Intelligent ear vagus nerve regulation glasses and control method, control device and regulation system thereof

Abstract

This invention relates to the field of smart wearable devices, and particularly to a smart ear vagus nerve modulation glasses and its control method, control device, and modulation system. The method includes: acquiring dynamic pupil feature parameters monitored by the smart ear vagus nerve modulation glasses; calling a trained machine learning model to analyze real-time physiological state information based on the pupil dynamic feature parameters; generating electrical stimulation control commands corresponding to the analyzed real-time physiological state information; and controlling the ear vagus electrodes in the smart ear vagus nerve modulation glasses to perform neural modulation based on the electrical stimulation control commands. This method can overcome the limitations of fixed parameter output in traditional ear vagus nerve stimulation devices, improve the individualization and accuracy of ear vagus nerve stimulation programs, and avoid problems such as insufficient treatment effect or discomfort.

Description

Technical Field

[0001] This invention relates to the field of smart wearable devices, and in particular to a smart ear vagus nerve modulation glasses and its control method, control device, and modulation system. Background Technology

[0002] Auricular vagus nerve stimulation, as a non-invasive neuromodulation method, has been clinically proven to have significant therapeutic value for various diseases such as depression, epilepsy, and autonomic nervous system disorders, providing an important technical direction for the non-invasive treatment of related diseases.

[0003] Traditional vagus nerve stimulation devices generally rely on manual setting of fixed parameters. That is, key parameters such as stimulation intensity, frequency, and pulse width are preset at the factory. Fixed parameter output modes cannot adapt to individual differences and real-time physiological changes of users, making it difficult for stimulation programs to achieve individualized optimal treatment effects. This may result in insufficient treatment effects or discomfort, affecting the efficiency of clinical treatment.

[0004] Therefore, how to overcome the limitations of fixed parameter output in traditional vagus nerve stimulation devices, improve the individualization and accuracy of vagus nerve stimulation programs, and avoid problems such as insufficient treatment effect or discomfort are urgent issues that need to be addressed by those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide at least one intelligent auricular vagus nerve modulation glasses and its control method, control device, and modulation system, which can overcome the limitations of fixed parameter output of traditional auricular vagus nerve stimulation devices, intelligently optimize stimulation parameters, maximize and personalize the therapeutic effect, and provide a more efficient and precise solution for clinical applications in the field of neuromodulation.

[0006] To address the aforementioned technical problems, at least one embodiment of this application provides a control method for intelligent vagus nerve modulation glasses, comprising: The pupil dynamic feature parameters are obtained by monitoring the intelligent ear vagus nerve modulation glasses; wherein, the pupil dynamic feature parameters are obtained by using a miniature camera installed on the lens of the intelligent ear vagus nerve modulation glasses to call real-time eye tracking technology combined with edge detection algorithm to dynamically monitor the pupil diameter at the sub-millimeter level; The trained machine learning model is invoked to analyze real-time physiological state information based on the pupil dynamic feature parameters; Based on the real-time physiological state information obtained from the analysis, an electrical stimulation control command corresponding to the real-time physiological state information is generated, and the vagus electrodes in the smart ear vagus nerve modulation glasses are controlled to perform neural modulation based on the electrical stimulation control command.

[0007] In one embodiment, the step of calling a trained machine learning model to analyze real-time physiological state information based on the pupil dynamic feature parameters includes: Effective features of pupil state changes are extracted from the pupil dynamic feature parameters; wherein, the effective features include at least one of the following: diameter fluctuation trend and light reflection latency change pattern; A lightweight LSTM machine learning model is invoked to predict physiological state information based on the mapping relationship between pupil dynamic features and physiological state information, generating real-time physiological state assessment results; wherein, the physiological state assessment results include physiological state type and state severity.

[0008] In one embodiment, the control method of the intelligent vagus nerve modulation glasses further includes: The system receives updated pupil dynamic feature parameters in real time and generates updated physiological state assessment results based on the updated pupil dynamic feature parameters. By comparing the physiological state assessment results before and after the update, a state comparison result is generated; Based on the state comparison results, an electrical stimulation control command corresponding to the updated pupil dynamic characteristic parameters is generated.

[0009] In one embodiment, the control method for the intelligent vagus nerve modulation glasses further includes: Read the built-in physiological safety threshold parameters; the physiological safety threshold parameters include at least one of the following: upper limit of stimulation intensity and upper limit of single continuous stimulation duration; Determine whether the control command exceeds the physiological safety threshold parameter; If the physiological safety threshold parameter is exceeded, the control command that exceeds the physiological safety threshold parameter will be adjusted to the threshold range before being output.

[0010] In one embodiment, the control method for the intelligent vagus nerve modulation glasses, wherein acquiring the pupil dynamic feature parameters monitored by the intelligent vagus nerve modulation glasses includes: Acquire dynamic pupil images monitored by the intelligent ear vagus nerve modulation glasses; Acquire pupil response data and eye movement data obtained by the intelligent ear vagus nerve modulation glasses through infrared sensing; Based on the pupil dynamic image, the pupil response data, and the eye movement data, the pupil dynamic characteristic parameters of the wearer's pupil are calculated.

[0011] In one embodiment, calculating the wearer's pupil characteristic parameters based on the pupil dynamic image, the pupil response data, and the eye movement data includes: Extract the region of interest (ROI) of the pupil from the dynamic pupil image; Based on the pupil response data, the brightness characteristics of the pixels within the region of interest of the pupil are calculated; the brightness characteristics include the average brightness and the standard deviation of brightness. An adaptive threshold adjustment strategy is used to calculate the high and low thresholds required for edge detection. In this strategy, the high threshold is set by combining the pupil diameter compensation coefficient with the sum of the average brightness and the brightness standard deviation at a first multiple. The low threshold is set by combining the texture complexity compensation coefficient with the sum of the average brightness and the brightness standard deviation at a second multiple. The first multiple is higher than the second multiple. Based on the high threshold and the low threshold, edge detection of the pupil is performed on the region of interest of the pupil, and the edge detection results are corrected by combining the eye movement data; The feature parameters of the pupil are extracted based on the corrected edge detection results.

[0012] In one embodiment, before performing pupil edge detection on the region of interest based on the high threshold and the low threshold, and correcting the edge detection result by combining the eye movement data, the method further includes: Ambient lighting conditions are determined based on the pupil response data and the average brightness value. If the ambient lighting conditions are strong light environments, adjust the high threshold and the low threshold upwards; If the ambient lighting conditions are low light conditions, adjust the high threshold and the low threshold downwards.

[0013] At least one embodiment of this application provides a control device for smart ear vagus nerve modulation glasses, comprising: The data acquisition module is used to acquire the dynamic pupil feature parameters monitored by the smart ear vagus nerve modulation glasses; wherein, the dynamic pupil feature parameters are obtained by using a miniature camera installed on the lens of the smart ear vagus nerve modulation glasses to call real-time eye-tracking technology combined with edge detection algorithm to dynamically monitor the pupil diameter at the sub-millimeter level; The state analysis module is used to call a trained machine learning model to analyze real-time physiological state information based on the pupil dynamic feature parameters; The instruction generation module is used to generate electrical stimulation control instructions corresponding to the real-time physiological state information based on the analyzed real-time physiological state information. The control and adjustment module is used to control the vagus electrodes in the smart ear vagus nerve modulation glasses to perform neural modulation based on the electrical stimulation control command.

[0014] At least one embodiment of this application provides a smart ear vagus nerve modulation glasses, including a frame and a control device for the smart ear vagus nerve modulation glasses as described above. The frame includes: a frame, lenses, and temples, and the control device of the smart ear vagus nerve modulation glasses is installed in the frame; Each of the lenses is equipped with a miniature camera, which is connected to the control device of the smart ear vagus nerve modulation glasses.

