Large-screen display system based on machine learning

Abstract

The invention belongs to the technical field of screen display, and particularly relates to a large-screen display system based on machine learning, which comprises an environment sensing module, a user sensing module, a calculation module, a compensation module and a display control module, the compensation module calculates the distance and the visual angle according to the positioning coordinate parameters, and when the distance and the visual angle exceed a preset distance threshold value and a preset visual angle threshold value, the initial brightness adjustment coefficient is compensated according to the distance, the visual angle and the pupil diameter parameters to generate a corrected brightness adjustment coefficient; according to the correction brightness adjustment coefficient, the distance between the viewer and the large screen and the viewing angle are considered, so that the large-screen display system can dynamically adjust the screen brightness more accurately according to the visual comfort level of the viewer and the influence of ambient light.

Description

Technical Field

[0001] The present invention belongs to the field of screen display technology, and in particular relates to a large-screen display system based on machine learning. Background Art

[0002] Large screens are used in transportation, smart manufacturing, energy management, production and construction, and other fields to monitor and display key data in real time. For example, in transportation, large screens can be used to monitor traffic flow and traffic conditions in real time, allowing for timely adjustments to traffic signals and improving road efficiency. In specific applications, they provide a wide field of view, allowing viewers standing in different positions to clearly see the screen's display content and easily grasp the information displayed on the large screen. In large-screen display systems, the viewer's position determines the angle at which they observe the screen (i.e., the viewing angle). According to optical principles, when light is reflected from the screen surface to the human eye, it is received by the human eye. The effective viewing angle of a large screen refers to the maximum angle range within which the viewer can clearly see the image. Beyond this range, the image may become blurred or difficult to recognize due to insufficient brightness or interference from reflections.

[0003] Even if the screen's display brightness remains consistent, the actual brightness perceived by the viewer will vary depending on their position. When the viewer is positioned to the side or at an angle to the screen, ambient light may reflect off the screen surface and enter the eye, causing glare or reflections. This reflection reduces the contrast of the image and affects the visual experience.

[0004] Therefore, it is necessary to adjust the brightness of the large screen in a targeted manner according to the changes in distance and viewing angle caused by the changes in the viewer's position. Summary of the Invention

[0005] In order to solve the above-mentioned problems existing in the prior art, the present invention provides a large-screen display system based on machine learning to solve the problem of blurred images caused by changes in distance and viewing angle caused by the viewer's standing position.

[0006] The purpose of the present invention can be achieved through the following technical solution: A large-screen display system based on machine learning, characterized in that: It includes an environment perception module, a user perception module, a calculation module, a compensation module and a display control module; The environment perception module is used to obtain environmental parameters, which include ambient light intensity; The user perception module is used to obtain the viewer's pupil diameter parameter and positioning coordinate parameter; The user perception module and the user perception module respectively transmit the acquired data to the calculation module, and the calculation module calculates the initial brightness adjustment coefficient according to the environmental parameters; The compensation module calculates the distance and viewing angle according to the positioning coordinate parameters, and when the distance and viewing angle exceed the preset distance threshold and viewing angle threshold, the compensation module compensates the initial brightness adjustment coefficient according to the distance, viewing angle and pupil diameter parameters to generate a modified brightness adjustment coefficient; The display control module dynamically adjusts the screen brightness according to the modified brightness adjustment coefficient.

[0007] Preferably, the compensation module compensates the initial brightness parameter including the following steps: Predict the viewer’s current visual comfort level V using a pre-trained machine learning model based on pupil diameter parameters; Based on the positioning coordinate parameters, the viewer's viewing angle is calculated and the ambient light impact factor R is evaluated; Compensating the initial brightness adjustment coefficient based on the viewer's current visual comfort level and viewing angle to generate a modified brightness adjustment coefficient; Here, compensating the initial brightness adjustment coefficient includes the following steps: Ccorrected = Cinitial ; Here, Ccorrect is the corrected brightness adjustment coefficient, Cinitial is the initial brightness adjustment coefficient, V is the visual quantization, and R is the ambient light intensity influence factor. is the weight coefficient.

