Method, apparatus, device and medium for safety risk control of ophthalmic optical equipment

Abstract

The application discloses a safety risk control method, device, equipment and medium of an ophthalmic optical device, and belongs to the technical field of optical devices. The method comprises the following steps: acquiring multi-modal data of a user's eye in the process of irradiating the user's eye according to target parameters by means of an irradiation module, wherein the multi-modal data is data for evaluating phototoxicity; for each modal data, extracting a feature vector of the modal data by means of an encoder corresponding to the modal data; fusing the feature vectors of the multi-modal data by means of a cross-modal attention fusion module to obtain fused features, and determining a phototoxicity risk score corresponding to the fused features as a current risk score; and adjusting the target parameters according to the current risk score. The application can reduce the phototoxicity risk and improve the safety in the process of photobiomodulation.

Description

Technical Field

[0001] This application belongs to the field of optical equipment technology, specifically relating to a method, device, equipment, and medium for controlling safety risks in ophthalmic optical equipment. Background Technology

[0002] Photobiological modulation using red light (630-680nm) and near-infrared light (810-850nm) at power densities of 8-50 mW / cm² has been shown to significantly control the progression of myopia in children and adolescents. However, there is a risk of phototoxicity to the eyes during the application of optical devices with this type of photobiological modulation, therefore monitoring of this risk is necessary.

[0003] In existing technologies, safety assessments of phototoxicity generally rely on static threshold lookup tables based on photobiological safety standards (such as IEC 62471). This approach fails to account for the increased actual light dose received by the retina due to the higher lens transparency (15-25% higher than adults) in children aged 6-18. Furthermore, this method ignores individual differences. The axial length (22-28 mm), pupil diameter (3-8 mm), and macular pigment density (0.1-0.8 OD) of different children significantly affect the actual retinal light dose, with individual differences reaching more than three times. This means that static threshold lookup tables cannot meet the safety requirements of different children, potentially leading to phototoxicity risks and resulting in low safety. Summary of the Invention

[0004] The purpose of this application is to provide a method, apparatus, device, and medium for controlling safety risks of ophthalmic optical devices, which can solve the problems of phototoxicity risks and low safety in the prior art.

[0005] To solve the above-mentioned technical problems, this application is implemented as follows: In a first aspect, embodiments of this application provide a method for controlling safety risks of ophthalmic optical devices, the method comprising: During the process of illuminating the user's eyes according to the target parameters through the illumination module, multimodal data of the user's eyes are acquired. The multimodal data is used to evaluate phototoxicity. For each modal data, the feature vector of the modal data is extracted through the encoder corresponding to the modal data; The feature vectors of the multimodal data are fused by the cross-modal attention fusion module to obtain fused features, and the phototoxicity risk score corresponding to the fused features is determined as the current risk score; The target parameters are adjusted based on the current risk score.

[0006] Secondly, embodiments of this application provide a safety risk control device for ophthalmic optical equipment, comprising: The data acquisition module is used to acquire multimodal data of the user's eyes during the process of illuminating the user's eyes according to the target parameters through the illumination module. The multimodal data is used to evaluate phototoxicity. The feature extraction module is used to extract the feature vector of each modal data through the encoder corresponding to the modal data. The risk assessment module is used to fuse the feature vectors of the multimodal data through the cross-modal attention fusion module to obtain fused features, and determine the phototoxicity risk score corresponding to the fused features as the current risk score; The parameter adjustment module is used to adjust the target parameters based on the current risk score.

[0007] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the first aspect.

[0008] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0009] Fifthly, embodiments of this application provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0010] In this embodiment, during the process of illuminating the user's eyes according to target parameters via the illumination module, multimodal data of the user's eyes is acquired. For each modal data, the feature vector of the modal data is extracted by the encoder corresponding to that modal data. The feature vectors of the multimodal data are fused by the cross-modal attention fusion module to obtain fused features, and the phototoxicity risk score corresponding to the fused features is determined as the current risk score. The target parameters are adjusted according to the current risk score. In this way, the user's multimodal data can be acquired in real time during the illumination output process, and phototoxicity risk assessment can be performed based on the multimodal data. This can improve the accuracy of phototoxicity risk assessment. Furthermore, the illumination output parameters of the illumination module can be adjusted in real time based on the risk assessment results, thereby reducing phototoxicity risk and improving the safety of the photobiological regulation process. Attached Figure Description

[0011] Figure 1This is a flowchart of a safety risk control method for ophthalmic optical devices provided in an embodiment of this application; Figure 2 This is a flowchart of a safety risk control method for ophthalmic optical devices provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a safety risk control device for ophthalmic optical equipment provided in an embodiment of this application; Figure 4 A schematic diagram of the hardware structure of an electronic device implementing an embodiment of this application. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0014] The safety risk control method, apparatus, device, and storage medium for ophthalmic optical equipment provided in this application will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios.