[0015] At least one embodiment of this application provides an intelligent ear vagus nerve modulation system, including: intelligent ear vagus nerve modulation glasses as described above, and an electrical stimulation system connected to a control device of the intelligent ear vagus nerve modulation glasses; The electrical stimulation system is worn on the wearer's ear and is used to deliver electrical stimulation to the wearer's vagus nerve.

[0016] The intelligent auricular vagus nerve modulation glasses and their control method, control device, and modulation system provided in the embodiments of this application achieve sub-millimeter-level dynamic monitoring of pupil diameter by integrating a miniature camera into the glasses lens and calling real-time eye-tracking technology and edge detection algorithms. This effectively overcomes the technical bottlenecks of traditional pupil measurement devices, which are bulky and have low sampling rates. It can accurately capture subtle dynamic changes in the pupil, providing high-quality raw data for subsequent feature parameter calculations. At the same time, the method constructs a complete logical chain of pupil dynamic image acquisition, feature parameter calculation, real-time physiological state analysis, and electrical stimulation control command generation. This solves the problems of traditional auricular vagus nerve stimulators relying on manual parameter settings and being disconnected from physiological signal monitoring, and avoids the limitations of existing neuromodulation devices with fixed parameter outputs that cannot adapt to individual real-time physiological differences. It can dynamically generate appropriate electrical stimulation commands based on the wearer's real-time physiological state, forming a closed-loop system of monitoring-analysis-modulation. This makes electrical stimulation modulation more in line with individual real-time physiological needs, helping to improve the accuracy and individualization of adjuvant treatment for diseases such as depression and epilepsy, and providing a more efficient and reliable technical solution for clinical applications in the field of neuromodulation. Attached Figure Description

[0017] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.

[0018] Figure 1 This is a flowchart illustrating a control method for intelligent vagus nerve modulation glasses provided in one embodiment of this application; Figure 2 This is a schematic diagram of a control device for an intelligent vagus nerve modulation glasses according to an embodiment of this application; Figure 3This is a three-dimensional structural schematic diagram of an intelligent ear vagus nerve modulation glasses provided in one embodiment of this application; Figure 4 This is a schematic diagram of the inner structure of a smart ear vagus nerve modulation glasses provided in one embodiment of this application; Figure 5 This is a schematic diagram of data transmission of a communication module provided in one embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the various embodiments of this application will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the various embodiments of this application to help readers better understand this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various changes and modifications based on the following embodiments. The division of the various embodiments below is for the convenience of description and should not constitute any limitation on the specific implementation of this application. The various embodiments can be combined with and referenced by each other without contradiction.

[0020] Dynamic changes in the pupil (such as pupillary light reflex, diameter fluctuations, and constriction / dilation rates) are sensitive and intuitive indicators of autonomic nervous system function, directly reflecting the balance between the sympathetic and parasympathetic nervous systems. However, traditional pupillary measurement devices are mostly used to assess autonomic nervous system function (such as pupillary light reflex), and suffer from drawbacks such as large size and low sampling rate, making it difficult to achieve accurate and real-time monitoring of dynamic pupillary changes. This invention proposes a control method for intelligent auricular vagus nerve modulation glasses. This method is based on real-time capture of the patient's pupillary changes through dynamic pupillary monitoring, intelligently optimizing stimulation parameters to maximize and personalize the treatment effect, providing a more efficient and accurate solution for clinical applications in the field of neuromodulation.

[0021] The following is a detailed description of the implementation details of the control method for the intelligent ear vagus nerve modulation glasses in this embodiment. The following content is only for the convenience of understanding and is not necessary for implementing this solution.

[0022] Example 1: The flowchart illustrating the control method of the intelligent vagus nerve modulation glasses provided in this embodiment can be seen as follows: Figure 1 As shown, its implementation steps include: Step 101: Obtain the dynamic pupil feature parameters obtained from the monitoring of the intelligent ear vagus nerve modulation glasses.

[0023] The pupil dynamic feature parameters are obtained by using a miniature camera mounted on the lens of the smart ear vagus nerve-controlled glasses to dynamically monitor the pupil diameter at the sub-millimeter level using real-time eye-tracking technology combined with edge detection algorithms. Specifically, dynamic pupil images are acquired by a miniature camera mounted on the lens of the smart ear vagus nerve-controlled glasses, and then the dynamic pupil feature parameters are obtained through image analysis. The image acquisition relies on the miniature camera mounted on the lens, which is located on the inner side of the lens close to the eye, directly facing the wearer's pupil, allowing for optimal capture of pupil dynamics.

[0024] The miniature camera utilizes real-time eye-tracking technology combined with edge detection algorithms to achieve sub-millimeter-level dynamic monitoring of pupil diameter, accurately capturing subtle changes in the wearer's pupils. This dynamic pupil diameter monitoring technology is real-time, continuously and dynamically acquiring pupil diameter information. Edge detection algorithms (such as the Canny algorithm) effectively filter environmental noise interference and highlight pupil edge features through gradient calculation, non-maximum suppression, and dual-threshold connection. It should be noted that the auricular vagus nerve modulation glasses used in this control method are primarily for patients with neuromodulation needs. These patients have unrestricted activity levels and can switch between various scenarios, including indoors and outdoors. Light intensity can range from hundreds to tens of thousands of lux, changing rapidly and unpredictably. Compared to AR and VR glasses, where the usage scenarios are relatively controllable, the dynamic threshold of the auricular vagus nerve modulation glasses must have rapid and precise adjustment capabilities to adapt to such drastic light intensity changes in a very short time, accurately capturing pupil edge information to ensure the accuracy of pupil diameter monitoring.

[0025] Furthermore, this step is not a single image capture, but rather a dynamic monitoring process achieved through the synergy of real-time eye-tracking technology and edge detection algorithms. On one hand, real-time eye-tracking technology can follow the wearer's eye movements in real time, ensuring that the miniature camera is always focused on the pupil area, avoiding pupil misses caused by head movements or eye rotations, and ensuring the continuity of image acquisition. On the other hand, edge detection algorithms (such as the Canny algorithm) can accurately identify the boundaries between the pupil and the iris and cornea in the image through gradient calculation, non-maximum suppression, and dual threshold connection, providing an algorithmic basis for subsequent extraction of pupil dynamic information. The combination of these two technologies upgrades image acquisition from static photography to precise monitoring through dynamic tracking and boundary recognition, solving the problems of low sampling rate and susceptibility to motion interference in traditional devices.

[0026] A miniature camera uses real-time eye-tracking technology to locate the pupil edge and calculate its diameter. One calculation process is as follows: 1. Ellipse Fitting: The pupil edge is elliptical. Least squares ellipse fitting (such as fitEllipse in OpenCV) is used to avoid edge breakage caused by eyelashes. 2. Subpixel Accuracy Optimization: By interpolation or gradient subpixel positioning (such as the Lucas-Kanade method), the edge positioning accuracy is improved to the 0.1 pixel level (corresponding to sub-millimeter physical accuracy). 3. Diameter Output: The major axis of the ellipse is taken as the pupil diameter, and dynamic changes are analyzed in conjunction with time series data. This embodiment only uses the above calculation process as an example. The implementation process of other algorithms can be referred to in this embodiment, but it is not limited in this embodiment. Furthermore, this embodiment does not limit the specific edge detection algorithm used in the miniature camera. For better understanding, the Canny algorithm is used as an example, with one calculation step as follows: 1. Gradient calculation: Use a 5×5 or 3×3 Sobel operator (with moderate computational cost) to obtain the horizontal and vertical gradients (Gx, Gy); 2. Non-maximum suppression: Only retain edge points with local maxima in the gradient direction to ensure that the edges are refined to single pixels; 3. Dual threshold connection: High threshold for strong edges (determining the pupil boundary); low threshold for weak edges (the part connected to the strong edge is retained). Of course, other algorithms can also be used, and all can refer to the description in this embodiment, which will not be repeated here.

