vibration itself, if the laser frequency cannot adapt to its dynamic changes, the system will be subject to more noise interference, resulting in the instability of the modulation structure of the interference signal. Therefore, the violent oscillations of the modulation signal intensity at different scales are an important indication of the mismatch between the laser frequency and the eyeball micro-vibration frequency. It reflects that the measurement system fails to correctly track the vibration mode of the target micro-vibration and requires real-time frequency adjustment through an adaptive control algorithm to optimize the quality of the interference signal.

[0064] The specific steps for analyzing the amplitude change of the modulation signal calculated at multiple time scales under the detection window to generate the multi-scale modulation intensity reference value are as follows:

[0065] First, perform multi-scale decomposition on the acquired interference reflection signal, analyze the amplitude change of the modulation signal at different time scales, and introduce the non-linear energy distribution factor and the signal envelope complexity ratio as characteristic indicators;

[0066] The non-linear energy distribution factor is used to measure the distribution of the interference signal energy at different scales, and the calculation expression is as follows:

[0067]

[0068] , where is the modulation signal intensity at the i-th moment at scale λ, is the modulation signal intensity at the (i - 1)-th moment at scale λ, that is, the modulation signal intensity at the previous moment, β is the adaptive non-linear weight factor used to enhance the change trend of the energy between scales, and E λ is the non-linear energy distribution factor;

[0069] When E λ is too large, it indicates that there are violent fluctuations in the energy distribution at different time scales, indicating a mismatch between the current laser frequency and the eyeball micro-vibration frequency.

[0070] The signal envelope complexity ratio is used to characterize the non-linear complexity of the signal envelope at different scales, and the calculation expression is as follows:

[0071]

[0072] , where is the modulation signal intensity at the j-th moment under the scale λ, β is the envelope characteristic index adjusted dynamically, controlling the asymmetry of the signal complexity, and C μ is the signal envelope complexity ratio;

[0073] If C μ abnormally increases in value, it indicates that the local envelope complexity of the signal has changed significantly, meaning that the stability of the interference signal is damaged, further reflecting the matching error between the laser frequency and the eye micro-vibration frequency.

[0074] After obtaining the non-linear energy distribution factor E λ and the signal envelope complexity ratio C μ , a multi-scale modulation intensity reference value is generated as the final matching evaluation index, and the generation formula is as follows:

[0075]

[0076] , where MSMI is the multi-scale modulation intensity reference value, Γ λ is the scale sensitivity factor, used to weigh the contribution of different time scales to the modulation intensity, γ is the modulation index control parameter, used for the contribution weight of the non-linear energy distribution factor E λ in the final multi-scale modulation intensity reference value, Λ λ is the non-linear modulation weight, used to enhance the contribution of the local complexity of the signal at a specific time scale, and δ is the envelope complexity control parameter, used to adjust the contribution ratio of the signal envelope complexity ratio C μ in the scale modulation intensity reference value.

[0077] By integrating the characteristic parameters (non-linear energy distribution factor E λ and signal envelope complexity ratio C μ ) extracted at different time scales, the multi-scale modulation intensity reference value MSMI is calculated to quantitatively evaluate the matching degree between the current laser frequency and the eye micro-vibration frequency. When MSMI abnormally increases, it indicates that the modulation intensity fluctuates violently at different time scales, meaning that the laser frequency and the eye micro-vibration frequency are mismatched, providing a decision basis for the adaptive adjustment of the laser emission frequency.

[0078] The larger the multi-scale modulation intensity reference value generated by analyzing the amplitude change of the modulation signal calculated on multiple time scales under the detection window, the greater the mismatch between the current laser frequency and the eyeball micro-vibration frequency. When the reference value is small or stable, it indicates a good match between the two. The principle lies in that when the laser frequency is precisely matched with the micro-vibration natural frequency, the modulation energy in the interference signal will be concentrated in a specific stable frequency band and maintain a relatively stable energy distribution on multiple time scales, resulting in a lower multi-scale modulation intensity reference value. If there is a mismatch between the laser frequency and the micro-vibration frequency, the main modulation components in the interference signal will be randomly dispersed between different frequency bands and be affected by environmental noise and measurement errors, causing instability in the modulation intensity at different scales, thus increasing the multi-scale modulation intensity reference value. In other words, the drastic change in the multi-scale modulation intensity reference value reflects the degree of non-linear coupling between the laser frequency and the target micro-vibration frequency. The higher the reference value, the more unstable the modulation intensity of the system at different time scales, indicating that the current measurement frequency cannot effectively capture the true characteristics of the eyeball micro-vibration.