[0015] Figure 1 This is a flowchart illustrating a safety risk control method for ophthalmic optical devices according to an embodiment of this application. This method can be applied to scenarios where phototoxicity that may occur during illumination is assessed and controlled within ophthalmic optical devices. The ophthalmic optical device can be a myopia control device, or it can be other types of ophthalmic optical devices. Figure 1 As shown, the safety risk control method for the ophthalmic optical device may include steps 110 to 140.

[0016] Step 110: During the process of illuminating the user's eyes according to the target parameters through the illumination module, multimodal data of the user's eyes is acquired. The multimodal data is used to evaluate phototoxicity.

[0017] The illumination module may include a red LED array and a near-infrared LED array. For the red LED array, the peak wavelength can be 660±5nm, the spectral bandwidth can be less than or equal to 20nm, and an exemplary array configuration can be a 6×6 matrix (i.e., 36 LEDs). The single-LED radiant power can be 50mW, the power density for the treatment surface can be adjustable within 8-50 mW / cm², the power density adjustment step can be 0.1 mW / cm², the typical drive current is 350mA, and the maximum drive current is 700mA. For the near-infrared LED array, the peak wavelength can be 830±10nm, the spectral bandwidth can be less than or equal to 30nm, and an exemplary array configuration can be a 4×4 matrix (i.e., 16 LEDs). The single-LED radiant power can be 150mW, the power density for the treatment surface can be adjustable within 8-50 mW / cm², the typical drive current is 500mA, and the maximum drive current is 1000mA. Optionally, the illumination module may also include a ring-shaped illumination optics component for providing a ring-shaped beam pattern in the central zero-intensity macular protection zone. The inner diameter of the ring-shaped illumination optics component can be adjusted from 8 mm to 6-12 mm, and the ring-shaped illumination optics component can prevent light from shining into the macular protection zone.

[0018] Phototoxicity is a skin or tissue damage reaction caused by specific chemicals excited by ultraviolet light. Its core mechanism lies in the direct destruction of cell structure by photosensitive substances after absorbing light energy. During the process of irradiating the eyes with light through a light module, the photosensitive substances in the eyes may absorb light energy and cause phototoxicity. The embodiments of this application can be used for risk assessment and safety control.

[0019] During the process of controlling the illumination module to illuminate the user's eyes according to target parameters for photobiological modulation, multimodal data of the user's eyes can be collected at a certain frequency using a multimodal sensor. This multimodal data can include spectral data, ocular surface temperature data, ocular biometric data (such as pupil diameter), and environmental parameters.

[0020] The multimodal sensor may include a real-time spectral monitoring sensor, a pupil monitoring camera, and an ocular surface temperature sensor. For example, the real-time spectral monitoring sensor may employ an integrating sphere photodetector with a spectral range of 400-1100 nm and a power range of 50 nW-50 mW; the pupil monitoring camera may employ an infrared-enhanced CMOS (Complementary Metal Oxide Semiconductor) with a resolution of 1920×1080 pixels, and the infrared illuminator may be a 940 nm LED ring light (invisible to the user), with a pupil detection accuracy of ±0.1 mm diameter; the ocular surface temperature sensor may employ an uncooled microbolometer with a resolution of 160×120 pixels, a NETD (Noise Equivalent Temperature Difference) less than or equal to 50 mK, an accuracy of ±0.1°C, a frame rate of 8.6 Hz, and a spectral band of 8-14 μm (LWIR).

[0021] Step 120: For each modal data, extract the feature vector of the modal data through the encoder corresponding to the modal data.

[0022] Feature vectors from various modalities can be extracted and fused using a multi-input fusion neural network to perform phototoxicity risk scoring based on the fused features. The input fusion neural network may include an encoder corresponding to each modality, a cross-modal attention fusion module, and a risk scoring prediction head.

[0023] Multimodal data are input into the encoder corresponding to each modality. For each modality, the encoder corresponding to that modality extracts features to obtain the feature vector of that modality.

[0024] Step 130: The feature vectors of the multimodal data are fused using the cross-modal attention fusion module to obtain fused features, and the phototoxicity risk score corresponding to the fused features is determined as the current risk score.