[0027] Furthermore, after acquiring the dynamic image of the pupil, the dynamic characteristic parameters of the wearer's pupil are calculated. The specific characteristic parameters may include, but are not limited to: pupil diameter, pupil diameter change rate, reflectivity parameters (such as the latency of light reflection), fluctuation pattern parameters (such as the dominant frequency of pupil oscillation), and stability parameters (such as the variance of pupil diameter change).

[0028] Step 102: Call the trained machine learning model to analyze real-time physiological state information based on pupil dynamic feature parameters.

[0029] Dynamic changes in the pupil (such as diameter fluctuations and light reflex speed) directly reflect the state of autonomic nervous system function and are related to the pathophysiological characteristics of diseases such as depression, epilepsy, and anxiety. Based on this, this step uses calculated pupil characteristic parameters (such as pupil diameter, diameter change rate, light reflex latency, and pupil oscillation frequency) as the core basis, and uses a trained machine learning model to associate physiological patterns related to autonomic nervous system function and disease state, thereby determining the wearer's real-time physiological state.

[0030] Step 103: Based on the real-time physiological state information obtained from the analysis, generate electrical stimulation control commands corresponding to the real-time physiological state information.

[0031] Optionally, based on the physiological state assessment results, a preset state-stimulation parameter mapping rule is matched to generate targeted electrical stimulation control instructions, which include key parameters such as stimulation frequency, pulse width, waveform, and intensity.

[0032] For example, if the assessment result is a depressive state (probability 72%) + autonomic nervous system hyperexcitability (8.2 / 10), the chip will match the regulatory target of increasing parasympathetic nerve activity and generate a stimulation frequency of 5Hz, a triangular waveform, and an intensity optimized based on historical treatment responses. If the assessment result is a high risk of epileptic seizures (probability 85%), a 25Hz square wave and a high-priority emergency inhibition instruction will be generated to ensure that the stimulation parameters are fully adapted to the current physiological state. This avoids both insufficient stimulation leading to ineffective treatment and excessive stimulation causing discomfort, ultimately achieving individualized and precise vagus nerve modulation.

[0033] Step 104: Based on the electrical stimulation control command, control the auricular vagus electrodes in the smart auricular vagus nerve modulation glasses to perform neural modulation.

[0034] Based on the above introduction, the control method of the intelligent ear vagus nerve modulation glasses provided in this embodiment achieves sub-millimeter-level dynamic monitoring of pupil diameter by integrating a miniature camera into the glasses lens and calling real-time eye-tracking technology and edge detection algorithm. This effectively overcomes the technical bottlenecks of traditional pupil measurement devices, which are bulky and have low sampling rates. It can accurately capture subtle dynamic changes in the pupil, providing high-quality raw data for subsequent feature parameter calculation. At the same time, this method constructs a complete logical chain of pupil dynamic image acquisition, feature parameter calculation, real-time physiological state analysis, and electrical stimulation control command generation. This not only solves the problem of traditional ear vagus nerve stimulators relying on manual parameter settings and being disconnected from physiological signal monitoring, but also avoids the limitations of existing neuromodulation devices with fixed parameter outputs that cannot adapt to individual real-time physiological differences. It can dynamically generate appropriate electrical stimulation commands based on the wearer's real-time physiological state, forming a closed-loop system of monitoring-analysis-modulation. This makes electrical stimulation modulation more in line with individual real-time physiological needs, helping to improve the accuracy and individualization of adjuvant treatment for diseases such as depression and epilepsy, and providing a more efficient and reliable technical solution for clinical applications in the field of neuromodulation.

[0035] Example 2: In the above embodiments, the specific type of machine learning model used in step 102 is not limited, and can be set according to actual usage requirements. In one embodiment, a lightweight LSTM machine learning model can be selected. Step 102 may specifically include the following sub-steps: Step 31: Extract effective features of pupil state changes from pupil dynamic feature parameters.

[0036] The feature parameters are filtered in a time series manner to remove abnormal data and then the effective features of pupil state changes are extracted. The effective features include at least one of the following: diameter fluctuation trend and light reflection latency change pattern.

[0037] Environmental interference during wear (such as overexposure of pupil images due to strong light, or monitoring deviations caused by head movement) or temporary equipment errors (such as momentary lag in the miniature camera) may generate abnormal data (such as sudden jumps in pupil diameter exceeding the physiological range, or abnormally high values ​​for pupillary light reflex latency). Temporal filtering (such as combining the changing patterns of adjacent frames to remove isolated data points deviating from the normal fluctuation range) ensures the reliability of the data input to the model. The final extracted diameter fluctuation trend (such as whether the pupil diameter is continuously expanding, contracting, or fluctuating periodically) and pupillary light reflex latency change pattern (such as whether the latency is gradually lengthening, shortening, or remaining stable) are key features directly related to the state of autonomic nervous system function. For example, in a depressive state, there is often continuous pupil diameter expansion with a smooth fluctuation trend; in an anxious state, there is often high-frequency fluctuation in pupil diameter with a shortened pupillary light reflex latency. These features provide accurate analytical basis for subsequent physiological state prediction.

[0038] Step 32: Call the lightweight LSTM machine learning model to predict the physiological state information of the effective features based on the mapping relationship between pupil dynamic features and physiological state information, and generate real-time physiological state assessment results.

[0039] A lightweight LSTM machine learning model is invoked to predict the physiological state based on the mapping relationship between pupil features and physiological state, generating real-time physiological state assessment results. The physiological state assessment results include the physiological state type and the severity of the state.

[0040] A lightweight LSTM machine learning model, based on a pre-defined mapping relationship between pupil features and physiological states (i.e., the model has been trained on a large amount of clinical data to establish a logical association between specific pupil features and specific physiological states), analyzes the aforementioned effective features and ultimately outputs physiological state assessment results that can be directly used for regulation. The results include the physiological state type (clearly identifying the wearer's current pathological / physiological state, such as depressive tendencies, anxiety, autonomic nervous system imbalance, increased risk of epileptic seizures, etc.) and the severity of the state (quantifying the degree of severity of the state, such as a 72% probability of depression, an autonomic nervous system excitability of 8.2 / 10, etc.). Combining these two aspects avoids the problem of determining the regulation intensity by only judging the state type, providing a basis for generating precise electrical stimulation commands.

[0041] To enhance understanding, this embodiment introduces a processing mechanism for a lightweight LSTM machine learning model based on input features. One input feature is as follows: features=[pupil_diameter, #current pupil diameter (mm)] diameter_velocity, # Rate of change of diameter (mm / s) light_reflex_lag, # Latency of light reflection (ms) oscillation_freq, # Pupil oscillation frequency (Hz) The output decision logic of a lightweight LSTM machine learning model: ; The dynamic mapping rule for a lightweight LSTM machine learning model is as follows: Anxiety level rises → Select 1-10Hz low-frequency triangular wave (parasympathetic nerve activation). Epilepsy seizure probability >80% → Emergency suppression with 25Hz square wave; Autonomic nervous system imbalance → random noise wave stimulation (avoid neural adaptation); The above-mentioned image processing algorithms and machine learning models work together through a collaborative mechanism; For the context of depression treatment, the following is an example of closed-loop regulation using a lightweight LSTM machine learning model: Image layer: Detected continuous pupil dilation (diameter increased by 0.4 mm, lasting for 5 seconds); Feature layer: Extract the features of "low oscillation frequency + high expansion speed" → depressive state; Model layer: LSTM output depression probability: 72%, autonomic nervous system excitability score: 8.2 / 10; Decision-making level: if depression_prob > 0.7 and arousal_score > 8 stim_freq=5Hz # Increases parasympathetic nerve activity wave_type="triangular wave" # Gentle stimulation to avoid discomfort intensity=auto_adjust_based(history) # Optimize based on historical response curves.