[0008] Preferably, the compensation module includes a model training submodule, and the model training submodule includes a data collection unit, a data preprocessing unit, a data feature unit and a data training unit; The data collection unit is used to obtain historical pupil diameter parameters and ambient light intensity parameters; The data preprocessing unit is used to perform a primary filtering on the historical pupil diameter parameter and the ambient light intensity parameter using the machine learning model using a wavelet transform method, and then perform a secondary filtering on the data after the primary filtering using a Kalman filter, and then perform noise reduction on the data after the secondary filtering, and use mean filling and KNN interpolation to supplement the missing data, and then use a linear transformation method to normalize the supplemented data; The data feature unit is used to determine the original features, use the principal component analysis method to reduce the dimension of the original features, and then classify and encode the original features into numerical features; The data training unit is used to train the neural network model to obtain a preset machine learning model, and the neural network model includes one of a convolutional neural network, a recurrent neural network and a graph neural network.

[0009] Preferably, the model training submodule also includes a data evaluation unit, which is used to evaluate the preset machine learning model using test data, and the evaluation indicators include at least one of mean absolute variance, mean square error, root mean square error and R variance.

[0010] Preferably, the test data is iteratively generated using a generative adversarial network, and the training set of the generative adversarial network is historical pupil diameter parameters and ambient light intensity parameters.

[0011] Preferably, it further includes a display feedback adjustment module, which is used to receive the feedback result of the current user after the screen brightness is adjusted.

[0012] Preferably, the model training submodule further includes an optimization unit, which is used to perform parameter optimization and structural optimization on a preset machine learning model according to the evaluation results and feedback results.

[0013] Preferably, it also includes a display recognition module, which is electrically connected to the display control module. The display recognition module transmits the recognition content to the display control module. The display content includes: dynamic and non-dynamic. The display control module compensates the brightness adjustment coefficient according to the display content.

[0014] Preferably, the compensation is that when it is determined that the screen content is in a dynamic state, the display control module increases the display brightness of the display screen.

[0015] The beneficial effects of the present invention are: The present invention uses the compensation module in this embodiment to compensate the initial brightness coefficient of the original large screen based on the ambient brightness data and the set parameter data based on the changes in the user's pupil diameter parameter and the station coordinates, to obtain a corrected brightness adjustment coefficient. The corrected brightness adjustment coefficient takes into account the viewer's distance from the large screen and the viewing angle, allowing the large-screen display system to more accurately adjust the screen brightness based on the viewer's visual comfort level and the influence of ambient light, thereby effectively reducing the viewer's visual fatigue and improving the viewing experience. At the same time, the machine learning-based method enables the system to continuously learn and optimize, adapting to different viewing scenarios and user needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0017] Figure 1 4 is a system block diagram of the display system of the present invention. DETAILED DESCRIPTION

[0018] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0019] See also Figure 1 , this embodiment provides a large-screen display system based on machine learning, including an environment perception module, a user perception module, a calculation module, a compensation module and a display control module.

[0020] Among them, the environmental perception module is used to obtain environmental parameters in real time. The environmental parameters include ambient light intensity, ambient color temperature, and the change rate of the two. The environmental perception module can use light sensors such as photoresistors or photodiodes to measure the ambient light intensity, and convert the light intensity signal into an electrical signal, and then calculate the light intensity value based on the electrical signal. An RGB color sensor or a dedicated color temperature sensor (such as Taos TCS34725) is used to measure the color composition of the ambient light, thereby deriving the color temperature information. The obtained light intensity value and color temperature information are subjected to time series analysis, and the differential algorithm is used to obtain the light intensity value-time curve and the color temperature-time curve, and the change rate of the two is calculated respectively.

[0021] User perception module, used to obtain the viewer's pupil diameter parameters, positioning coordinate parameters, and viewing time parameters in real time: Among them, the combination of infrared cameras, infrared sensors, eye tracking and other equipment combined with computer vision technology can capture the user's pupil diameter parameters in real time. Edge detection technology can also be used to extract the pupil area from the collected video stream and further measure its diameter size, and use positioning sensors to detect the position coordinate parameters of multiple viewers.