[0025] The cross-modal attention fusion module employs a cross-modal attention mechanism to concatenate the feature vectors of multiple modal data into a single overall feature vector. This overall feature vector is then fused using a multi-layer (e.g., two-layer) Transformer encoder to obtain a fused feature. This fused feature is input into a risk scoring prediction head, which performs a phototoxicity risk score on the fused feature. For example, the cross-modal attention mechanism can use an 8-head self-attention mechanism to concatenate the outputs of multiple encoders into a 448-dimensional vector, which is then processed by a two-layer Transformer encoder to output a 256-dimensional fused feature.

[0026] Phototoxicity risk assessments can be performed at shorter intervals, such as every 100ms, allowing for timely monitoring of phototoxicity risks and further mitigation. The current risk score can be displayed in real-time on the user interface.

[0027] Step 140: Adjust the target parameters based on the current risk score.

[0028] The current risk score can be compared with a safety threshold, and based on the comparison result, target parameters can be adjusted, such as reducing power density or shutting off the illumination module, to avoid phototoxicity. Alternatively, based on the comparison result, the risk level corresponding to the current risk score can be determined, and target parameters can be adjusted based on that risk level, such as reducing power density or shutting off the illumination module, to avoid phototoxicity. Optionally, the age range of 6-18 years can be divided into several different age groups, with higher safety thresholds for older age groups and stricter safety thresholds for younger age groups.

[0029] For example, when the current risk score is greater than the first scoring threshold and less than or equal to the second scoring threshold, the power density in the target parameter is reduced by a first proportion; when the current risk score is greater than the second scoring threshold and less than or equal to the third scoring threshold, the power density in the target parameter is reduced by a second proportion; when the current risk score is greater than the third scoring threshold, the light output of the light module is turned off through a hardware interlock device.

[0030] The safety risk control method for ophthalmic optical devices provided in this application acquires multimodal data of the user's eyes during the process of illuminating the user's eyes according to target parameters through the illumination module. For each modal data, the feature vector of the modal data is extracted by the encoder corresponding to the modal data. The feature vectors of the multimodal data are fused by the cross-modal attention fusion module to obtain fused features, and the phototoxicity risk score corresponding to the fused features is determined as the current risk score. The target parameters are adjusted according to the current risk score. In this way, the user's multimodal data can be acquired in real time during the illumination output process, and phototoxicity risk assessment can be performed based on the multimodal data. This can improve the accuracy of phototoxicity risk assessment. Furthermore, the illumination output parameters of the illumination module can be adjusted in real time based on the risk assessment results, thereby reducing phototoxicity risk and improving the safety of the photobiological regulation process.

[0031] Figure 2 This is a flowchart of a safety risk control method for ophthalmic optical devices provided in this application embodiment. Based on the above embodiments, this embodiment can also predict the evolution trend of phototoxicity risk after the current time, and then adjust the target parameters based on the current risk score and the evolution trend of phototoxicity risk. For example... Figure 2As shown, the safety risk control method for the ophthalmic optical device may include steps 210 to 250.

[0032] Step 210: During the process of illuminating the user's eyes according to the target parameters through the illumination module, multimodal data of the user's eyes is acquired. The multimodal data is used to evaluate phototoxicity.

[0033] Step 220: For each modal data, extract the feature vector of the modal data through the encoder corresponding to the modal data.

[0034] Step 230: The feature vectors of the multimodal data are fused using the cross-modal attention fusion module to obtain fused features, and the phototoxicity risk score corresponding to the fused features is determined as the current risk score.

[0035] Step 240: Based on the current risk score, predict the evolution trend of phototoxicity risk within a target time period after the current time using the time-series risk prediction module. The evolution trend of phototoxicity risk includes the predicted risk scores at multiple time points within the target time period.

[0036] The current risk score, historical risk scores, and time-series data of multimodal data can be input into the risk prediction module. Based on this input data, the module can predict the risk score at multiple time points within a target timeframe after the current time, thus obtaining the predicted phototoxicity risk evolution trend. For example, the time-series risk prediction module can use a Transformer decoder architecture to predict the phototoxicity risk evolution trend for the next 5-30 seconds, such as predicting the phototoxicity risk evolution trend for the next 5 seconds, 15 seconds, and 30 seconds.

[0037] Step 250: Adjust the target parameters based on the current risk score and each of the predicted risk scores.