[0042] The decision-making team uses conditional judgment (probability of depression > 0.7 and excitement level > 8) to match the stimulation parameters of a 5Hz triangular wave (5Hz is in the low frequency range of 1-10Hz, which can activate the parasympathetic nervous system to relieve excitement, and the triangular wave stimulation is gentler and avoids discomfort). At the same time, the stimulation intensity is optimized by combining historical treatment response curves, and finally an electrical stimulation command is generated and transmitted to the electrical stimulation system to complete the closed-loop treatment of monitoring-analysis-regulation.

[0043] Example 3: To improve the dynamic adaptation performance of the electrical stimulation protocol, the following steps can be further performed based on the above embodiments: Step 105: Receive the updated pupil dynamic feature parameters in real time, and generate an updated physiological state assessment result based on the updated pupil dynamic feature parameters.

[0044] Step 106: Compare the physiological state assessment results before and after the update to generate state comparison results.

[0045] The physiological state assessment results before and after the update are quantitatively compared to generate state comparison results, which include both whether the state type has changed (such as from anxiety tendency to autonomic nervous balance) and the fluctuation range of state severity (such as the probability of epileptic seizures increasing from 60% to 75% and the dominant frequency of pupillary oscillations changing from 0.5Hz to 0.8Hz), accurately capturing the dynamic trend of physiological state.

[0046] Step 107: Generate an electrical stimulation control command corresponding to the updated pupil dynamic characteristic parameters based on the state comparison results.

[0047] Based on the state comparison results, the chip determines whether the current electrical stimulation program is suitable for the latest state. For example, if the state improves (such as a decrease in the probability of depression), it generates control instructions to reduce the stimulation intensity and adjust the frequency (such as changing from a 5Hz triangular wave to a 3Hz triangular wave); if the state deteriorates (such as an increase in the risk of epilepsy), it generates instructions to strengthen the stimulation parameters (such as changing from a 10Hz square wave to a 25Hz square wave); if the state is stable, it maintains or fine-tunes the parameters to ensure that the stimulation program always matches the physiological state.

[0048] Based on the above description, this embodiment receives updated feature parameters in real time and generates the latest physiological state assessment results based on the built-in image processing algorithm and machine learning model. Then, by comparing the physiological state assessment results before and after the update, the changing trends and differences in the wearer's physiological state are clarified, forming an intuitive state comparison result. Finally, based on this state comparison result, control commands for adjusting electrical stimulation are precisely generated. This method can dynamically track real-time changes in the wearer's physiological state, avoiding the problem of insufficient adaptability of electrical stimulation protocols caused by relying on fixed parameters or delayed assessments. It ensures that electrical stimulation modulation is always highly matched to the wearer's current physiological state, further improving the accuracy and individualization of neuromodulation therapy. Furthermore, through dynamic adjustments based on state changes, it ensures that the treatment plan can be optimized in a timely manner according to the progression of the disease or physiological fluctuations, providing more adaptive technical support for the efficient treatment of patients undergoing neuromodulation.

[0049] Example 4: To prevent electrical stimulation from exceeding physiological safety limits and affecting user experience, a safety protection mechanism can be implemented based on the above embodiments. Specifically, the following steps can be further performed: Step 108: Read the built-in physiological safety threshold parameters.

[0050] Pre-setting physiological safety threshold parameters includes upper limits for stimulation intensity (such as the maximum safe value of electrical stimulation current / voltage to avoid discomfort or nerve damage to ear tissues caused by excessively high-intensity stimulation) and upper limits for the duration of a single continuous stimulation (such as the longest duration of a single stimulation to prevent nerve over-excitation or fatigue caused by prolonged stimulation). These threshold parameters are set based on clinical physiological safety research data and are adapted to the human body's tolerance range to vagus nerve stimulation.

[0051] Step 109: Determine whether the control command exceeds the physiological safety threshold parameter.

[0052] After generating electrical stimulation control instructions (including parameters such as stimulation intensity and duration), the parameters in the instructions are compared with the built-in physiological safety thresholds in real time to determine whether the current instructions exceed the upper limit for stimulation intensity or the upper limit for single continuous stimulation. For example, if the stimulation intensity in the instructions is 15mA (while the upper limit is 12mA) and the single continuous stimulation duration is 30 minutes (while the upper limit is 20 minutes), it is determined that the safety threshold is exceeded.

[0053] Step 110: If the physiological safety threshold parameter is exceeded, adjust the control command that exceeds the physiological safety threshold parameter to the threshold range.

[0054] If the electrical stimulation control command is detected to exceed the physiological safety threshold, the parameters exceeding the limit are adjusted and corrected to the threshold range before being output to the electrical stimulation system. For example, the stimulation intensity of 15mA is reduced to 12mA (the upper limit), and the duration of a single continuous stimulation of 30 minutes is shortened to 20 minutes (the upper limit), ensuring that the final electrical stimulation operation always meets the physiological safety standards and avoids safety risks.

[0055] Based on the above introduction, this embodiment constructs a key safety control mechanism by pre-setting and reading physiological safety threshold parameters, including the upper limit of stimulation intensity and the upper limit of single continuous stimulation duration. After the control device of the intelligent auricular vagus nerve modulation glasses generates an electrical stimulation control command, this mechanism first judges whether the command exceeds the preset physiological safety threshold parameters. If the command is detected to exceed the threshold range, the excess parameters are automatically adjusted to within the safety threshold before being output to the electrical stimulation system. On the one hand, by clearly setting physiological safety threshold parameters, a strict safety boundary is defined for electrical stimulation therapy, avoiding physiological risks such as nerve damage to patients caused by excessive stimulation intensity or excessively long single continuous stimulation time from the source. On the other hand, with the help of automatic judgment and parameter adjustment functions, it can ensure that every electrical stimulation command meets safety standards without manual intervention, thus ensuring the treatment safety of patients undergoing long-term, dynamic use of the device and further improving its reliability and applicability in clinical applications.

[0056] Example 5: To improve the accuracy and stability of the collected data, in one embodiment, the pupil dynamic feature parameters monitored by the intelligent ear vagus nerve modulation glasses are obtained, including: acquiring the pupil dynamic image monitored by the intelligent ear vagus nerve modulation glasses; acquiring pupil reaction data (pupil light reflection amplitude, pupil light reflection latency, etc.) and eye movement data (including eyeball rotation angle and relative eye displacement) collected by the intelligent ear vagus nerve modulation glasses through infrared sensing, so as to compensate for the monitoring defects of a single miniature camera in strong light (corneal overexposure) and weak light (blurred pupil edge) environments, and help improve the pupil monitoring accuracy. Based on the pupil dynamic image, pupil reaction data and eye movement data, the characteristic parameters of the wearer's pupil are calculated.