[0022] The viewing time parameter here uses the duration of time the viewer's eyes remain fixed on the screen as the starting point, and calculates the longest viewing time among all viewers. In the aforementioned description, the viewer's location coordinates are obtained to first locate the viewer, and then the pupil diameter parameters of the viewer at different positions are obtained. Because the detection equipment that detects the viewer's location coordinates and pupil diameter parameters carries a timestamp when recording the test results, the location coordinate parameters here can be used with a general detection equipment that can locate the viewer's position.

[0023] Therefore, on this basis, Shapelet extraction methods, downsampling techniques, or other corresponding extraction techniques can be used to extract a continuous sequence of a specified time length from the infrared recognition sensor's detection parameters. This sequence is then used as the cumulative viewing time of the large screen at the current moment. After the continuous time sequence ends, the previously accumulated viewing time is reset and recalculated. Because when a user's eyes move away from the electronic screen and look at objects farther away, their eye focus adjusts, thus alleviating eye fatigue. Therefore, only the data of the viewer with the longest viewing time needs to be calculated. That is, after the user perception module detects the viewer's parameters, the subsequent detection data for each viewer is assigned a unique label, enabling personalized understanding of multiple viewers. Based on the labels and the accumulated viewing time, the calculation module presets a viewing time threshold and can associate the viewing time threshold with the number of viewers who have reached the threshold using a piecewise function. When the calculation module detects that the viewing time of the viewer with the longest viewing time exceeds the viewing time threshold, it reduces the screen refresh rate when displaying static content to reduce screen flicker and increases the screen refresh rate when displaying dynamic content to achieve a smoother image.

[0024] The user perception module and the user perception module respectively transmit the acquired data to the calculation module. The calculation module calculates the initial brightness adjustment coefficient according to the environmental parameters. The calculation module first adjusts the display brightness of the large screen according to the ambient light intensity and color temperature parameters provided by the environmental perception module, combined with a predefined ambient light response model (such as a linear or nonlinear function). Here, the calculation module adjusts the display brightness of the large screen according to the parameters obtained by the perception module, and uses a general large screen to perform environmental adaptive brightness adjustment technology according to the environmental parameters and its own set operating parameters.

[0025] The compensation module calculates the distance and viewing angle based on the positioning coordinate parameters. When the distance and viewing angle exceed the preset distance threshold and viewing angle threshold, the compensation module compensates the initial brightness adjustment coefficient based on the distance, viewing angle, and pupil diameter parameters to generate a corrected brightness adjustment coefficient. The compensation based on the initial brightness adjustment coefficient includes the following steps: S1: Predict the viewer's current visual comfort level V based on pupil diameter parameters using a pre-trained machine learning model, including the following sub-steps: S11: Machine Learning Model Selection and Training: Select a neural network model. During the training phase, collect a large amount of pupil diameter parameters and corresponding viewer visual comfort evaluation data as a training set. Visual comfort evaluation can be collected through subjective questionnaires or objective methods such as eye tracking. S12: Model training process: The neural network model is trained using the pupil diameter parameters in the training set as input and the visual comfort level as output. By continuously adjusting the model's weights and biases, the model is able to accurately predict the visual comfort level based on the pupil diameter parameters. S13: Predicting visual comfort level: Inputting the currently acquired viewer pupil diameter parameter into a pre-trained machine learning model, and the model outputs the viewer's current visual comfort level.

[0026] S2: Calculate the viewer's viewing angle based on the positioning coordinate parameters and evaluate the ambient light impact factor R, including the following sub-steps: S21: Calculate viewing angle: Based on the positioning coordinate parameters, use the geometric calculation method to calculate the viewer's viewing angle relative to the large screen. The center point coordinates of the large screen can be used as the origin, and the viewer's positioning coordinates can be , the viewing angle can be obtained through vector calculation and trigonometric function relationship; S22: Evaluate the ambient light impact factor R: Evaluate the degree to which ambient light affects the viewer's vision based on the viewing angle and ambient light intensity. For example, when the viewer is positioned to the side of a large screen and the ambient light is strong, the reflected ambient light may have a significant impact on vision, resulting in a larger R value. Conversely, when the viewing angle is straight on and the ambient light is dim, the R value is smaller. A large amount of experimental data can be obtained through eyeball simulation experiments to establish a mapping relationship between viewing angle, ambient light intensity, and R value.