[0038] The system can determine the current risk score and the highest risk score among all predicted risk scores. This highest risk score is compared to a safety threshold, and based on the comparison result, target parameters are adjusted, such as reducing power density or shutting down the illumination module, to avoid phototoxicity. Alternatively, based on the comparison result, the risk level corresponding to the highest risk score is determined, and target parameters are adjusted accordingly, such as reducing power density or shutting down the illumination module, to avoid phototoxicity. When it is predicted that the risk score will exceed the safety threshold within a future target time (30 seconds), preventative power reduction is implemented in advance to avoid phototoxicity.

[0039] In some embodiments of this application, adjusting the target parameter based on the current risk score and each of the predicted risk scores includes: obtaining a level score threshold corresponding to the user's age, the level score threshold including score thresholds corresponding to multiple risk levels respectively; determining the target risk level corresponding to the current risk score and the predicted risk score respectively based on the level score threshold; and adjusting the target parameter according to the highest risk level among the target risk levels.

[0040] Risk rating thresholds can be set separately for different age groups, with lower thresholds for younger age groups. For example, risk levels, from lowest to highest, could be: Attention, Warning, Danger, and Emergency. For the 6-9 age group, the thresholds could be: Attention 40-50, Warning 55-70, Danger 70-85, and Emergency 85 or higher. For the 10-14 age group, the thresholds could be: Attention 50-65, Warning 65-80, Danger 80-90, and Emergency 90 or higher. For the 15-18 age group, the thresholds could be: Attention 55-70, Warning 70-85, Danger 85-95, and Emergency 95 or higher.

[0041] The system determines the user's age group and obtains the corresponding risk rating threshold. The current and predicted risk scores are compared to this threshold to determine the target risk level for each score. The highest risk level is then identified, and the adjustment strategy corresponding to the highest risk level is implemented to reduce power density or shut down the illumination module. Generally, phototoxicity risk increases with illumination duration; therefore, the predicted risk score should be higher than the current risk score to allow for earlier adjustment of target parameters and thus reduce phototoxicity risk.

[0042] In some embodiments of this application, the multiple risk levels, in ascending order of risk, include: safe, warning, danger, and emergency. The adjustment of the target parameters according to the highest risk level among the target risk levels includes one of the following: when the highest risk level is warning, reducing the power density of the target parameters by a first proportion; when the highest risk level is danger, reducing the power density of the target parameters by a second proportion, the second proportion being greater than the first proportion; when the highest risk level is emergency, turning off the light output of the light module.

[0043] For example, a level of concern can be included between the safety and warning levels of the risk level, meaning that the risk level can include five levels.

[0044] When the highest risk level determined based on the current and predicted risk scores is "Warning," the power density in the target parameters can be reduced by a first percentage. When the highest risk level is "Dangerous," the power density in the target parameters can be reduced by a second percentage, which is greater than the first percentage, because the risk level is high at this point, and a greater reduction in power density is needed to reduce phototoxicity risk. When the highest risk level is "Emergency," the illumination output of the illumination module can be shut off via a hardware interlock, which can respond within 1 millisecond. For example, when the risk score reaches the "Warning" level, the light source power density is automatically reduced by 20%; when it reaches the "Dangerous" level, it is forcibly reduced by 50%; and when it reaches the "Emergency" level, the hardware interlock shuts off the light output within 1 millisecond.

[0045] By implementing target parameter adjustments corresponding to the highest risk level, the risk of phototoxicity can be further reduced, and the safety of photobiological regulation processes can be further improved.

[0046] In some embodiments of this application, determining the target risk level corresponding to the current risk score and the predicted risk score respectively based on the rating threshold may include: if the user's medication list includes photosensitizing drugs, lowering the rating threshold by a target value; and determining the target risk level corresponding to the current risk score and the predicted risk score respectively based on the rating threshold after the target value reduction.

[0047] Before using ophthalmic optical devices for photobiological modulation, users can input their medication list. If the user's medication list includes photosensitizing drugs, which would lead to increased light energy absorption, the determined risk level can be lowered by a target value to reduce the risk of phototoxicity. Based on the reduced risk level, the target risk level corresponding to the current risk score and the predicted risk score can be determined. For example, if a user's medication list contains photosensitizing drugs, the risk level threshold for the user's age group can be lowered by 15 points.

[0048] Based on the above technical solution, the multimodal data includes: spectral data, ocular biometric data, time-series monitoring data, and environmental parameters. The ocular biometric data includes pupil diameter, and the time-series monitoring data includes ocular surface temperature. The step of extracting the feature vector of each modal data using the encoder corresponding to the modal data includes: extracting the feature vector of the spectral data using a spectral feature encoder for the spectral data; extracting the feature vector of the ocular biometric data using a biomarker encoder for the ocular biometric data; extracting the feature vector of the time-series monitoring data using a time-series pattern encoder for the time-series monitoring data; and extracting the feature vector of the environmental parameters using an environmental perception encoder for the environmental parameters.