[0057] In this embodiment, the specific algorithm used to calculate the wearer's pupil feature parameters based on pupil dynamic images, pupil response data, and eye movement data is not limited. To enhance scene adaptability and meet complex usage requirements, it can be implemented according to the following steps: Step 21: Extract the region of interest (ROI) of the pupil from the dynamic pupil image.

[0058] Step 22: Combine the pupil response data to calculate the brightness characteristics of the pixels within the region of interest of the pupil; the brightness characteristics include the average brightness and the standard deviation of brightness.

[0059] Step 23: Calculate the high and low thresholds required for edge detection using an adaptive threshold adjustment strategy.

[0060] In the adaptive threshold adjustment strategy, the high threshold is set by combining the pupil diameter compensation coefficient with the superposition result of the average brightness and the brightness standard deviation of the first multiple; the low threshold is set by combining the texture complexity compensation coefficient with the superposition result of the average brightness and the brightness standard deviation of the second multiple; the first multiple is higher than the second multiple.

[0061] Step 24: Perform edge detection of the pupil based on the high and low thresholds for the region of interest of the pupil, and correct the edge detection results by combining eye motion data.

[0062] Step 25: Extract the feature parameters of the pupil based on the corrected edge detection results.

[0063] First, the region of interest (ROI) of the pupil is located from the dynamic pupil image to eliminate irrelevant background interference and focus on the core analysis object. Then, combined with pupil response data, the average brightness of pixels in the region (reflecting the overall illumination level of the region) and the standard deviation of brightness (reflecting the degree of brightness difference in the region) are calculated, transforming the image information into quantifiable brightness features. Subsequently, an adaptive threshold adjustment strategy is used to generate high and low thresholds required for edge detection. The high threshold is set by the pupil diameter compensation coefficient combined with the average brightness plus a first multiple of the standard deviation of brightness, while the low threshold is set by the texture complexity compensation coefficient combined with the average brightness plus a second multiple of the standard deviation of brightness (with the first multiple being greater than the second multiple), ensuring that the thresholds are adapted to the actual state of the pupil. Then, edge detection is performed on the ROI based on these high and low thresholds, while eye movement data is used to correct detection deviations caused by eye movement. Finally, key feature parameters such as pupil diameter and constriction / dilation speed are extracted from the corrected edge detection results to provide data support for subsequent physiological state analysis and electrical stimulation modulation.

[0064] Specifically, a dynamic threshold generation mechanism takes as input the brightness features (mean, standard deviation, information entropy) of the pupil ROI region in an infrared image and outputs high and low thresholds.

[0065] One adjustment rule:

[0066] in, Pupil diameter compensation coefficient (diameter d↓→k_1↑ suppresses specular noise) Texture complexity compensation coefficient (entropy E↑→k_2↑ enhances weak edge response) This represents the average brightness of all pixels within the defined region of interest (ROI) in the infrared image of the current frame. It is a direct quantification of the current ambient light level and is the primary basis for threshold adjustment.

[0067] : Represents the standard deviation of the brightness of all pixels within the defined region of interest (ROI) in the infrared image of the current frame. It reflects the dispersion or contrast of brightness within that region.

[0068] Thigh: This is the dynamically calculated high threshold in the Canny edge detection operator. Pixels with gradient strengths higher than this value are immediately identified as strong edge pixels (most likely pupil edges).

[0069] Tlow: This is a dynamically calculated low threshold in the Canny edge detection operator. Pixels with gradient strengths below this value are suppressed (not considered edges). Pixels with gradient strengths between Tlow and Thigh are marked as weak edge pixels and are only preserved if they are connected to strong edge pixels.

[0070] Furthermore, before performing pupil edge detection on the region of interest based on the high and low thresholds in step 24, and correcting the edge detection results by combining eye movement data, the ambient lighting conditions can be determined based on the pupil response data and the average brightness. If the ambient lighting conditions are strong light, the high and low thresholds are adjusted upwards; if the ambient lighting conditions are weak light, the high and low thresholds are adjusted downwards.

[0071] Specifically, one ambient light compensation scheme is as follows:

[0072] This embodiment uses the above scheme as an example only. The specific implementation of other schemes can be referred to the description in this embodiment, and no limitation is made here.

[0073] Furthermore, the threshold can be increased for certain areas, such as the area covered by eyelashes, to enhance anti-interference capabilities. This embodiment does not limit this.

[0074] Furthermore, the traditional eight-direction NMS (i.e., processing gradients in eight directions: 0°, 45°, 90°, 135°, and the opposite directions) has two major drawbacks in pupil monitoring: over-refinement leading to edge breakage and gradient direction discretization not matching pupil geometry. To address this, this embodiment proposes abandoning the traditional eight directions and retaining only four main directions: 0° (horizontal), 45° (lower right to upper left diagonal), 90° (vertical), and 135° (upper right to lower left diagonal). Since the gradient directions at the pupil edge are mainly concentrated in these four main directions (the tangent directions of the elliptical contour mostly coincide with the main directions), there is no need for additional subdivision of the opposite directions.

[0075] Specifically, the actual gradient direction of the pupil edge pixels can be calculated first, and then mapped to the nearest principal direction (for example, if the actual gradient direction is 10°, it is assigned to 0°; if it is 40°, it is assigned to 45°). Only the local maxima with the largest gradient intensity in each principal direction are retained as valid edge points. This design avoids over-refinement in eight directions, reduces edge breakage, and makes the gradient direction determination more consistent with the elliptical geometry of the pupil through principal direction matching.

[0076] Furthermore, this embodiment proposes to detect isolated edge points (i.e., single pixels not connected to other edge points) in the edge detection results using an algorithm, and then perform interpolation calculations along the tangent direction of the ellipse based on the preset geometric model of the pupil ellipse to supplement the missing edge pixels and connect the fragmented edges into a continuous contour. At the same time, artifact removal is performed, such as directly filtering isolated edge chains (i.e., short fragment edges) with a length of less than 5 pixels. These edges are mostly pseudo-edges caused by eyelash obscuring or noise interference, rather than real pupil edges. Filtering them can further improve the purity of edge detection.

[0077] Furthermore, based on the physiological characteristics of different anatomical regions of the eye, different weight priorities can be assigned to the four directions—within a certain region, the edge points corresponding to the directions with higher priority will be retained first, ensuring that the edge detection results fit the real structure of the eye.

[0078] Specifically, one method of weight adjustment guided by anatomical structure is as follows:

[0079] Accordingly, after acquiring infrared sensing data, the following steps are taken to calculate the characteristic parameters of the wearer's pupil based on the pupil dynamic image, pupil reaction data, and eye movement data: 1. Preprocessing: Infrared and visible light image fusion (infrared sensor (8)), adaptive histogram equalization, motion blur compensation; 2. Edge detection: Canny operator (primary choice), dynamic threshold: adjust the high and low thresholds adaptively through infrared image brightness, non-maximum suppression direction optimization (0° / 45° / 90° / 135°); 3. Pupil fitting: Ellipse fitting (least square method); 4. Diameter calculation: Subpixel accuracy optimization: improve the edge positioning accuracy to 0.1 pixel level (corresponding to sub-millimeter physical accuracy) through interpolation or gradient subpixel positioning (such as Lucas-Kanade method). Diameter output: take the major axis of the ellipse as the pupil diameter, and combine it with time series analysis of dynamic changes; 5. Dynamic feature extraction: capture pupil oscillation, contraction / expansion speed, light reflection latency, and diameter change variance using the time difference method.

[0080] In this embodiment, by focusing on the region of interest, quantifying brightness features, and correcting motion deviations, the influence of background interference, illumination fluctuations, eye movement, and other factors on the detection is effectively eliminated. This ensures that the extracted pupil feature parameters can truly reflect the actual state of the patient's pupils, providing accurate data support for subsequent physiological state assessment (such as the risk assessment of depression and epilepsy) and adjustment of electrical stimulation parameters, and avoiding control errors caused by parameter deviations.