[0079] The micro-vibration signal steadily decreases within a fixed time while the fluctuation increases in a short time, indicating a mismatch between the current laser frequency and the eyeball micro-vibration frequency. This phenomenon shows that the system fails to accurately capture the natural frequency of the micro-vibration, resulting in a decrease in the coherence of the interference signal and affecting the stability of the measurement. When the laser frequency matches the micro-vibration frequency, the reflected signal should exhibit relatively stable periodic fluctuations and the energy distribution should remain consistent within a specific frequency band. However, when a mismatch occurs, the energy of the micro-vibration signal will partially or significantly attenuate, leading to an overall decrease in the signal on a longer time scale. At the same time, due to the failure to accurately lock the micro-vibration frequency, the measurement fluctuation in a short time will increase, showing characteristics such as enhanced instability, spectral drift, and increased noise. This increase in fluctuation in a short time may be due to the system attempting to capture the micro-vibration characteristics near multiple incorrect frequencies, causing the signal to jump between different frequencies instead of stably focusing on the true dynamic characteristics of the eyeball micro-vibration. Therefore, this phenomenon can be used as a key indicator of the mismatch between the laser emission frequency and the micro-vibration natural frequency and can be used to dynamically adjust the system to make the laser frequency more accurately match the micro-vibration characteristics to optimize the measurement accuracy and stability.

[0080] The specific steps to generate the eyeball micro-vibration non-steady state ratio reference value by analyzing the dynamic characteristics of the micro-vibration signal remaining stable within a fixed time under the detection window are as follows:

[0081] First, perform a non-steady state characteristic analysis on the micro-vibration signal to evaluate its dynamic change trend over time. Since the eyeball micro-vibration signal is a complex non-linear system, its stability in the time series can be quantified by the degree of dispersion of the energy distribution and the characteristics of the instantaneous frequency change; introduce the discrete spectrum offset degree and the instantaneous modulation inflation factor to measure the non-steady state degree of the micro-vibration signal;

[0082] The discrete spectrum shift degree is used to measure the degree of shift of the spectrum center of the microseismic signal with respect to its previous state at a fixed time. The calculation expression is as follows:

[0083]

[0084] , where DSS is the discrete spectrum shift degree, and f q is the frequency value of the q-th frequency component, and F ref is the center frequency reference value of the previous detection window, and P q is the power spectral density of the q-th frequency component;

[0085] The larger the discrete spectrum shift degree DSS is, it indicates that the main frequency of the microseismic signal has a large drift within the window, suggesting that the matching degree between the current laser frequency and the eye microseismic decreases, and the interference effect of the system deteriorates.

[0086] The instantaneous modulation expansion factor is used to measure the degree of change in the modulation mode of the microseismic signal in a short time, reflecting the expansion trend of the non-steady signal. The calculation expression is as follows:

[0087]

[0088] , where IMEF is the instantaneous modulation expansion factor, and M k is the instantaneous modulation amplitude at the k-th moment, and M k-1 is the instantaneous modulation amplitude at the previous moment, and ∈ is a small constant to prevent the denominator from being zero;

[0089] The larger the instantaneous modulation expansion factor is, it indicates that the modulation mode of the microseismic signal has a sudden change in a short time, suggesting that the instability of the microseismic signal increases, and the tracking accuracy of the system for eye microseismic decreases.

