Self-adjusting liquid crystal display based on multi-source data analysis and adjusting method thereof

Abstract

The application relates to the field of display adjustment, in particular to a self-adjusting liquid crystal display based on multi-source data analysis and an adjusting method thereof, which comprises the following steps: constructing an eyeball recognition model based on a YOLOv8 framework, and constructing a comprehensive loss function, wherein a weighted Focal loss term is determined according to environmental light intensity and a real eyeball diameter, an eyeball boundary constraint term is determined based on the real eyeball diameter and a predicted eyeball diameter, a binocular constraint term is determined based on a binocular distance, and the predicted eyeball diameter and the binocular distance are determined by calculating an eyeball recognition area; determining an eyeball change area and a blink frequency value of a user through the eyeball recognition model; calculating an expected brightness value, and performing brightness self-adjusting control of the liquid crystal display based on the expected brightness value. The application realizes a self-adjusting liquid crystal display based on environment light and vision, can combine multi-source data such as a user eyeball state and environmental light intensity, realizes accurate brightness adjustment, and thus improves user experience and reduces visual fatigue.

Description

Technical Field

[0001] This invention relates to the field of display adjustment, and more particularly to a self-adjusting liquid crystal display based on multi-source data analysis and its adjustment method. Background Technology

[0002] With the rapid development of display technology, liquid crystal displays (LCDs) have become widely used in daily life and work. However, prolonged use of LCDs can lead to eye strain, especially when ambient light changes, as the monitor's brightness may not match the ambient light, causing discomfort to the eyes. Traditional methods of adjusting monitor brightness typically rely on manual adjustment or simple ambient light sensor feedback mechanisms, which are difficult to provide the optimal visual experience in complex and changing lighting environments.

[0003] In recent years, attempts have been made to achieve adaptive brightness adjustment of displays using ambient light sensors. However, this method typically relies on a single ambient light intensity signal and lacks comprehensive analysis of the user's visual state, making it difficult to achieve precise brightness adjustment.

[0004] Therefore, existing display brightness adjustment methods still have shortcomings in terms of comprehensive analysis of multi-source data and intelligent adjustment. There is an urgent need for self-adjusting liquid crystal displays based on multi-source data analysis, which can combine multi-source data such as user eye status and ambient light intensity to achieve brightness adjustment that improves visual experience and reduces visual fatigue from long-term use. Summary of the Invention

[0005] To address this, the present invention provides a self-adjusting liquid crystal display and its adjustment method based on multi-source data analysis. By constructing a comprehensive loss function, it suppresses false detections caused by strong or weak light interference, ensures the geometric consistency between the predicted eye diameter and the true value, avoids single-eye missed detections or deviations, and significantly improves the eye positioning accuracy of the YOLOv8 model under complex lighting conditions. This enables a self-adjusting liquid crystal display based on ambient light and vision, which can combine multi-source data such as user eye state and ambient light intensity to achieve precise brightness adjustment, thereby improving user experience and reducing visual fatigue.

[0006] To achieve the above objectives, this invention proposes an adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis, comprising:

[0007] The system acquires user images from the visual sensor of a self-adjusting liquid crystal display and ambient light intensity from the ambient light sensor. A sample test set is constructed, including the correspondence between user images and ambient light intensity. An eye recognition model based on the YOLOv8 framework is built. A comprehensive loss function with a weighted Focal loss term, an eye boundary constraint term, and a binocular constraint term is constructed. The eye recognition model is trained using the comprehensive loss function and the sample test set. The weighted Focal loss term is determined based on the ambient light intensity and the true diameter of the eyeball. The eye boundary constraint term is determined based on the true eyeball diameter and the predicted eyeball diameter. The binocular constraint term is determined based on the interocular distance. The predicted eyeball diameter and the interocular distance are determined by calculating the eye recognition area.

[0008] The current user image is used to generate a user eyeball annotation image through the eyeball recognition model, and the user's eyeball change area and blink frequency value are determined based on the eyeball annotation image;

[0009] The ambient light intensity, the change area of ​​the eyeball, and the blink frequency value are used to calculate the desired brightness value using a brightness conversion formula. Based on the desired brightness value, the brightness self-adjustment control of the liquid crystal display is performed.

[0010] Furthermore, the process of constructing the comprehensive loss function includes:

[0011] The weighted Focal loss term is constructed based on the product of the logarithmic function of the positive confidence of the eyeballs in the sample test set, the negative confidence of the eyeballs with the focusing parameter as the exponent, and the dynamic weighting coefficients.