[0049] Spectral data can be real-time spectral vectors acquired currently. Ocular biometric data can include corneal thickness, lens density, macular pigment optical density (MPOD), choroidal thickness, pupil diameter, axial length, iris color index, and fitzpatrick type. It can also include cumulative sessions, cumulative dose, and last session interval. Temporal monitoring data can be a sequence of ocular surface temperatures acquired chronologically, such as ocular surface temperature data acquired at 10Hz over the last 60 seconds. Environmental parameters can include ambient light intensity, ambient temperature, ambient humidity, device power density, spot area, and coordinates of the spot center point.

[0050] Spectral feature encoders can employ a one-dimensional convolutional neural network structure, for example, a four-layer one-dimensional convolutional neural network with 64-128-256-512 channels, outputting a 128-dimensional spectral feature vector after global average pooling. Biomarker encoders can employ multilayer perceptrons, for example, a three-layer fully connected layer using the ReLU activation function. Temporal pattern encoders can employ a bidirectional long short-term memory network structure, with hidden layers, for example, using 256 units, processing multidimensional temporal data sampled at 10Hz over the last 60 seconds. Environmental perception encoders can employ lightweight fully connected layers, such as layers with 6→32→64 units, outputting 64-dimensional features.

[0051] By extracting features from the corresponding modal data using encoders for different modal data, the accuracy of feature extraction can be improved, thereby enhancing the accuracy of phototoxicity risk assessment.

[0052] Based on the above technical solution, the method may further include: before controlling the illumination module to illuminate, acquiring the baseline pupil diameter of the user's eye through a pupil camera, and acquiring the baseline ocular surface temperature of the user's eye through a temperature sensor, wherein the baseline pupil diameter and the ocular biometric data are used together as input data of the biomarker encoder, and the baseline ocular surface temperature and the time-series monitoring data are used together as input data of the time-series pattern encoder.

[0053] Before photobiomodulation of the user's eye, or when the user first uses the ophthalmic optical device, a user identifier can be obtained. Based on the user identifier, the system can retrieve the user's age, previous photobiomodulation records, cumulative dose, etc. The pupil camera is activated to acquire the baseline pupil diameter of the user's eye, and a temperature sensor acquires the baseline ocular surface temperature. The baseline pupil diameter and ocular biometric data serve as input data for a biomarker encoder. This allows the biomarker encoder to encode the difference between the pupil diameter in the ocular biometric data and the baseline pupil diameter, thereby capturing features related to pupil changes and phototoxicity risk, further improving the accuracy of phototoxicity risk assessment. The baseline ocular surface temperature and time-series monitoring data serve as input data for a time-series pattern encoder. This allows the time-series pattern encoder to encode the difference between each ocular surface temperature in the time-series monitoring data and the baseline ocular surface temperature, thereby capturing features related to ocular surface temperature changes and phototoxicity risk, further improving the accuracy of phototoxicity risk assessment.

[0054] Based on the above technical solution, the method may further include: after the illumination module finishes irradiation, generating a phototoxicity safety report, the phototoxicity safety report including a risk score time series curve and the multimodal data, the risk score time series curve including risk scores at different time points during the illumination process.

[0055] After the controlled illumination module's illumination ends, a risk score time-series curve can be generated based on the risk scores at different time points during the illumination process. This generates a phototoxicity safety report that includes the risk score time-series curve and multimodal data acquired at each time point during the illumination process, allowing users to understand whether there are phototoxic risks during their photobiological regulation. The phototoxicity safety report can also include peak risk factors and cumulative dose statistics. When the multi-input fusion neural network performs risk assessment and outputs the current risk score, it simultaneously outputs the peak risk factors from the multimodal data.

[0056] The safety risk control method for ophthalmic optical devices provided in this application embodiment can be executed through an embedded AI computing platform in the safety risk control system of the ophthalmic optical device. Optionally, the embedded AI computing platform includes a GPU with at least 1024 CUDA (Compute Unified Device Architecture) cores and 16GB of memory, configured to complete a single risk assessment cycle with a latency of no more than 50 milliseconds. The safety risk control system for ophthalmic optical devices may further include: the aforementioned illumination module, multimodal sensor data acquisition module, multi-level safety response control module, and personalized risk model update module; the multimodal sensor data acquisition module integrates a real-time spectral monitoring sensor, a pupil monitoring camera, and an ocular surface temperature sensor; the multi-level safety response control module includes a hardware interlock device to shut off the illumination output of the illumination module when the highest risk level is emergency. The components of the safety risk control system for ophthalmic optical equipment can be integrated into the shell of a desktop myopia control device. For example, the shell of this desktop myopia control device measures 350×280×200mm, is made of anodized aluminum alloy 6061-T6, and is fitted with a medical-grade ABS plastic cover, which complies with IEC60601-1 and IEC 60601-1-2 standards.