[0090] After constructing the discrete spectrum shift degree DSS and the instantaneous modulation expansion factor IMEF, an eye microseismic non-steady ratio reference value is generated to comprehensively evaluate the matching degree between the laser frequency and the microseismic signal frequency. The generation formula is as follows:

[0091]

[0092] , where MSNSR is the eye microseismic non-steady ratio reference value, ω is the weight coefficient of the discrete spectrum shift degree DSS, which determines the influence weight of the discrete spectrum shift degree DSS in the calculation of MSNSR, is the weight coefficient of the instantaneous modulation expansion factor IMEF, which determines the influence of the instantaneous modulation expansion factor IMEF in the MSNSR calculation. τ is the interaction weight of the discrete spectrum offset DSS and the instantaneous modulation expansion factor IMEF, which controls the interaction effect between the discrete spectrum offset DSS and the instantaneous modulation expansion factor IMEF, that is, when the spectrum offset and modulation expansion occur at the same time, their influence on MSNSR. σ is a small constant to prevent the denominator from being zero, usually 10-6.

[0093] MSNSR weightedly fuses the discrete spectrum shift DSS and the instantaneous modulation expansion factor IMEF, and introduces an additional interaction term to enable it to take into account both the overall spectrum shift and the short-term modulation mutation.

[0094] If the discrete spectrum deviation DSS and the instantaneous modulation expansion factor IMEF have large values, it means that the current laser frequency is not well matched with the eye microtremor frequency, and the stability of the system interference signal is reduced.

[0095] If the MSNSR performance value is small, it means that the steady state of the microseismic signal is high, which means that the current laser frequency has well matched the microseismic frequency of the eye and the interference measurement effect is better.

[0096] The larger the reference value of the non-steady-state ratio of eye microseisms generated by analyzing the dynamic characteristics of the microseismic signal that remains stable within a fixed time under the detection window, the greater the mismatch between the current laser frequency and the eye microseismic frequency, and vice versa. The non-steady-state ratio reference value of eye microseisms evaluates its stability over time by monitoring the microseismic signal in the window, that is, whether it maintains consistent dynamic characteristics. When the laser frequency matches the eye microseismic frequency, the microseismic signal should present a stable pattern, with relatively consistent time evolution characteristics and a small fluctuation amplitude, resulting in a low performance value of the non-steady-state ratio reference value of the eye microseismic. However, when a mismatch occurs, the laser signal fails to accurately lock the true frequency of the microseismic signal, resulting in unstable signals in the interferometric measurement, and the occurrence of random drift, spectral energy discreteness, and short-term mutation enhancement. The increase in these non-steady-state characteristics will cause the non-steady-state ratio reference value of the eye microseismic to increase, indicating that there is a significant deviation between the current laser frequency and the eye microseismic frequency.

[0097] The analyzed key indicators are input into a pre-trained machine learning model, and the matching degree between the current laser frequency and the natural frequency of microseismicity is identified through the machine learning model to determine whether there is a mismatch between the laser frequency and the eye microseismic frequency;

[0098] Input the multi-scale modulation intensity reference value and the eyeball micro-vibration non-steady state ratio reference value after analysis into a pre-trained machine learning model. Generate a frequency mismatch risk coefficient through the machine learning model, and identify the matching degree between the current laser frequency and the micro-vibration natural frequency through the frequency mismatch risk coefficient, and determine whether there is a mismatch between the laser frequency and the eyeball micro-vibration frequency.

[0099] The pre-trained machine learning model refers to a model that has been learned and optimized through a large amount of historical data before the system is officially run to have the ability of accurate prediction and classification. A large number of labeled data sets are used in the training stage of the model, including various situations of the matching and mismatch between the laser frequency and the eyeball micro-vibration frequency, and through supervised learning, deep learning or reinforcement learning methods, learn the patterns and features of these data. During the training process, the model will continuously optimize its internal parameters (such as weights, biases, activation functions, etc.) so that when new test data is input, it can efficiently and accurately predict whether the current laser frequency matches the eyeball micro-vibration natural frequency. This training process usually involves large-scale computing, cross-validation, and hyperparameter optimization to ensure that the model still has high generalization ability and high accuracy in a complex environment. Once the model is trained and verified, it can be deployed as an inference model into the actual system to process the input data in real time and give the corresponding matching degree evaluation result.