[0081] Example 6: This embodiment relates to a control device for intelligent ear vagus nerve modulation glasses. A schematic diagram of the control device for this embodiment is shown below. Figure 2 As shown, it includes: a data acquisition module 201, a status analysis module 202, an instruction generation module 203, and a control adjustment module 204.

[0082] The data acquisition module 201 is used to acquire the dynamic pupil feature parameters monitored by the smart ear vagus nerve modulation glasses; wherein, the dynamic pupil feature parameters are obtained by using a miniature camera installed on the lens of the smart ear vagus nerve modulation glasses to call real-time eye-tracking technology combined with edge detection algorithm to dynamically monitor the pupil diameter at the sub-millimeter level. The state analysis module 202 calls a trained machine learning model to analyze real-time physiological state information based on the pupil dynamic feature parameters; The instruction generation module 203 is used to generate an electrical stimulation control instruction corresponding to the real-time physiological state information based on the analyzed real-time physiological state information. The control and adjustment module is used to control the vagus electrodes in the smart ear vagus nerve modulation glasses to perform neural modulation based on the electrical stimulation control command.

[0083] It should be noted that the contents of the control device of the intelligent ear vagus nerve modulation glasses provided in this embodiment can be referred to in conjunction with the control method of the intelligent ear vagus nerve modulation glasses provided in the above embodiments, and the repeated parts will not be described again in this embodiment.

[0084] Furthermore, it should be noted that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, to highlight the innovative aspects of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed in this application; however, this does not mean that other units do not exist in this embodiment.

[0085] In the control device of the intelligent auricular vagus nerve modulation glasses provided in this embodiment, the data acquisition module relies on a miniature camera on the lens, combined with real-time eye-tracking technology and edge detection algorithms, to accurately acquire sub-millimeter-level dynamic images of the pupil, providing a high-precision data foundation for subsequent analysis; the feature calculation module extracts pupil feature parameters based on the image, realizing the quantitative transformation of pupil dynamics; the state analysis module calls a machine learning model to deeply correlate feature parameters with physiological states, ensuring the scientificity and accuracy of real-time physiological state analysis; the instruction generation module generates electrical stimulation control instructions based on the analysis results, realizing a precise response to auricular vagus nerve modulation. Through the close integration of each module, the overall device deeply integrates high-precision pupil monitoring with intelligent analysis and precise modulation, ensuring both the reliability of physiological state judgment and the targeted nature of electrical stimulation modulation. This effectively supports the efficient operation of the monitoring-analysis-treatment closed-loop system, providing a more precise and intelligent treatment control solution for patients undergoing neuromodulation, meeting the needs of non-invasive and individualized clinical applications.

[0086] Example 7: This embodiment proposes an intelligent ear vagus nerve modulation glasses. The three-dimensional structural schematic diagram of the intelligent ear vagus nerve modulation glasses in this embodiment is shown below. Figure 3 As shown in the diagram, the inner structure is as follows: Figure 4 As shown, Figure 3 and Figure 4 The reference numerals in the attached figures represent: 1 miniature camera, 2 control device for smart ear vagus nerve modulation glasses, 3 battery module, 4 LCD screen, 5 ear vagus electrode, 6 electrode wire, 7 power switch, 8 infrared sensor, and 9 electrode base.

[0087] The intelligent ear vagus nerve modulation glasses of this embodiment adopt the form of a traditional eyeglass frame, including a frame and a control device for the intelligent ear vagus nerve modulation glasses.

[0088] The frame includes a frame, lenses, and temples. The control device for the smart ear vagus nerve modulation glasses is installed in the frame. Each lens is equipped with a miniature camera, which is connected to the control device of the smart ear vagus nerve modulation glasses.

[0089] The frame is located at the front of the glasses. On one hand, it is used to fix the two lenses to ensure that the lens position is adapted to the wearer's eyes and to ensure the accuracy of pupil monitoring. On the other hand, it is hinged to the temples on both sides, so that the temples can be opened and closed to fit different head shapes and are easy to store and carry.

[0090] The control device 2 of the intelligent vagus nerve modulation glasses is installed in the frame of the glasses. The specific installation location is not limited in this embodiment; it can be as follows: Figure 3 It is installed in the middle of the frame as shown.

[0091] The device has two lenses, each mounted within a frame. Each lens is equipped with a miniature camera 1, positioned close to the wearer's eye to directly capture the pupil's field of vision. The miniature camera 1 connects to the control device 2 of the smart ear vagus nerve modulation glasses. The miniature camera 1 uses real-time eye-tracking technology combined with edge detection algorithms to dynamically monitor the pupil diameter at sub-millimeter levels. It can capture subtle changes in pupil diameter, such as fluctuations of 0.1mm, in response to emotional or physiological states. The detected dynamic pupil images are then transmitted to the control device 2 of the smart ear vagus nerve modulation glasses, providing precise physiological signal input for electrical stimulation modulation.

[0092] In this embodiment, the miniature camera 1 uses real-time eye-tracking technology to locate the pupil edge and calculate its diameter. One calculation process is as follows: 1. Ellipse Fitting: The pupil edge is elliptical. Least squares ellipse fitting, such as in OpenCV's fitEllipse, is used to avoid edge breakage caused by eyelash obstruction. 2. Sub-pixel Accuracy Optimization: Through interpolation or gradient sub-pixel positioning, such as the Lucas-Kanade method, the edge positioning accuracy is improved to the 0.1 pixel level (corresponding to sub-millimeter physical accuracy). 3. Diameter Output: The major axis of the ellipse is taken as the pupil diameter, and dynamic changes are analyzed in conjunction with time series data. This embodiment only uses the above calculation process as an example. Implementation processes based on other algorithms can refer to this embodiment, but are not limited thereto. In addition, this embodiment does not limit the specific edge detection algorithm used in the miniature camera (1). To deepen understanding, the Canny algorithm is used as an example. One calculation step is as follows: 1. Gradient calculation: Use a 5×5 or 3×3 Sobel operator (with moderate computational load) to obtain the horizontal and vertical gradients (Gx, Gy); 2. Non-maximum suppression: Only retain the edge points with local maximum values ​​in the gradient direction to ensure that the edge is refined to a single pixel; 3. Dual threshold connection: High threshold, strong edge (determines the pupil boundary); low threshold, weak edge (the part connected to the strong edge is retained). Of course, other algorithms can also be used, and all can refer to the description in this embodiment, which will not be repeated here.

[0093] Accordingly, the control device 2 of the intelligent vagus nerve modulation glasses is used to receive dynamic pupil images and calculate the wearer's pupil feature parameters. These feature parameters may include, but are not limited to: pupil diameter, pupil diameter change rate, reflex characteristic parameters (such as pupillary light reflex latency), fluctuation pattern parameters (such as pupillary oscillation frequency), and stability parameters (such as pupil diameter change variance). Based on the feature parameters, the wearer analyzes the real-time physiological state and generates electrical stimulation control commands for modulating the vagus nerve. The control device 2 of the intelligent vagus nerve modulation glasses can analyze pupil data through image processing algorithms and machine learning models, automatically generating electrical stimulation control commands to achieve closed-loop control of monitoring-analysis-treatment, avoiding human intervention errors, and improving the personalization and accuracy of treatment plans.