[0057] Optionally, the personalized risk model update module maintains a sensitivity coefficient (range 0.5-2.0) and a cumulative effect weight parameter for each user. The sensitivity coefficient and cumulative effect weight parameter can be updated after each session (i.e., each photobiological regulation) using an online learning algorithm. That is, after each session, the user's sensitivity coefficient and cumulative effect weight are updated based on the actual results (risk score and target parameter adjustment results during photobiological regulation, etc.) using an online learning algorithm.

[0058] Optionally, the multimodal sensor acquisition module also includes an axial length measurement unit based on partial coherence interferometry, with an accuracy of ±0.01 mm, for personalized retinal dose estimation.

[0059] Optionally, the method may further include: transmitting the de-identified risk assessment data (de-identified multimodal data and corresponding risk scores) to a central server for global model optimization via a federated learning framework, with a privacy budget ε≤1.0.

[0060] In practice, the embodiments of this application can improve the accuracy of risk assessment, increasing it from 61% (i.e., the traditional method, static threshold lookup) to 96.5%; the assessment delay is <50ms, meeting the needs of real-time control; the personalized risk model adapts to individual differences among children; and predictive intervention reduces the incidence of phototoxic events by 85%.

[0061] It should be noted that the safety risk control method for ophthalmic optical devices provided in this application embodiment can be executed by an ophthalmic optical device safety risk control device, or a control module within that device for executing the safety risk control method for loading the ophthalmic optical device. This application embodiment uses the execution of the safety risk control method for loading the ophthalmic optical device by an ophthalmic optical device safety risk control device as an example to illustrate the safety risk control method for ophthalmic optical devices provided in this application embodiment.

[0062] Figure 3 This is a schematic diagram of the structure of a safety risk control device for ophthalmic optical equipment provided in an embodiment of this application, as shown below. Figure 3 As shown, the safety risk control device for this ophthalmic optical equipment includes: The data acquisition module 310 is used to acquire multimodal data of the user's eyes during the process of illuminating the user's eyes according to the target parameters through the illumination module. The multimodal data is used to evaluate phototoxicity. The feature extraction module 320 is used to extract the feature vector of each modal data through the encoder corresponding to the modal data. The risk assessment module 330 is used to fuse the feature vectors of the multimodal data through the cross-modal attention fusion module to obtain fused features, and determine the phototoxicity risk score corresponding to the fused features as the current risk score; The parameter adjustment module 340 is used to adjust the target parameters according to the current risk score.

[0063] Optionally, the device further includes: The time-series risk prediction module is used to predict the evolution trend of phototoxicity risk within a target time after the current time based on the current risk score. The phototoxicity risk evolution trend includes the predicted risk scores at multiple time points within the target time. The parameter adjustment module is specifically used for: The target parameters are adjusted based on the current risk score and each of the predicted risk scores.

[0064] Optionally, the parameter adjustment module includes: A threshold acquisition unit is used to acquire a level rating threshold corresponding to the user's age, wherein the level rating threshold includes rating thresholds corresponding to multiple risk levels respectively. A risk level determination unit is used to determine the target risk level corresponding to the current risk score and the predicted risk score respectively, based on the level scoring threshold. The parameter adjustment unit is used to adjust the target parameters according to the highest risk level among the target risk levels.

[0065] Optionally, the multiple risk levels, in ascending order, include: safe, warning, danger, and emergency. The parameter adjustment unit is used to perform one of the following: When the highest risk level is a warning, the power density in the target parameter is reduced by a first proportion; When the highest risk level is dangerous, the power density in the target parameter is reduced by a second proportion, where the second proportion is greater than the first proportion; When the highest risk level is emergency, the light output of the lighting module is turned off.

[0066] Optionally, the risk level determination unit is specifically used for: If the user's medication list includes photosensitizing drugs, the target value for the grade rating threshold will be lowered. Based on the rating threshold after reducing the target value, the target risk level corresponding to the current risk score and the predicted risk score are determined respectively.