[0100] In this system, the pre-trained machine learning model will receive feature inputs such as the multi-scale modulation intensity reference value and the eyeball micro-vibration non-steady state ratio reference value, and combine other auxiliary features (such as the interference phase drift index, the harmonic distortion index, etc.), and use algorithms such as deep neural network (DNN), support vector machine (SVM) or random forest (RF) to generate a frequency mismatch risk coefficient. This risk coefficient represents the matching degree between the current laser frequency and the micro-vibration natural frequency. If the coefficient is high, it means there is an obvious frequency mismatch, and the system needs to execute an adaptive regulation mechanism to adjust the laser emission frequency. If the risk coefficient is low, it indicates that the current laser frequency and the eyeball micro-vibration frequency are highly matched, and the measurement result is stable and reliable. In this way, the pre-trained machine learning model can identify the relationship between the laser frequency and the micro-vibration natural frequency in real time and efficiently, avoid the limitations of traditional rule-based or manual adjustment, and improve the accuracy and intelligence level of micro-motion detection.

[0101] The machine learning model is not limited here. Any machine learning model that can comprehensively analyze the multi-scale modulation intensity reference value MSMI and the eyeball micro-vibration non-steady state ratio reference value MSNSR to generate a frequency mismatch risk coefficient FMRC can be used. To implement the technical solution of the present invention, the present invention provides a specific implementation method;

[0102] The formula for generating the Frequency Mismatch Risk Coefficient (FMRC) is as follows: FMRC = k1·MSMI + k2·MSNSR, where k1 and k2 are the preset proportionality coefficients of the Multi-Scale Modulation Intensity Reference (MSMI) and the Micro-Saccade Non-Steady State Ratio Reference (MSNSR) respectively, and both k1 and k2 are greater than 0.

[0103] The preset proportionality coefficients refer to the weight factors artificially set to ensure that the contribution degrees of different characteristic parameters (Multi-Scale Modulation Intensity Reference (MSMI) and Micro-Saccade Non-Steady State Ratio Reference (MSNSR)) to the final calculation result conform to the actual situation when calculating the Frequency Mismatch Risk Coefficient (FMRC). These coefficients k1 and k2 are used to adjust and optimize the influence ratios of the Multi-Scale Modulation Intensity Reference (MSMI) and the Micro-Saccade Non-Steady State Ratio Reference (MSNSR) when calculating FMRC, so as to conform to the physical model of the system, experimental data or actual application requirements. For example, if MSMI has a greater contribution to FMRC in the measurement system, the value of k1 should be appropriately increased to ensure a higher influence weight of MSMI on the overall frequency matching degree. Similarly, if MSNSR has a greater influence on the mismatch risk, the value of k2 needs to be appropriately increased. Such preset proportionality coefficients can be obtained through experimental data fitting, statistical analysis or machine learning model training, and optimized and adjusted in actual applications to improve the accuracy and robustness of the measurement system.

[0104] From the Frequency Mismatch Risk Coefficient, it can be seen that the larger the Multi-Scale Modulation Intensity Reference value generated by analyzing the amplitude change of the modulation signal calculated on multiple time scales under the detection window, and the larger the Micro-Saccade Non-Steady State Ratio Reference value generated by analyzing the dynamic characteristics of the micro-saccade signal remaining stable within a fixed time under the detection window, it indicates that the Frequency Mismatch Risk Coefficient generated when identifying the matching degree between the current laser frequency and the micro-saccade inherent frequency through a pre-trained machine learning model is larger, indicating that there is a mismatch between the current laser frequency and the micro-saccade frequency. On the contrary, it indicates that there is no mismatch between the current laser frequency and the micro-saccade frequency.

[0105] Compare and analyze the Frequency Mismatch Risk Coefficient generated when identifying the matching degree between the current laser frequency and the micro-saccade inherent frequency through a pre-trained machine learning model with the preset Frequency Mismatch Risk Coefficient reference threshold to determine whether there is a mismatch between the laser frequency and the micro-saccade frequency. The determination process is as follows:

[0106] If the Frequency Mismatch Risk Coefficient is greater than the preset Frequency Mismatch Risk Coefficient reference threshold, it is determined that there is a mismatch between the current laser frequency and the micro-saccade frequency; if the Frequency Mismatch Risk Coefficient is less than or equal to the preset Frequency Mismatch Risk Coefficient reference threshold, it is determined that there is no mismatch between the current laser frequency and the micro-saccade frequency.