[0012] Calculate the exponential function of the difference between the predicted eye diameter and the standard eye diameter, and construct the eye boundary constraint term based on the product of the exponential function and the eye expansion adjustment amount;

[0013] The binocular constraint term is constructed based on the activation function values ​​of the predicted difference in binocular diameter and the standard deviation of binoculars;

[0014] The comprehensive loss function is constructed by weighted summation of the weighted Focal loss term, the eyeball boundary constraint term, and the binocular constraint term.

[0015] Furthermore, the process of constructing the weighted Focal loss term includes:

[0016] The dynamic weighting coefficients are adjusted based on a comparison between the actual eye diameter of the sample test set used to train the eye recognition model and the normal upper limit threshold and normal lower limit threshold of the eye diameter.

[0017] The focusing parameters are adjusted based on a comparison of the ambient light intensity with the strong light threshold and the weak light threshold.

[0018] Furthermore, the process of constructing the eyeball boundary constraint term includes:

[0019] Obtain the standard eyeball diameter corresponding to neutral light in the sample test set, and the user's historical blink frequency corresponding to the blinking time sequence information;

[0020] The user adaptation delay time is calculated based on the user's historical blink frequency.

[0021] Calculate the ratio of the standard eye diameter to the actual eye diameter in the sample test set, and calculate the eye dilation adjustment amount with the user adaptation delay time as the exponent and the ratio as the base.

[0022] Furthermore, the process of constructing the binocular constraint term includes:

[0023] The difference between the predicted difference in binocular diameter and the standard deviation of both eyes is used to construct the binocular constraint term through the ReLU activation function value.

[0024] In the above scheme, the weighted Focal loss term dynamically adjusts the focusing parameters by ambient light intensity to solve the false detection problem caused by sudden changes in illumination. It adaptively adjusts the weights according to the actual diameter of the eyeball to improve the model's ability to generalize to physiological differences. The eyeball boundary constraint term introduces an exponential function to strengthen the penalty for the deviation between the predicted eyeball diameter and the standard diameter, ensuring the geometric consistency of pupil dilation changes. It also avoids over-adjustment by combining the user adaptation delay time. The binocular constraint term forces binocular collaborative detection to reduce the false negative rate.

[0025] Furthermore, the process of determining the user's eyeball change area and blink frequency value based on the eyeball annotation image includes:

[0026] The binary mask of the eye region is determined based on the eyeball annotation image, and the least squares method is used to fit an ellipse to the binary mask to calculate the eyeball diameter.

[0027] The change area of ​​the eyeball is calculated using the sensor size conversion formula;

[0028] If the current change area of ​​the eyeball is less than the average change area of ​​the eyeball in the previous set number of frames, it is determined to be a blinking event;

[0029] The blink frequency value is determined based on the number of blink events of a set duration.

[0030] Furthermore, the process of calculating the desired brightness value using the ambient light intensity, the change area of ​​the eyeball, and the blink frequency value through a brightness conversion formula includes:

[0031] Calculate the logarithm of the ambient light intensity plus the light intensity offset to determine the basic expected brightness value;

[0032] Based on the change in eyeball area and the blink frequency value, determine whether the eyeball is fatigued; if so, adjust the zoom sensitivity.

[0033] Calculate the ratio of the standard area of ​​the eyeball to the change area of ​​the eyeball, and calculate the brightness response gain value with the ratio of the change area of ​​the eyeball as the base and the scaling sensitivity as the exponent.

[0034] The sum of the base expected brightness value and the brightness response gain value is taken as the expected brightness value.

[0035] Furthermore, the process of performing brightness self-adjustment control of the liquid crystal display based on the desired brightness value includes:

[0036] The dynamic brightness factor is calculated by summing the Sigmoid function values ​​of the desired brightness value and the current brightness value with the user's habitual terms, so that the dynamic brightness factor forms an S-shaped transition.

[0037] The display brightness value is determined by multiplying the dynamic brightness factor, the user-set brightness value, and the current display mode fixed factor, and the brightness of the liquid crystal display is adjusted to the display brightness value.

[0038] Furthermore, the user habit item is determined based on the average historical brightness preference of users with time decay weight.

[0039] The present invention also provides a self-adjusting liquid crystal display based on multi-source data analysis, which is equipped with a visual sensor for data acquisition and an ambient light sensor, and is equipped with an adjustment method for the self-adjusting liquid crystal display based on multi-source data analysis.