[0067] Optionally, the multimodal data includes: spectral data, ocular biometric data, time-series monitoring data, and environmental parameters, wherein the ocular biometric data includes pupil diameter, and the time-series monitoring data includes ocular surface temperature; The feature extraction module is specifically used for: For the spectral data, a feature vector is extracted from the spectral data using a spectral feature encoder; For the aforementioned eye biometric data, a feature vector is extracted from the eye biometric data using a biomarker encoder; For the time-series monitoring data, a feature vector is extracted from the time-series monitoring data using a time-series pattern encoder; For the environmental parameters, feature vectors of the environmental parameters are extracted by an environmental perception encoder.

[0068] Optionally, the device further includes: The baseline data acquisition module is used to acquire the baseline pupil diameter of the user's eye through a pupil camera and the baseline ocular surface temperature of the user's eye through a temperature sensor before controlling the illumination module to perform illumination. The baseline pupil diameter and the ocular biometric data are used together as the input data of the biomarker encoder, and the baseline ocular surface temperature and the time-series monitoring data are used together as the input data of the time-series pattern encoder.

[0069] Optionally, the illumination module includes a ring-shaped illumination optics assembly for providing a ring-shaped beam pattern in the central zero-intensity macular protection zone.

[0070] Optionally, the device further includes: A safety report generation module is used to generate a phototoxicity safety report after the illumination module has finished irradiating the area. The phototoxicity safety report includes a risk score time series curve and the multimodal data. The risk score time series curve includes risk scores at different time points during the illumination process.

[0071] The safety risk control device for ophthalmic optical equipment in this application embodiment can be a device, or it can be a component, integrated circuit, or chip in the ophthalmic optical equipment. The device can be a mobile electronic device or a non-mobile electronic device. For example, mobile electronic devices can be mobile phones, tablets, laptops, PDAs, in-vehicle electronic devices, wearable devices, ultra-mobile personal computers (UMPCs), netbooks, or personal digital assistants (PDAs), etc., while non-mobile electronic devices can be servers, network attached storage (NAS), personal computers (PCs), televisions (TVs), ATMs, or self-service machines, etc. This application embodiment does not specifically limit the scope.

[0072] The safety risk control device for ophthalmic optical equipment provided in this application embodiment can achieve... Figures 1 to 2 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0073] The safety risk control device for ophthalmic optical equipment provided in this application acquires multimodal data of the user's eyes during the process of illuminating the user's eyes according to target parameters through the illumination module. For each modal data, the feature vector of the modal data is extracted by the encoder corresponding to the modal data. The feature vectors of the multimodal data are fused by the cross-modal attention fusion module to obtain the fused feature, and the phototoxicity risk score corresponding to the fused feature is determined as the current risk score. The target parameters are adjusted according to the current risk score. In this way, the user's multimodal data can be acquired in real time during the illumination output process, and phototoxicity risk assessment can be performed based on the multimodal data. This can improve the accuracy of phototoxicity risk assessment. Furthermore, the illumination output parameters of the illumination module can be adjusted in real time based on the risk assessment results, thereby reducing phototoxicity risk and improving the safety of the photobiological regulation process.

[0074] Optionally, this application also provides an electronic device, including a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the various processes of the above-described safety risk control method embodiment for ophthalmic optical devices and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0075] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0076] Figure 4 A schematic diagram of the hardware structure of an electronic device implementing an embodiment of this application.

[0077] The electronic device 400 includes, but is not limited to, components such as: a radio frequency unit 401, a network module 402, an audio output unit 403, an input unit 404, a sensor 405, a display unit 406, a user input unit 407, an interface unit 408, a memory 409, and a processor 410. The input unit 404 may include an image processor 4041 and a microphone 4042; the display unit 406 may include a display panel 4061; and the user input unit 407 may include a touch panel 4071 and / or other input devices (such as a keyboard) 4072.

[0078] Those skilled in the art will understand that the electronic device 400 may also include a power supply (such as a battery) for supplying power to various components. The power supply may be logically connected to the processor 410 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. Figure 4 The electronic device structure shown does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here. The memory 409 stores a program or instruction that can run on the processor. When the program or instruction is executed by the processor 410, it implements the various processes of the above-described embodiment of the safety risk control method for ophthalmic optical devices and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0079] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described embodiments of the safety risk control method for ophthalmic optical devices and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0080] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0081] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described safety risk control method embodiment for ophthalmic optical devices, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0082] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0083] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0084] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0085] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for controlling safety risks in ophthalmic optical devices, characterized in that, include: During the process of illuminating the user's eyes according to the target parameters through the illumination module, multimodal data of the user's eyes are acquired. The multimodal data is used to evaluate phototoxicity. For each modal data, the feature vector of the modal data is extracted through the encoder corresponding to the modal data; The feature vectors of the multimodal data are fused by the cross-modal attention fusion module to obtain fused features, and the phototoxicity risk score corresponding to the fused features is determined as the current risk score; The target parameters are adjusted based on the current risk score.