[0107] When the machine learning model determines that there is a mismatch between the current laser frequency and the eyeball micro-vibration frequency, through an adaptive control algorithm, based on the spectral information of the eyeball micro-vibration, and through the eye movement vibration characteristics captured in real time, the laser emission frequency is dynamically modulated. At the same time, the laser measurement position is adjusted to intelligently explore the optical reflection characteristics of different physiological regions of the eyeball, and the measurement focus is dynamically switched according to the real-time feature distribution of the eyeball movement.

[0108] By optimizing the measurement accuracy and matching degree of the laser interference measurement system in real time, it is ensured that the laser emission frequency and the inherent frequency of the eyeball micro-vibration maintain the best resonance state, thereby improving the detection quality of the micro-vibration signal. When the machine learning model determines that there is a mismatch between the current laser frequency and the eyeball micro-vibration frequency, the system needs to use an adaptive control algorithm to perform dynamic adjustment based on the real-time spectral information of the eyeball micro-vibration to ensure that the laser and the micro-vibration movement maintain the optimal interference matching relationship. The adaptive control algorithm can combine the eyeball vibration characteristics collected in real time and continuously optimize the laser emission frequency through a feedback mechanism to make it more accurately lock the natural vibration mode of the eyeball micro-vibration and avoid signal distortion or misjudgment caused by a fixed frequency.

[0109] In addition, while ensuring frequency matching, the laser measurement position is synchronously adjusted to intelligently explore the optical reflection characteristics of different physiological regions of the eyeball (such as the cornea, iris, and sclera). The optical reflectivity and interference signal characteristics of different physiological regions are different. By dynamically switching the measurement area, the area with the most stable reflection signal and the easiest micro-vibration detection can be selected to improve the reliability of data acquisition. The cornea has high transmittance and is suitable for optical interference, while the scattering characteristics of the iris and sclera may affect the measurement accuracy. Therefore, the system needs to intelligently switch the measurement points to find the optimal reflection signal area and reduce the impact of environmental interference on the measurement.

[0110] At the same time, an analysis of the real-time feature distribution based on the eyeball movement is introduced to ensure that the laser measurement focus can be dynamically adjusted following the eyeball movement. When the line-of-sight direction of an individual changes or the micro-vibration mode undergoes a short-term change, if the measurement focus remains fixed, it may lead to the loss of the interference signal or an increase in measurement error. Therefore, the system dynamically adjusts the measurement focus by real-time tracking the movement trajectory of the eyeball and combining an adaptive focusing strategy, so that the laser measurement point is always in the area with the strongest and most stable micro-vibration signal, thereby maximizing the measurement accuracy and robustness.

[0111] Through frequency adaptive control, dynamic optimization of the measurement area, and intelligent adjustment of the measurement focus, it is ensured that the laser interference measurement system can efficiently, accurately, and stably obtain the eyeball micro-vibration signal, providing more reliable data support for applications such as micro-motion analysis, lie detection, and cognitive state monitoring.

[0112] When the machine learning model determines that there is a mismatch between the current laser frequency and the eyeball micro-vibration frequency, the specific steps of dynamically modulating the laser emission frequency and adjusting the laser measurement position through the adaptive control algorithm are as follows:

[0113] When it is detected that there is a mismatch between the current laser frequency and the eyeball micro-vibration frequency, first update the laser emission frequency through the adaptive control algorithm. Let the current laser frequency be f L , and use the real-time captured eye movement vibration spectrum characteristics (such as the center of nystagmus energy or spectral centroid) and the frequency mismatch risk coefficient FMRC, combined with the preset frequency mismatch risk reference threshold to calculate the frequency modulation amount. The calculation formula is as follows:

[0114]

[0115] , where f L ′ is the adjusted laser emission frequency, f L is the current laser emission frequency, θ is the control gain parameter used to adjust the sensitivity of frequency modulation, FMRC ref is the frequency mismatch risk coefficient reference threshold, S E is the eyeball micro-vibration spectrum characteristic parameter, representing the core frequency distribution of the current micro-vibration activity;

[0116] The function of this step is to dynamically correct the laser frequency according to the real-time vibration information, making it more consistent with the natural frequency of the eyeball micro-vibration, thereby improving the accuracy and stability of the interference signal.