[0040] The above scheme achieves dynamic calibration of basic brightness. The basic expected brightness is calculated by adding the logarithm of the ambient light intensity and the light intensity offset. This conforms to the human eye's perception characteristics of light intensity according to the Weber-Fechner law, avoids the problem of excessive brightness / darkness caused by linear adjustment, increases sensitivity under high fatigue conditions, and rapidly improves the brightness response gain value to relieve eye fatigue in a timely manner.

[0041] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0042] 1. By constructing a comprehensive loss function, false detections caused by strong or weak light interference are suppressed, ensuring geometric consistency between the predicted eye diameter and the true value, avoiding single-eye missed detections or deviations, and significantly improving the eye localization accuracy of the YOLOv8 model under complex lighting conditions. This enables a self-adjusting liquid crystal display based on ambient light and vision, which can combine multi-source data such as user eye status and ambient light intensity to achieve precise brightness adjustment, thereby improving user experience and reducing visual fatigue.

[0043] 2. The weighted Focal loss term dynamically adjusts the focusing parameters by ambient light intensity to solve the false detection problem caused by sudden changes in illumination. It adaptively adjusts the weights according to the actual diameter of the eyeball to improve the model's ability to generalize to physiological differences. The eyeball boundary constraint term introduces an exponential function to strengthen the penalty for the deviation between the predicted eyeball diameter and the standard diameter, ensuring the geometric consistency of pupil dilation and contraction changes. It also avoids over-adjustment by combining the user adaptation delay time. The binocular constraint term forces binocular collaborative detection to reduce the false negative rate.

[0044] 3. It achieves dynamic calibration of basic brightness. The basic expected brightness is calculated by adding the logarithm of ambient light intensity and light intensity offset. This conforms to the Weber-Fechner law of human eye perception of light intensity, avoids the problem of excessive brightness / darkness caused by linear adjustment, increases sensitivity under high fatigue conditions, and rapidly improves the brightness response gain value to relieve eye fatigue in a timely manner. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating the adjustment method of a self-adjusting liquid crystal display based on multi-source data analysis according to an embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram illustrating the process of constructing the comprehensive loss function for the self-adjusting liquid crystal display adjustment method based on multi-source data analysis, according to an embodiment of the present invention.

[0047] Figure 3 This is a schematic diagram illustrating the calculation process of the eyeball change area of ​​the liquid crystal display in the self-adjusting liquid crystal display adjustment method based on multi-source data analysis according to an embodiment of the present invention.

[0048] Figure 4 This is a schematic diagram of the process for determining the display brightness of a liquid crystal display in the self-adjusting liquid crystal display adjustment method based on multi-source data analysis, according to an embodiment of the present invention. Detailed Implementation

[0049] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0050] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0051] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0052] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0053] like Figures 1 to 4 As shown, this invention provides a self-adjusting liquid crystal display and its adjustment method based on multi-source data analysis. By constructing a comprehensive loss function, it suppresses false detections caused by strong or weak light interference, ensures the geometric consistency between the predicted eye diameter and the true value, avoids single-eye missed detections or deviations, and significantly improves the eye positioning accuracy of the YOLOv8 model under complex lighting conditions. This enables a self-adjusting liquid crystal display based on ambient light and vision, which can combine multi-source data such as user eye state and ambient light intensity to achieve precise brightness adjustment, thereby improving user experience and reducing visual fatigue.

[0054] like Figures 1 to 4 As shown, this embodiment proposes an adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis, including:

[0055] The system acquires user images from the visual sensor of a self-adjusting liquid crystal display and ambient light intensity from the ambient light sensor. A sample test set is constructed, including the correspondence between user images and ambient light intensity. An eye recognition model based on the YOLOv8 framework is built. A comprehensive loss function with a weighted Focal loss term, an eye boundary constraint term, and a binocular constraint term is constructed. The eye recognition model is trained using the comprehensive loss function and the sample test set. The weighted Focal loss term is determined based on the ambient light intensity and the true diameter of the eyeball. The eye boundary constraint term is determined based on the true eyeball diameter and the predicted eyeball diameter. The binocular constraint term is determined based on the interocular distance. The predicted eyeball diameter and the interocular distance are determined by calculating the eye recognition area.

[0056] The current user image is used to generate a user eyeball annotation image through the eyeball recognition model, and the user's eyeball change area and blink frequency value are determined based on the eyeball annotation image;

[0057] The ambient light intensity, the change area of ​​the eyeball, and the blink frequency value are used to calculate the desired brightness value using a brightness conversion formula. Based on the desired brightness value, the brightness self-adjustment control of the liquid crystal display is performed.