[0027] S3: compensating the initial brightness adjustment coefficient based on the viewer's current visual comfort level and viewing angle to generate a modified brightness adjustment coefficient; Here, the corrected brightness adjustment coefficient is calculated using the following formula in combination with the initial brightness adjustment coefficient: Ccorrected = Cinitial ; Here, Ccorrect is the corrected brightness adjustment coefficient, Cinitial is the initial brightness adjustment coefficient, V is the visual quantization, and R is the ambient light intensity influence factor. is the weight coefficient.

[0028] The compensation module sends the calculated corrected brightness adjustment coefficient Ccorrection to the display control module, and the display control module dynamically adjusts the screen brightness according to the coefficient to improve the viewer's visual experience.

[0029] The compensation module in this embodiment compensates the initial brightness coefficient of the original large screen based on ambient brightness data and set parameter data to obtain a correction coefficient. This correction coefficient takes into account the changes in light perception caused by the viewer's perspective. This allows the large-screen display system to more accurately dynamically adjust the screen brightness based on the viewer's visual comfort level and the influence of ambient light, effectively reducing viewer visual fatigue and improving the viewing experience. At the same time, the machine learning-based approach enables the system to continuously learn and optimize, adapting to different viewing scenarios and user needs.

[0030] Large screens display different content in different usage scenarios, including static and dynamic content, such as images and text, and videos. Furthermore, the human eye has varying sensitivities to different types of visual information. When faced with complex visual stimuli (such as fast-moving motion in a video), the human eye tends to focus on overall dynamic changes rather than local details. Therefore, in such situations, appropriately high brightness helps enhance contrast and make the image more vivid. Conversely, reading static text requires higher levels of legibility, and appropriately lowering background brightness can reduce glare and thus alleviate eye strain.

[0031] Therefore, in order to optimize the user's cognitive process when the user is faced with different contents displayed on the display screen, in one embodiment, it also includes a display recognition module. The display recognition module and the display control module are electrically connected. The display recognition module transmits the recognition content to the display control module for identifying the display content of the display screen. The display content includes: dynamic and non-dynamic, that is, whether the display content of the current electronic screen is dynamically displayed or statically displayed. The display control module compensates for the brightness adjustment coefficient when the content transmitted by the display recognition module is a dynamic signal. The compensation includes correspondingly increasing the display brightness, including when the display recognition module recognizes that the display content of the electronic screen is dynamic, the display control module increases the screen brightness by a certain proportion. The proportion can be obtained through corresponding experimental research. The display control module displays the display brightness of the adjusted screen on the display screen.

[0032] By classifying the screen's real-time display content into dynamic and non-dynamic (static) content, the screen's display brightness can be customized. This ensures that when playing video, the screen provides sufficient brightness to allow users to easily track multiple visual focal points. In text reading mode, unnecessary visual interference is avoided, and the original brightness adjustment logic is maintained. This allows for differentiated control of display brightness based on the screen's display content.