[0094] In this embodiment, the connection method between the control device of the smart ear vagus nerve modulation glasses and the external electrical stimulation system is not limited. In order to ensure the stability and real-time performance of command transmission, at least one temple is provided with an electrode wire 6 connected to the control device 2 of the smart ear vagus nerve modulation glasses. The other end of the electrode wire 6 is provided with an interface structure for establishing a signal connection with the electrical stimulation system. When the control device 2 of the smart ear vagus nerve modulation glasses analyzes the wearer's real-time physiological state (such as depressive tendencies, autonomic nervous system imbalance, etc.) through pupil dynamic monitoring and generates electrical stimulation control commands (including parameters such as stimulation frequency, pulse width, and waveform), the electrical stimulation control commands can be transmitted to the electrical stimulation system through the interface structure to drive the electrical stimulation system to perform electrical stimulation operation on the wearer's ear vagus nerve.

[0095] The electrode wire 6 is built into the temple, forming a functional integration design with the battery module 3 inside the temple and the external switch 7, etc. This eliminates the need for an additional independent command transmitter or connecting wire. At the same time, it ensures that the stimulation parameters output by the control device 2 of the smart ear vagus nerve modulation glasses are accurately transmitted to the electrical stimulation system, avoiding insufficient stimulation due to command distortion, and ensuring the safety and effectiveness of treatment.

[0096] The specific type of device selected for the miniature camera 1 is not limited in this embodiment. To reduce the size and weight of the device, a near-eye embedded miniature optical module can be used. Simultaneously, the miniature camera 1 can be connected to the control device 2 of the smart ear vagus nerve modulation glasses via an integrated circuit. This ensures image quality while reducing camera size, lens thickness, and wearing comfort, and also shortens data transmission latency to guarantee real-time monitoring. Specifically, the near-eye embedded miniature optical module can use a high-resolution miniature camera of 1280x800, such as the OmniVision OV9281, with a frame rate of 60fps. The miniature camera is mounted at the front of the glasses (i.e., near the inner side of the lens), directly facing the wearer's eyes. This position ensures that the miniature camera can obtain the optimal viewing angle of the wearer's eyes. This allows the camera to accurately capture pupil size, ensuring high accuracy and stability of the data.

[0097] The glasses have two temples, each connected to the frame. Their core functions are to power the device, provide a human-machine interface for electrical stimulation control, and link with the control device 2 of the smart ear vagus nerve modulation glasses inside the frame via internal components. The two temples extend from the posts on both sides of the frame, wrap around the ears to secure the glasses. The posts of the frame are hinged to the temples, allowing the temples to open and close for easy storage and carrying. They also adapt to different head shapes, improving the ergonomics of the device and reducing discomfort during prolonged wear.

[0098] At least one temple has a battery module 3 installed on its inner side. The battery module 3 is electrically connected to the control device 2 of the smart ear vagus nerve modulation glasses inside the frame and the miniature camera 1 on the lens to provide power and ensure the continuous operation of the built-in components such as the control device and the miniature camera of the smart ear vagus nerve modulation glasses.

[0099] In addition, at least one power switch 7 is installed on the outside of the temple. The power switch 7 is connected to the battery module 3 and is used to control the power supply to realize the start and stop operation of the device.

[0100] Furthermore, an LCD screen 4 can be installed on the outer side of at least one temple. The LCD screen 4 is connected to the control device 2 of the intelligent vagus nerve modulation glasses to display the battery module 3's power information or the parameter information of the electrical stimulation control command, or simultaneously display both power and parameter information. The LCD screen displays the power and stimulation parameters in real time; it supports Bluetooth / WiFi wireless transmission, enabling interconnection with external devices for data synchronization and remote control. This design integrates pupil monitoring, intelligent analysis, and electrical stimulation therapy, allowing users to intuitively monitor the device status and treatment progress, and supports parameter adjustment as needed, improving ease of use and treatment confidence.

[0101] Based on the above introduction, the intelligent ear vagus nerve modulation glasses provided in this embodiment use traditional glasses as a carrier and the frame as the structural center, which not only realizes the connection and fixation between the lens and the temple, but also integrates the control device 2 of the intelligent ear vagus nerve modulation glasses to ensure centralized control functions; the lens is embedded with a miniature camera 1, which combines real-time eye-tracking technology and edge detection algorithm to achieve sub-millimeter-level dynamic monitoring of pupil diameter, which can capture subtle changes in the pupil and provide high-precision data support for physiological state analysis; after receiving the dynamic image of the pupil, the control device 2 of the intelligent ear vagus nerve modulation glasses generates electrical stimulation control commands by calculating feature parameters and analyzing real-time physiological state, thus integrating pupil monitoring with... The deep integration of intelligent analysis avoids errors caused by human intervention, forming an automated process of data collection and intelligent decision-making. Compared with the traditional control method that relies on manual parameter setting, it significantly improves the accuracy and individualized adaptation of neural modulation. At the same time, the battery module 3 installed on the temple provides unified power supply for the core power components such as the control device 2 and the miniature camera 1 of the intelligent ear vagus nerve modulation glasses, ensuring continuous operation of the equipment. The external switch 7 is directly connected to the battery module 3, realizing convenient control of power on and off. The device can be started and stopped without complicated operations, which lowers the threshold for use. At the same time, the modular energy design facilitates subsequent maintenance and replacement, further improving the practicality of the equipment and the user experience.

[0102] The modular design of the frame, lenses, and temples in this solution compactly integrates the control device, miniature camera, and other components of the intelligent ear vagus nerve modulation glasses. It deeply combines medical-grade pupil monitoring, intelligent analysis, and ear vagus nerve modulation functions with the wearable glasses form. It conforms to the conventional form of glasses, is lightweight and small, and is easy to wear and carry in daily life. It is suitable for various usage scenarios such as home and travel, and provides an innovative and feasible solution for the clinical translation and daily application of non-invasive neuromodulation technology.

[0103] To improve the accuracy and stability of the collected data, in one embodiment, an infrared sensor 8 can be further provided. The infrared sensor 8 can be installed inside the two lenses and close to the miniature camera 1, and its function is to assist in improving the accuracy of pupil monitoring.

[0104] In this embodiment, the infrared sensor and the miniature camera work together to enhance the ability to capture pupil and eye movement signals (such as improving detection stability in light-changing scenarios), which can reduce environmental interference and further improve the accuracy and reliability of pupil diameter detection.

[0105] Furthermore, a communication module can be further incorporated into the smart ear vagus nerve modulation glasses. The glasses can then wirelessly transmit information to external devices, including but not limited to mobile phones and tablets.

[0106] like Figure 5The diagram illustrates data transmission of a communication module that supports Bluetooth or WiFi. Bluetooth / WiFi communication enables data synchronization between the glasses and devices such as mobile phones and tablets (e.g., treatment records, pupil data reports). Users can remotely view data or adjust parameters via an app, providing technical support for telemedicine or home health management.

[0107] Example 8: This embodiment relates to an intelligent ear vagus nerve modulation system. A connection diagram of the intelligent ear vagus nerve modulation system in this embodiment is shown below. Figure 3 or Figure 4 As shown, it includes intelligent ear vagus nerve modulation glasses and an electrical stimulation system.

[0108] The electrical stimulation system includes a vagus electrode 5 for stimulating the cymba conchae of the ear. The vagus electrode 5 is connected to the control device 2 of the intelligent vagus nerve modulation glasses via electrode wires 6. The stimulation frequency of the vagus electrode 5 is 0–200 Hz, the stimulation waveform is 100–1000 μs, and the stimulation waveform can be any of the following: triangular wave, square wave, sawtooth wave, or noise wave. The electrical stimulation system uses cymba conchae electrodes to provide non-invasive stimulation with multiple waveforms at frequencies of 0–200 Hz and waveforms of 100–1000 μs, avoiding the risks of invasive treatment. It can be used as an adjunct therapy for neurological and psychological disorders, providing users with a safe and comfortable treatment method.