[0058] like Figure 2 As shown, the process of constructing the comprehensive loss function further includes:

[0059] The weighted Focal loss term is constructed based on the product of the logarithmic function of the positive confidence of the eyeballs in the sample test set, the negative confidence of the eyeballs with the focusing parameter as the exponent, and the dynamic weighting coefficients.

[0060] Calculate the exponential function of the difference between the predicted eye diameter and the standard eye diameter, and construct the eye boundary constraint term based on the product of the exponential function and the eye expansion adjustment amount;

[0061] The binocular constraint term is constructed based on the activation function values ​​of the predicted difference in binocular diameter and the standard deviation of binoculars;

[0062] The comprehensive loss function is constructed by weighted summation of the weighted Focal loss term, the eyeball boundary constraint term, and the binocular constraint term.

[0063] Specifically, the comprehensive loss function is:

[0064] L total =ω d ·FL(p t ,γ dynamic )+λ1L bound (d pred )+λ2L sym (d predleft ,d predright )

[0065] In the formula, L total Represents the comprehensive loss function, ω d ·FL(p t ,γ dynamic ) indicates that p is the positive confidence level. t and focusing parameter γ dynamic For the weighted Focal loss term of the variable, ω d L represents the dynamic weighting coefficient. bound (d pred ) indicates the prediction of eyeball diameter d pred L is the eyeball boundary constraint term for the variable. sym (d predleft ,d predright ) represents the predicted value d of the diameter of both eyes. predleft ,d predrightThe binocular constraint term is defined as a variable, with λ1 and λ2 being two weighting coefficients, preferably 0.7 and 0.3. Therefore, by constructing a comprehensive loss function that combines the weighted Focal loss term, the eye boundary constraint term, and the binocular constraint term, the eye localization accuracy of the YOLOv8 model under complex lighting conditions is significantly improved, the recognition error caused by sudden changes in lighting is reduced, and the localization error is significantly reduced.

[0066] Furthermore, the process of constructing the weighted Focal loss term includes:

[0067] The dynamic weighting coefficients are adjusted based on a comparison between the actual eye diameter of the sample test set used to train the eye recognition model and the normal upper limit threshold and normal lower limit threshold of the eye diameter.

[0068] The focusing parameters are adjusted based on a comparison of the ambient light intensity with the strong light threshold and the weak light threshold.

[0069] Specifically, the weighted Focal loss term is:

[0070]

[0071] In the formula, ω d ·FL(p t ,γ dynamic ) represents the weighted Focal loss term, ω d Represents the dynamic weighting coefficient, γ dynamic p represents the focusing parameter. t Indicates positive confidence level of the eyeball, 1-p t Indicates the negative confidence level of the eyeball, α t This indicates the category balance weight.

[0072] Specifically, the process of adjusting the dynamic weighting coefficients is as follows:

[0073]

[0074] In the formula, ω d d represents the dynamic weighting coefficient. true This represents the actual eyeball diameter in the sample test set. Therefore, a 50% increase in the dynamic weighting coefficient for actual eyeball diameters ≤7mm compensates for the learning difficulty of small targets, while a 20% decrease in the dynamic weighting coefficient for diameters ≥9mm prevents large targets from dominating training.

[0075] Specifically, the process of adjusting the focusing parameters is as follows:

[0076]

[0077] In the formula, γ dynamic E represents the focusing parameter. lightThis indicates the ambient light intensity; therefore, by increasing γ under strong light... dynamic Increase attention to small pupils and reduce gamma rays in dark environments. dynamic Avoid over-suppressing easy samples.

[0078] Understandably, the conventional Focal Loss function is insufficient for focusing on small targets of 5 to 8 mm in the eye. Therefore, dynamic weighting coefficients are used to compensate for the learning difficulty of small targets, and focusing parameters are used to improve classification accuracy in strong light and reduce classification accuracy in dark environments to avoid over-suppression leading to classification errors.

[0079] Furthermore, the process of constructing the eyeball boundary constraint term includes:

[0080] Obtain the standard eyeball diameter corresponding to neutral light in the sample test set, and the user's historical blink frequency corresponding to the blinking time sequence information;

[0081] The user adaptation delay time is calculated based on the user's historical blink frequency.

[0082] Calculate the ratio of the standard eye diameter to the actual eye diameter of the sample test set, and calculate the eye dilation adjustment amount with the user adaptation delay time as the exponent and the ratio as the base;

[0083] The eyeball boundary constraint term is constructed based on the product of the eyeball dilation adjustment amount and the exponential function.