[0117] After the laser frequency is modulated, to optimize signal acquisition, the laser measurement area and focusing position will be intelligently adjusted according to the real-time eye movement tracking data and optical feedback information. Let the current laser measurement position be P L and the current laser focus position be F L , and at the same time introduce the regional optical reflection characteristic index to reflect the reflection characteristics of different regions such as the cornea, sclera, and iris, and jointly adjust the measurement position and focusing. The adjustment formula is as follows:

[0118]

[0119] , where P L ′ is the updated laser measurement position, and the laser measurement area is adjusted through optical feedback to adapt to the best signal reflection point. P L is the current laser measurement position, reflecting the irradiation area of the laser beam on the eyeball, which may be different physiological regions such as the cornea, iris, and sclera, is the position adjustment gain parameter, controlling the sensitivity of the measurement position adjustment and affecting the response rate of the laser beam to the measurement area, R OIt is the real-time regional optical reflection characteristic, reflecting the optical reflection ability of the current laser irradiation area (such as the cornea, iris, sclera), and is calculated based on the feedback of the photodetector. It is the preset optimal regional reflection reference value, reflecting the optical reflection level of each physiological region under ideal conditions, F L F′ is the updated laser focus position, which is the focus position after adaptive adjustment to ensure signal stability, F L It is the current laser focus position, μ is the focusing adjustment gain parameter, which controls the rate of laser focus adjustment and determines the response ability of the system to eye movement, Δ E It is the focusing error caused by eye movement, calculating the deviation of the current eye movement distribution relative to the ideal state and determining whether it is necessary to adjust the laser focus.

[0120] The function of this step is to adjust the laser irradiation position and focusing state in real time according to eye movement vibration and regional optical feedback, ensuring that the laser beam works at the optimal position and focus, thereby further improving the accuracy and robustness of interference measurement.

[0121] The present invention adopts the initial laser interference measurement and adaptive dynamic modulation strategy, which can match the frequency characteristics of eye micro-vibration in real time and accurately, and significantly improve the detection sensitivity and stability of the laser interference system to the high-frequency and micro-amplitude movement of eye micro-vibration. By introducing the quantitative analysis of key indicators after preprocessing and the intelligent matching judgment of machine learning, and combining the eye movement vibration characteristics and the optical feedback information of the eye physiological region, the present invention realizes the efficient, real-time and adaptive adjustment of the laser emission frequency and measurement focus, effectively overcoming the problem that it is difficult for the fixed-frequency laser in the prior art to capture eye micro-vibration. Therefore, the present invention can greatly improve the accuracy of lie detection and has important application value and social significance.

[0122] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0123] Only some exemplary embodiments of the present invention are described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, without departing from the spirit and scope of the present invention, the described embodiments can be modified in various different ways. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

[0124] It should be noted that in this text, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0125] It should be understood that in various embodiments of the present application, the magnitude of the sequence numbers of the above processes does not imply the order of execution, and the order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0126] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this text can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0127] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0128] The unit described as a separate component may or may not be physically separated, and the component shown as a unit may or may not be a physical unit, that is, it may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0129] In addition, in each embodiment of the present application, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0130] As described above, it is only the specific implementation manner of the present application. However, the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims described above.

[0131] Only some exemplary embodiments of the present invention have been described above by way of illustration. Without doubt, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. A psychological science analysis method based on facial and body micro-movement recognition, characterized in that: The following steps are involved: The laser interferometry system emits a light beam to the target eye area at an initially set laser emission frequency, and obtains the original reflection signal in real time through a photoelectric detector; After the original reflection signal is obtained, the original reflection signal is preprocessed, and key indicators reflecting the mismatch between the laser frequency and the eye microseismic frequency are extracted from the preprocessed data. The extracted key indicators are analyzed to quantify the mismatch between the laser emission frequency and the eye microseismic frequency; The analyzed key indicators are input into a pre-trained machine learning model, and the matching degree between the current laser frequency and the natural frequency of microseismicity is identified through the machine learning model to determine whether there is a mismatch between the laser frequency and the eye microseismic frequency; When the machine learning model determines that there is a mismatch between the current laser frequency and the eye microtremor frequency, an adaptive control algorithm is used to dynamically modulate the laser emission frequency based on the spectral information of the eye microtremor and the real-time captured eye movement vibration characteristics. At the same time, the laser measurement position is adjusted to intelligently explore the optical reflection characteristics of different physiological areas of the eye and dynamically switch the measurement focus according to the real-time characteristic distribution of eye movement.