2. The method according to claim 1, characterized in that, Before adjusting the target parameter based on the current risk score, the method further includes: Based on the current risk score, the time-series risk prediction module predicts the evolution trend of phototoxicity risk within a target time period after the current time. The phototoxicity risk evolution trend includes the predicted risk scores at multiple time points within the target time period. The step of adjusting the target parameters based on the current risk score includes: The target parameters are adjusted based on the current risk score and each of the predicted risk scores.

3. The method according to claim 2, characterized in that, The step of adjusting the target parameters based on the current risk score and each of the predicted risk scores includes: Obtain a rating threshold corresponding to the user's age, wherein the rating threshold includes rating thresholds corresponding to multiple risk levels; Based on the rating threshold, the target risk levels corresponding to the current risk score and the predicted risk score are determined respectively. The target parameters are adjusted according to the highest risk level among the various target risk levels.

4. The method according to claim 3, characterized in that, The risk levels, in ascending order of severity, are: safe, warning, danger, and emergency. The adjustment of the target parameters according to the highest risk level among the target risk levels includes one of the following: When the highest risk level is a warning, the power density in the target parameter is reduced by a first proportion; When the highest risk level is dangerous, the power density in the target parameter is reduced by a second proportion, where the second proportion is greater than the first proportion; When the highest risk level is emergency, the light output of the lighting module is turned off.

5. The method according to claim 3, characterized in that, The step of determining the target risk level corresponding to the current risk score and the predicted risk score respectively based on the rating threshold includes: If the user's medication list includes photosensitizing drugs, the target value for the grade rating threshold will be lowered. Based on the rating threshold after reducing the target value, the target risk level corresponding to the current risk score and the predicted risk score are determined respectively.

6. The method according to any one of claims 1-5, characterized in that, The multimodal data includes: spectral data, ocular biometric data, time-series monitoring data, and environmental parameters. The ocular biometric data includes pupil diameter, and the time-series monitoring data includes ocular surface temperature. The step of extracting the feature vector of each modal data through the encoder corresponding to the modal data includes: For the spectral data, a feature vector is extracted from the spectral data using a spectral feature encoder; For the aforementioned eye biometric data, a feature vector is extracted from the eye biometric data using a biomarker encoder; For the time-series monitoring data, a feature vector is extracted from the time-series monitoring data using a time-series pattern encoder; For the environmental parameters, feature vectors of the environmental parameters are extracted by an environmental perception encoder.

7. The method according to claim 6, characterized in that, Also includes: Before controlling the illumination module to provide illumination, the baseline pupil diameter of the user's eye is obtained through a pupil camera, and the baseline ocular surface temperature of the user's eye is obtained through a temperature sensor. The baseline pupil diameter and the ocular biometric data are used together as input data for the biomarker encoder, and the baseline ocular surface temperature and the time-series monitoring data are used together as input data for the time-series pattern encoder.

8. The method according to any one of claims 1-5, characterized in that, The illumination module includes a ring-shaped illumination optics component for providing a ring-shaped beam pattern in the central zero-intensity macular protection zone.

9. The method according to any one of claims 1-5, characterized in that, Also includes: After the illumination module finishes irradiation, a phototoxicity safety report is generated. The phototoxicity safety report includes a risk score time series curve and the multimodal data. The risk score time series curve includes the risk score at different time points during the illumination process.

10. A safety risk control device for ophthalmic optical equipment, characterized in that, include: The data acquisition module is used to acquire multimodal data of the user's eyes during the process of illuminating the user's eyes according to the target parameters through the illumination module. The multimodal data is used to evaluate phototoxicity. The feature extraction module is used to extract the feature vector of each modal data through the encoder corresponding to the modal data. The risk assessment module is used to fuse the feature vectors of the multimodal data through the cross-modal attention fusion module to obtain fused features, and determine the phototoxicity risk score corresponding to the fused features as the current risk score; The parameter adjustment module is used to adjust the target parameters based on the current risk score.

11. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the safety risk control method for ophthalmic optical devices as described in any one of claims 1-9.

12. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the safety risk control method for ophthalmic optical devices as described in any one of claims 1-9.

Images