[0084] Specifically, the eyeball boundary constraint term is:

[0085]

[0086] In the formula, L bound (d pred ) represents the eyeball boundary constraint term, d true d represents the true eye diameter of the sample test set. st The standard eyeball diameter is obtained by taking the standard eyeball diameter corresponding to neutral light in the sample test set, for example, 9. μ represents the user adaptation delay time. Let be the exponential function, where d pred This indicates a predicted eyeball diameter.

[0087] Specifically, the process of calculating the user adaptation delay duration based on the user's historical blink frequency is as follows:

[0088] μ = μ base +k·(HBR-BR norm )

[0089] In the formula, μ represents the user adaptation delay time, μ baseThe physiological baseline value is set to 1.5, and k represents the adjustment coefficient, preferably an empirical value of 0.02. BR norm The medical baseline blink rate is preferably 15 blinks per minute, while the HBR represents the user's historical blink rate, which is the number of blinks in the most recent minute. Preferably, the user's historical blink rate HBR is recalculated every 30 days to avoid habit drift.

[0090] Therefore, a stronger penalty is implemented for predicted values ​​exceeding the standard diameter of the eyeball through the eyeball boundary constraint term, and through... Adjusting the sensitivity of the eyeball boundary constraint term to the degree of deviation of the predicted value, when the user's historical blinking frequency is high, it indicates that eye discomfort is very likely to occur. Increasing the sensitivity will increase the model's recognition accuracy, improve the accuracy and reliability of the prediction, and make it more consistent with the normal medical condition of the eyeball.

[0091] Furthermore, the process of constructing the binocular constraint term includes:

[0092] The difference between the predicted difference in binocular diameter and the standard deviation of both eyes is used to construct the binocular constraint term through the ReLU activation function value.

[0093] Specifically, the binocular constraint term is:

[0094] L sym (d predleft ,d predright ) = ReLU(|d predleft -d predright -0.5)

[0095] In the formula, L sym (d predleft ,d predright ) represents the binocular constraint term, d predleft d predright represents the predicted diameter of the left eye and the predicted diameter of the right eye, respectively; 0.5 represents the binocular standard deviation; and ReLU represents the ReLU activation function. Therefore, the binocular constraint term penalizes predicted values ​​exceeding the normal binocular standard deviation of 0.5 mm, ensuring the prediction results conform to normal binocular conditions.

[0096] Specifically, the training process of the eye recognition model involves using data collected from the official repository (git) and the visual sensor and ambient light sensor of the LCD monitor to jointly construct a YOLO standard training and test set labeled with eye regions. A pre-trained YOLOv8n model with weights is loaded, and the training parameters are set as follows: training epochs = 100, input resolution imgsz = 640, batch size = 16, GPUID using dual cards device = [0,1], data loading threads workers = 2, automatic data augmentation (Mosaic / MixUp) augment = True, confidence threshold conf = 0.25, and IoU threshold iou = 0.6. The built-in cosine annealing strategy of the YOLOv8n model, which does not require manual settings, is used. The YOLOv8n model is trained and tested based on the comprehensive loss function until the value of the comprehensive loss function meets the confidence threshold and the IoU meets the IoU threshold, thus generating the eye recognition model.

[0097] In the above scheme, the weighted Focal loss term dynamically adjusts the focusing parameters by ambient light intensity to solve the false detection problem caused by sudden changes in illumination. It adaptively adjusts the weights according to the actual diameter of the eyeball to improve the model's ability to generalize to physiological differences. The eyeball boundary constraint term introduces an exponential function to strengthen the penalty for the deviation between the predicted eyeball diameter and the standard diameter, ensuring the geometric consistency of pupil dilation changes. It also avoids over-adjustment by combining the user adaptation delay time. The binocular constraint term forces binocular collaborative detection to reduce the false negative rate.

[0098] like Figure 3 As shown, further, the process of determining the user's eyeball change area and blink frequency value based on the eyeball annotation image includes:

[0099] The binary mask of the eye region is determined based on the eyeball annotation image, and the least squares method is used to fit an ellipse to the binary mask to calculate the eyeball diameter.

[0100] The change area of ​​the eyeball is calculated using the sensor size conversion formula;

[0101] If the current change area of ​​the eyeball is less than the average change area of ​​the eyeball in the previous set number of frames, it is determined to be a blinking event;

[0102] The blink frequency value is determined based on the number of blink events of a set duration.