2. The psychological science analysis method based on facial and body micro-movement recognition according to claim 1, characterized in that: The laser interferometry system emits a beam to the target eye area at the initially set laser emission frequency and obtains the original reflection signal in real time through the photoelectric detector. The specific steps are as follows: First, the laser interferometry system is initialized and the initial laser emission frequency is set; Secondly, the laser emits a stable and highly coherent beam at this set frequency, precisely focused on the target eye area; Next, the photodetector captures the reflected signal from the target surface in real time; the reflected signal is then transmitted to the data processing module to provide raw data for subsequent filtering, denoising and time-frequency analysis; Finally, these data are stored and organized to form a multi-dimensional feature set, laying the foundation for subsequent feature extraction and dynamic adjustment.

3. The psychological science analysis method based on facial and body micro-movement recognition according to claim 1, characterized in that: Key indicators reflecting the mismatch between laser frequency and eye microseismic frequency are extracted from the preprocessed data. The extracted key indicators include the amplitude change of the modulation signal calculated on multiple time scales and the dynamic characteristics of the microseismic signal remaining stable within a fixed time. The amplitude change of the modulation signal calculated on multiple time scales and the dynamic characteristics of the microseismic signal remaining stable within a fixed time are analyzed under the detection window to generate multi-scale modulation intensity reference values ​​and eye microseismic non-steady-state ratio reference values, and the mismatch between the laser emission frequency and the eye microseismic frequency is quantified by the multi-scale modulation intensity reference values ​​and the eye microseismic non-steady-state ratio reference values.

4. The psychological science analysis method based on facial and body micro-movement recognition according to claim 3 is characterized in that: Firstly, the obtained interference reflection signal is decomposed into multiple scales to analyze the amplitude variation of the modulation signal on different time scales, and the nonlinear energy distribution factor and signal envelope complexity ratio are introduced as characteristic indicators. The nonlinear energy distribution factor is used to measure the distribution of interference signal energy at different scales. The calculation expression is as follows: ， In the formula, is the modulated signal strength at the i-th moment under scale λ, is the modulation signal strength at the i-1th moment under scale λ, that is, the modulation signal strength at the previous moment, β is the adaptive nonlinear weight factor, and Eλ is the nonlinear energy distribution factor; The signal envelope complexity ratio is used to characterize the nonlinear complexity of the signal envelope at different scales. The calculation expression is as follows: ， In the formula, is the modulation signal strength at the jth moment under scale λ, β is the dynamically adjusted envelope characteristic index, C μ is the signal envelope complexity ratio; In order to obtain the nonlinear energy distribution factor E λ Sum signal envelope complexity ratio C μ Finally, a multi-scale modulation intensity reference value is generated as the final matching evaluation index. The generation formula is as follows: ， Where MSMI is the multi-scale modulation intensity reference value, Γ λ is the scale sensitivity factor, γ is the modulation index control parameter, Λ λ is the nonlinear modulation weight, and δ is the envelope complexity control parameter.