[0109] In addition, the electrical stimulation system also includes an electrode base 9 that can be worn in the cymba conchae and conchae cavity. The auricular abductor electrode 5 is connected to the electrode base 9. The electrode base 9 can be designed to conform to the shape of the cymba conchae and conchae cavity (such as arc-shaped, forked, etc.) and fits the side wall of the cymba conchae and conchae cavity with flexible materials such as silicone and memory foam to avoid electrode displacement due to head movement.

[0110] The electrical stimulation system achieves treatment through stimulation of the vagus nerve in the ear. Compared to invasive methods, it avoids surgical risks, reduces patient suffering, and is suitable for long-term or repeated treatment. Furthermore, the electrodes in the cymba conchae are precisely positioned, ensuring that the stimulation signal effectively targets the vagus nerve and minimizes interference with other tissues. With a frequency range of 0–200 Hz, a wavelength range of 100–1000 μs, and multiple waveform options, it can adapt to the treatment needs of various diseases (such as anxiety, depression, epilepsy, etc.), improving the compatibility of treatment plans.

[0111] The intelligent vagus nerve modulation system provided in this embodiment uses a miniature camera equipped with an edge detection algorithm to capture sub-millimeter-level dynamic changes in pupil diameter in real time and transmits the data to the control device of the intelligent vagus nerve modulation glasses inside the frame. The control device of the intelligent vagus nerve modulation glasses integrates image processing algorithms and machine learning models to analyze pupil signals in real time and automatically generate electrical stimulation modulation commands based on physiological characteristics, driving the electrical stimulation system inside the temples to perform non-invasive electrical stimulation on the vagus nerve. The energy management system adopts a modular design, with a battery module embedded in the inner side of the temples to provide continuous power to the miniature camera and the control device of the intelligent vagus nerve modulation glasses. An external switch can control the power on and off with one button, ensuring stable operation of the device in various environments such as home treatment and mobile scenarios, realizing the deep integration of medical-grade modulation technology and wearable devices.

[0112] It should be noted that the intelligent ear vagus nerve modulation glasses in the intelligent ear vagus nerve modulation system provided in this embodiment can be referred to in conjunction with the intelligent ear vagus nerve modulation glasses provided in the above embodiments, and the repeated parts will not be described again in this embodiment.

[0113] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.

Claims

1. A control device of intelligent ear vagus nerve regulation glasses, characterized in that, include: The data acquisition module is used to acquire the dynamic pupil feature parameters monitored by the smart ear vagus nerve modulation glasses; wherein, the dynamic pupil feature parameters are obtained by using a miniature camera installed on the lens of the smart ear vagus nerve modulation glasses to call real-time eye-tracking technology combined with edge detection algorithm to dynamically monitor the pupil diameter at the sub-millimeter level; The state analysis module is used to call a trained machine learning model to analyze real-time physiological state information based on the pupil dynamic feature parameters; The instruction generation module is used to generate electrical stimulation control instructions corresponding to the real-time physiological state information based on the analyzed real-time physiological state information. The control and adjustment module is used to control the vagus electrodes in the smart ear vagus nerve modulation glasses to perform neural modulation based on the electrical stimulation control command; The step of calling a trained machine learning model to analyze real-time physiological state information based on the pupil dynamic feature parameters includes: Effective features of pupil state changes are extracted from the pupil dynamic feature parameters; wherein, the effective features include at least one of the following: diameter fluctuation trend and light reflection latency change pattern; A lightweight LSTM machine learning model is invoked to predict physiological state information based on the mapping relationship between pupil dynamic features and physiological state information, generating real-time physiological state assessment results; wherein, the physiological state assessment results include physiological state type and state severity.

2. The control device of the intelligent ear vagus nerve regulation glasses according to claim 1, wherein, Also includes: The system receives updated pupil dynamic feature parameters in real time and generates updated physiological state assessment results based on the updated pupil dynamic feature parameters. By comparing the physiological state assessment results before and after the update, a state comparison result is generated; Based on the state comparison results, an electrical stimulation control command corresponding to the updated pupil dynamic characteristic parameters is generated.

3. The control device for the intelligent vagus nerve modulation glasses according to claim 1, characterized in that, Also includes: Read the built-in physiological safety threshold parameters; The physiological safety threshold parameters include at least one of the following: upper limit of stimulation intensity and upper limit of single continuous stimulation duration; Determine whether the control command exceeds the physiological safety threshold parameter; If the physiological safety threshold parameter is exceeded, the control command that exceeds the physiological safety threshold parameter will adjust the parameter to the threshold range.

4. The control device for the intelligent vagus nerve modulation glasses according to any one of claims 1 to 3, characterized in that, The acquisition of pupil dynamic feature parameters obtained by the intelligent vagus nerve modulation glasses includes: Acquire dynamic pupil images monitored by the intelligent ear vagus nerve modulation glasses; Acquire pupil response data and eye movement data obtained by the intelligent ear vagus nerve modulation glasses through infrared sensing; Based on the pupil dynamic image, the pupil response data, and the eye movement data, the wearer's pupil dynamic characteristic parameters are calculated.

5. The control device for the intelligent vagus nerve modulation glasses according to claim 4, characterized in that, The calculation of dynamic characteristic parameters of the wearer's pupil based on the pupil dynamic image, the pupil response data, and the eye movement data includes: Extract the region of interest (ROI) of the pupil from the dynamic pupil image; Based on the pupil response data, the brightness characteristics of the pixels within the region of interest of the pupil are calculated; the brightness characteristics include the average brightness and the standard deviation of brightness. An adaptive threshold adjustment strategy is used to calculate the high and low thresholds required for edge detection. In this strategy, the high threshold is set by combining the pupil diameter compensation coefficient with the sum of the average brightness and the brightness standard deviation at a first multiple. The low threshold is set by combining the texture complexity compensation coefficient with the sum of the average brightness and the brightness standard deviation at a second multiple. The first multiple is higher than the second multiple. Based on the high threshold and the low threshold, edge detection of the pupil is performed on the region of interest of the pupil, and the edge detection results are corrected by combining the eye movement data; The feature parameters of the pupil are extracted based on the corrected edge detection results.

6. The control device of the intelligent ear vagus nerve regulation glasses according to claim 5, wherein, Before performing pupil edge detection on the region of interest based on the high threshold and the low threshold, and correcting the edge detection result by combining the eye motion data, the method further includes: Ambient lighting conditions are determined based on the pupil response data and the average brightness value. If the ambient lighting conditions are strong light environments, adjust the high threshold and the low threshold upwards; If the ambient lighting conditions are low light conditions, adjust the high threshold and the low threshold downwards.

7. Intelligent ear vagus nerve modulation glasses, characterized in that, The glasses include a frame and a control device for the smart ear vagus nerve modulation glasses as described in any one of claims 1-6. The frame includes: a frame, lenses, and temples, and the control device of the smart ear vagus nerve modulation glasses is installed in the frame; Each of the lenses is equipped with a miniature camera, which is connected to the control device of the smart ear vagus nerve modulation glasses.

8. An intelligent ear vagus nerve modulation system, comprising: include: The intelligent ear vagus nerve modulation glasses as described in claim 7, and the electrical stimulation system connected to the control device of the intelligent ear vagus nerve modulation glasses; The electrical stimulation system is worn on the wearer's ear and is used to deliver electrical stimulation to the wearer's vagus nerve.

Images