[0103] Specifically, the process of calculating the diameter of the eyeball is as follows:

[0104]

[0105] In the formula, x c ,y cLet represent the coordinates of the ellipse center, a and b represent the eye diameter (including the major and minor axes of the ellipse), θ represent the rotation angle of the ellipse, and x and y represent the coordinates of the eye region belonging to the binary mask M. Therefore, the above formula can be used to calculate and determine a and b as the eye diameter.

[0106] Specifically, the sensor size conversion formula is as follows:

[0107]

[0108] In the formula, A t d represents the area of ​​change in the eyeball. pixel The average diameter of the eyeball is represented by s, which is the average of the major radius *a* and the minor axis radius *b* of the ellipse. sensor This indicates the size of the vision sensor, for example, a camera sensor is 50mm. 2 f represents the focal length of the vision sensor (in mm), r image Indicates image resolution.

[0109] It is understandable that the process of determining the average eyeball diameter described above is also used for labeling the actual eyeball diameter of the sample test set, as well as the predicted eyeball diameter output by the model.

[0110] Specifically, a blinking event is defined as an event where the area of ​​change of the eyeball is less than the average area of ​​change of the eyeball in the previous 30 frames.

[0111] like Figure 4 As shown, the process of calculating the desired brightness value using the ambient light intensity, the change area of ​​the eyeball, and the blink frequency value through a brightness conversion formula further includes:

[0112] Calculate the logarithm of the ambient light intensity plus the light intensity offset to determine the basic expected brightness value;

[0113] Based on the change in eyeball area and the blink frequency value, determine whether the eyeball is fatigued; if so, adjust the zoom sensitivity.

[0114] Calculate the ratio of the standard area of ​​the eyeball to the change area of ​​the eyeball, and calculate the brightness response gain value with the ratio of the change area of ​​the eyeball as the base and the scaling sensitivity as the exponent.

[0115] The sum of the base expected brightness value and the brightness response gain value is taken as the expected brightness value.

[0116] Specifically, the desired brightness value is:

[0117]

[0118] In the formula, B exp Represents the desired brightness value, α·ln(E)env +β) represents the basic expected brightness value, α is a dynamic coefficient that can be adjusted according to user habits, E env β represents the ambient light intensity, and β represents the light intensity offset. The preferred base light compensation value is 10 lux. This represents the luminance response gain value, where A st A represents the standard area of ​​the eyeball. t The variable area of ​​the eyeball is represented by γ, which represents the eye response gain, preferably 0.1-0.3, to suppress brightness oscillations caused by overcompensation. δ represents scaling sensitivity. It can be understood that the eye response gain γ (0.1-0.3) can suppress brightness oscillations caused by overcompensation, and the scaling sensitivity δ is used to match the desired brightness value with the Weber-Fechner law of human brightness perception. The Weber-Fechner law states that the brightness intensity perceived by the human eye is proportional to the logarithm of the stimulus intensity, and the basic desired brightness value (representing the human eye's perception of brightness based on a certain ambient light background) embodies this logarithmic relationship. That is, the setting of the basic desired brightness value is based on the logarithmic response characteristics of the human eye to ambient light intensity. The brightness response gain value... This reflects the nonlinear (logarithmically correlated) response mechanism of the human eye to relative changes in perceived brightness, as described in the Weber-Fechner law. By using the eye response gain γ and scaling sensitivity δ, the expected brightness value calculated by the entire formula conforms to the nonlinear, relatively sensitive, and logarithmically similar characteristics of the human eye when actually perceiving brightness.

[0119] Specifically, the fatigue state is triggered when the change in eyeball area and the blinking frequency value exceed a baseline value by 30% for 5 seconds. The specific calculation process for the scaling sensitivity δ is as follows:

[0120]

[0121] like Figure 4 As shown, further, the process of self-adjusting the brightness of the liquid crystal display based on the desired brightness value includes:

[0122] The dynamic brightness factor is calculated by summing the Sigmoid function values ​​of the desired brightness value and the current brightness value with the user's habitual terms, so that the dynamic brightness factor forms an S-shaped transition.

[0123] The display brightness value is determined by multiplying the dynamic brightness factor, the user-set brightness value, and the current display mode fixed factor, and the brightness of the liquid crystal display is adjusted to the display brightness value.

[0124] Specifically, the value of the Sigmoid function is:

[0125]

[0126] In the formula, B(t) represents the Sigmoid function value, that is, the ambient light intensity adjustment value at time t. start Indicates the current monitor brightness, B exp Let t0 represent the desired brightness value, t0 represent the center switching time point of the S-shaped transition, t represent the current time, and k represent the steepness of the curve, preferably 0.8. Therefore, by using an S-shaped curve (Sigmoid) gradient, the step-like feel of linear adjustment can be eliminated.