5. The psychological science analysis method based on facial and body micro-movement recognition according to claim 3 is characterized in that: The specific steps of analyzing the dynamic characteristics of the microseismic signal that remains stable within a fixed time under the detection window to generate the reference value of the non-steady-state ratio of the eye microseismic signal are as follows: Firstly, the non-steady-state characteristics of microseismic signals are analyzed to evaluate their dynamic change trend over time, and the discrete spectrum deviation and instantaneous modulation expansion factor are introduced to measure the non-steady-state degree of microseismic signals. Discrete spectrum deviation is used to measure the deviation of the spectrum center of the microseismic signal from its previous state at a fixed time. The calculation expression is as follows: ， Where DSS is the discrete spectrum deviation, f q is the frequency value of the qth frequency component, F ref is the center frequency reference value of the previous detection window, P q is the power spectral density of the qth frequency component; The instantaneous modulation expansion factor is used to measure the degree of change of the modulation mode of the microseismic signal in a short period of time and reflects the expansion trend of the non-steady-state signal. The calculation expression is as follows: ， Where IMEF is the instantaneous modulation expansion factor, M k is the instantaneous modulation amplitude at the kth moment, M k-1 is the instantaneous modulation amplitude at the previous moment, ∈ is a small constant to prevent the denominator from being zero; After constructing the discrete spectrum deviation DSS and the instantaneous modulation expansion factor IMEF, the reference value of the non-steady-state ratio of eye microseismicity is generated to comprehensively evaluate the matching degree between the laser frequency and the microseismic signal frequency. The generation formula is as follows: ， Where, MSNSR is the reference value of the non-steady-state ratio of microstremors, ω is the weight coefficient of the discrete spectrum deviation DSS, is the weight coefficient of the instantaneous modulation expansion factor IMEF, τ is the interaction weight between the discrete spectrum shift DSS and the instantaneous modulation expansion factor IMEF, and σ is a small constant to prevent the denominator from being zero.

6. The psychological science analysis method based on facial and body micro-movement recognition according to claim 3, characterized in that: The analyzed multi-scale modulation intensity reference value and the eye microseismic non-steady-state ratio reference value are input into the pre-trained machine learning model, and the frequency mismatch risk coefficient is generated by the machine learning model. The frequency mismatch risk coefficient is used to identify the degree of matching between the current laser frequency and the microseismic inherent frequency, and to determine whether there is a mismatch between the laser frequency and the eye microseismic frequency.

7. The psychological science analysis method based on facial and body micro-movement recognition according to claim 6, characterized in that: The frequency mismatch risk coefficient generated when the pre-trained machine learning model identifies the matching degree between the current laser frequency and the microseismic natural frequency is compared with the pre-set frequency mismatch risk coefficient reference threshold to determine whether there is a mismatch between the laser frequency and the eye microseismic frequency. The judgment process is as follows: If the frequency mismatch risk coefficient is greater than the pre-set frequency mismatch risk coefficient reference threshold, it is judged that there is a mismatch between the current laser frequency and the eye microseismic frequency; if the frequency mismatch risk coefficient is less than or equal to the pre-set frequency mismatch risk coefficient reference threshold, it is judged that there is no mismatch between the current laser frequency and the eye microseismic frequency.

8. The psychological science analysis method based on facial and body micro-movement recognition according to claim 1, characterized in that: When the machine learning model determines that there is a mismatch between the current laser frequency and the eye microtremor frequency, the specific steps of dynamically modulating the laser emission frequency and adjusting the laser measurement position through the adaptive control algorithm are as follows: When a mismatch between the current laser frequency and the eye microtremor frequency is detected, the laser emission frequency is first updated through an adaptive control algorithm. Suppose the current laser frequency is f L , using the real-time captured eye movement vibration spectrum characteristics and the frequency mismatch risk coefficient FMRC, combined with the preset frequency mismatch risk reference threshold to calculate the frequency modulation amount, the calculation expression is as follows: ， In the formula, f L ′ is the adjusted laser emission frequency, f L is the current laser emission frequency, θ is the control gain parameter, FMRC ref is the reference threshold of the frequency mismatch risk factor, S E It is the characteristic parameter of eye microtremor spectrum; After laser frequency modulation, in order to optimize signal acquisition, the laser measurement area and focus position will be intelligently adjusted according to real-time eye tracking data and optical feedback information. The current laser measurement position is set as P L and the current laser focus position is F L , and introduce the regional optical reflection characteristic index to jointly adjust the measurement position and focus. The adjustment formula is as follows: ， Where P L ′ is the updated laser measurement position, P L is the current laser measurement position, is the position adjustment gain parameter, R O is the real-time regional optical reflection characteristic, is the preset optimal area reflection reference value, F L ′ is the updated laser focus position, F L is the current laser focus position, μ is the focus adjustment gain parameter, Δ E It is the focusing error caused by eye movement.

Images