[0127] Specifically, the calculation process for the dynamic luminance factor is DBF = B(t) + h(U). p U h ), where DBF represents the dynamic brightness factor, B(t) represents the Sigmoid function value, and h(U p U h () indicates user habits.

[0128] Specifically, the calculation process for the display brightness value is as follows: B actua =B set ×F fixed ×DBF, where B actua Indicates the display brightness value, B set This indicates the user-defined brightness value, F. fixed This indicates the current display mode's fixed factor, such as the eye protection mode F. fixed The value is 0.8, where DBF represents the dynamic luminance factor.

[0129] Furthermore, the user habit item is determined based on the average historical brightness preference of users with time decay weight.

[0130] Specifically, the calculation process for user habit items is as follows:

[0131]

[0132] In the formula, h(U) p U h ) indicates user habits item, U p Indicates the current screen brightness value, U h This represents the user's historical brightness preference value, σ represents the Sigmoid activation function, and w i The time decay weight at time point i, U is used to calculate the mean of user historical brightness preferences based on time decay weights for time period n. i Let i be the user's historical brightness preference value at time point i, where the time decay weight is... t now t represents the current time point. i Let i represent a historical time point, such that the further the historical time point is from the current time point, the smaller the time decay weight.

[0133] Furthermore, the MonitorControl tool is used to adjust the display brightness to the specified display brightness value via the DDC / CI protocol.

[0134] This embodiment also provides a self-adjusting liquid crystal display based on multi-source data analysis, which is equipped with a visual sensor for data acquisition and an ambient light sensor, and is equipped with an adjustment method for the self-adjusting liquid crystal display based on multi-source data analysis.

[0135] The above scheme achieves dynamic calibration of basic brightness. The basic expected brightness is calculated by adding the logarithm of the ambient light intensity and the light intensity offset. This conforms to the human eye's perception characteristics of light intensity according to the Weber-Fechner law, avoids the problem of excessive brightness / darkness caused by linear adjustment, increases sensitivity under high fatigue conditions, and rapidly improves the brightness response gain value to relieve eye fatigue in a timely manner.

[0136] In this embodiment, by constructing a comprehensive loss function, false detections caused by strong or weak light interference are suppressed, ensuring geometric consistency between the predicted eye diameter and the true value, avoiding single-eye missed detections or deviations, and significantly improving the eye localization accuracy of the YOLOv8 model under complex lighting conditions. This enables a self-adjusting liquid crystal display based on ambient light and vision, combining multi-source data such as user eye state and ambient light intensity to achieve precise brightness adjustment, thereby improving user experience and reducing visual fatigue. The weighted Focal loss term dynamically adjusts the focusing parameters based on ambient light intensity to solve the false detection problem caused by sudden changes in lighting. It adaptively adjusts the weights according to the true eye diameter, improving the model's generalization ability to physiological differences. The eye boundary constraint term introduces an exponential function to strengthen the penalty for deviation between the predicted eye diameter and the standard diameter, ensuring geometric consistency of pupil dilation changes. It also avoids over-adjustment by incorporating user adaptation delay time. The binocular constraint term forces binocular collaborative detection, reducing the false detection rate. It achieves dynamic calibration of basic brightness, calculates the basic expected brightness by adding the logarithm of ambient light intensity and light intensity offset, conforms to the Weber-Fechner law of human eye perception of light intensity, avoids the problem of excessive brightness / darkness caused by linear adjustment, increases sensitivity under high fatigue state, and rapidly improves the brightness response gain value to relieve eye fatigue in time.

[0137] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0138] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A self-adjusting liquid crystal display adjustment method based on multi-source data analysis, characterized in that, include: The system acquires user images from the visual sensor of a self-adjusting liquid crystal display and ambient light intensity from the ambient light sensor. A sample test set is constructed, including the correspondence between user images and ambient light intensity. An eye recognition model based on the YOLOv8 framework is built. A comprehensive loss function with a weighted Focal loss term, an eye boundary constraint term, and a binocular constraint term is constructed. The eye recognition model is trained using the comprehensive loss function and the sample test set. The weighted Focal loss term is determined based on the ambient light intensity and the true diameter of the eyeball. The eye boundary constraint term is determined based on the true eyeball diameter and the predicted eyeball diameter. The binocular constraint term is determined based on the interocular distance. The predicted eyeball diameter and the interocular distance are determined by calculating the eye recognition area. The current user image is used to generate a user eyeball annotation image through the eyeball recognition model, and the user's eyeball change area and blink frequency value are determined based on the eyeball annotation image; The ambient light intensity, the change area of ​​the eyeball, and the blink frequency value are used to calculate the desired brightness value using a brightness conversion formula. Based on the desired brightness value, the brightness self-adjustment control of the liquid crystal display is performed.

2. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to claim 1, characterized in that, The process of constructing the comprehensive loss function includes: The weighted Focal loss term is constructed based on the product of the logarithmic function of the positive confidence of the eyeballs in the sample test set, the negative confidence of the eyeballs with the focusing parameter as the exponent, and the dynamic weighting coefficients. Calculate the exponential function of the difference between the predicted eye diameter and the standard eye diameter, and construct the eye boundary constraint term based on the product of the exponential function and the eye expansion adjustment amount; The binocular constraint term is constructed based on the activation function values ​​of the predicted difference in binocular diameter and the standard deviation of binoculars; The comprehensive loss function is constructed by weighted summation of the weighted Focal loss term, the eyeball boundary constraint term, and the binocular constraint term.

3. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to claim 2, characterized in that, The process of constructing the weighted Focal loss term includes: The dynamic weighting coefficients are adjusted based on a comparison between the actual eye diameter of the sample test set used to train the eye recognition model and the normal upper limit threshold and normal lower limit threshold of the eye diameter. The focusing parameters are adjusted based on a comparison of the ambient light intensity with the strong light threshold and the weak light threshold.

4. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to claim 2, characterized in that, The process of constructing the eyeball boundary constraint term includes: Obtain the standard eyeball diameter corresponding to neutral light in the sample test set, and the user's historical blink frequency corresponding to the blinking time sequence information; The user adaptation delay time is calculated based on the user's historical blink frequency. Calculate the ratio of the standard eye diameter to the actual eye diameter in the sample test set, and calculate the eye dilation adjustment amount with the user adaptation delay time as the exponent and the ratio as the base.

5. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to claim 2, characterized in that, The process of constructing the binocular constraint terms includes: The difference between the predicted difference in binocular diameter and the standard deviation of both eyes is used to construct the binocular constraint term through the ReLU activation function value.

6. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to claim 1, characterized in that, The process of determining the user's eyeball change area and blink frequency values ​​based on the eyeball annotation image includes: The binary mask of the eye region is determined based on the eyeball annotation image, and the least squares method is used to fit an ellipse to the binary mask to calculate the eyeball diameter. The change area of ​​the eyeball is calculated using the sensor size conversion formula; If the current change area of ​​the eyeball is less than the average change area of ​​the eyeball in the previous set number of frames, it is determined to be a blinking event; The blink frequency value is determined based on the number of blink events of a set duration.

7. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to claim 1, characterized in that, The process of calculating the desired brightness value using the ambient light intensity, the change area of ​​the eyeball, and the blink frequency value through a brightness conversion formula includes: Calculate the logarithm of the ambient light intensity plus the light intensity offset to determine the basic expected brightness value; Based on the change in eyeball area and the blink frequency value, determine whether the eyeball is fatigued; if so, adjust the zoom sensitivity. Calculate the ratio of the standard area of ​​the eyeball to the change area of ​​the eyeball, and calculate the brightness response gain value with the ratio of the change area of ​​the eyeball as the base and the scaling sensitivity as the exponent. The sum of the base expected brightness value and the brightness response gain value is taken as the expected brightness value.

8. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to any one of claims 1 to 7, characterized in that, The process of self-adjusting the brightness of the liquid crystal display based on the desired brightness value includes: The dynamic brightness factor is calculated by summing the Sigmoid function values ​​of the desired brightness value and the current brightness value with the user's habitual terms, so that the dynamic brightness factor forms an S-shaped transition. The display brightness value is determined by multiplying the dynamic brightness factor, the user-set brightness value, and the current display mode fixed factor, and the brightness of the liquid crystal display is adjusted to the display brightness value.

9. The adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis according to claim 8, characterized in that, The user habit item is determined based on the average historical brightness preference of users with time decay weight.

10. A self-adjusting liquid crystal display based on multi-source data analysis, characterized in that, The device includes a visual sensor for data acquisition and an ambient light sensor, and is equipped with an adjustment method for a self-adjusting liquid crystal display based on multi-source data analysis as described in any one of claims 1 to 9.

Images