[0033] In one embodiment, the compensation module includes a model training submodule, and the model training submodule includes a data collection unit, a data preprocessing unit, a data feature unit, a data evaluation unit, an optimization unit, and a data training unit; The data collection unit is used to obtain historical pupil diameter parameters and historical ambient light intensity parameters; The data preprocessing unit is used to perform the following processing on the historical pupil diameter parameter and the historical ambient light intensity parameter using a machine learning model: Initial filtering: Use wavelet transform to perform initial filtering on the raw data. Wavelet transform can effectively remove high-frequency noise and retain the main features of the data. For example, for the instantaneous fluctuations in the ambient light intensity data, wavelet transform can filter out these noises and obtain a relatively smooth ambient light intensity change curve. Secondary filtering: The data after primary filtering is filtered again using a Kalman filter, extended Kalman filter (EKF), or unscented Kalman filter (UKF). The UKF excels at handling nonlinear systems and can better estimate the true state of the data. For example, pupil diameter, as its changes are affected by multiple factors and exhibit nonlinear characteristics, the UKF can more accurately track the changing trends of pupil diameter, further improving data accuracy. Noise reduction and missing data processing: To remove residual minor noise, the data after secondary filtering needs to be denoised. Missing data is also supplemented using mean filling and KNN interpolation. For example, if pupil diameter data at a certain moment is missing, the KNN algorithm will calculate a reasonable estimate based on multiple data points at adjacent moments to fill the missing value. Normalization: The completed data is normalized using a linear transformation. All data is mapped to the [0, 1] interval, ensuring that data with different features are on the same scale, facilitating subsequent model training and analysis. For example, ambient light intensity data may vary over a wide range. After normalization, it is processed on the same scale as data such as pupil diameter parameters, improving the stability and accuracy of model training. Data feature unit, used for the following functions: Determine the original features, including ambient light intensity, pupil diameter parameters, and viewer position coordinate parameters. These original features are directly extracted from the collected data and are the basis for subsequent analysis and processing; Dimensionality reduction: Principal component analysis (PCA) is used to reduce the dimensionality of the original features. PCA can convert high-dimensional data into low-dimensional data while preserving the data's key information. For example, when processing a large amount of viewer location information, PCA can extract several key principal components, reducing the data's dimensionality and computational complexity without losing too much information. Feature conversion: categorizes and encodes raw features into numerical features. For example, the viewer's location information can be classified into categories such as front row, middle row, and back row. This information is then converted into numerical features using methods such as one-hot encoding to facilitate processing by the neural network model.

[0034] Data training unit: For model selection, we chose to build a hybrid neural network model, combining the strengths of convolutional neural networks (CNNs), recurrent neural networks (RNNs), and graph neural networks (GNNs). CNNs excel at processing image and spatial data, and are particularly effective at extracting spatial distribution features from ambient light intensity data. RNNs can process time series data, and are particularly advantageous for analyzing trends in pupil diameter over time. GNNs are suitable for processing data with complex relationships, such as the positional relationships between viewers. Model training involves training the hybrid neural network model using preprocessed and feature-engineered data. During training, appropriate hyperparameters, such as the learning rate and number of iterations, are set. By continuously adjusting the model's parameters, the model accurately learns the relationship between factors like ambient light intensity and pupil diameter and visual comfort, ultimately achieving a pre-defined machine learning model.

[0035] In one embodiment, the data evaluation unit is configured to: Evaluation metrics, such as mean squared error (MSE) and accuracy, are used to measure model performance. MSE reflects the error between the model's predicted value and the true value, while accuracy assesses the model's accuracy in classification tasks. Evaluation: Use the validation set to evaluate the trained model. Input the validation set data into the model and calculate metrics such as the mean square error (MSE) and accuracy between the model's predictions and the true values. Based on the evaluation results, determine whether the model meets the expected performance requirements. If the model performs poorly, the evaluation results are fed back to the optimization unit for further optimization.

[0036] To maximize the use of existing data and improve the model's ability to learn realistic distributions, the evaluation unit uses a generative adversarial network (GAN) to iteratively learn historical environmental parameters, pupil diameter parameters, and positioning coordinate parameters. This test data simulates the distribution characteristics of historical data while introducing a certain degree of diversity or perturbation. This generated test dataset is used to validate the main model and evaluate its performance, such as accuracy and loss. Because the test data is generated by a GAN, it contains samples that are similar to but slightly different from the historical data, while also introducing new sample variants.

[0037] Because measuring pupil changes requires the test subject to move in front of the screen at different angles or distances, the process is cumbersome. Prior art eye biosimulation technologies exist that, using experimental data and models, can biologically display specific human eye movements or reactions. For example, publication number CN102698442B describes a medical manikin eye, in which a drive motor and a support plate for the blinking transmission mechanism and pupil change mechanism are fixed to a base. The blinking drive motor transmits power to the eyelids via a crankshaft transmission mechanism, causing blinking. The pupil drive motor drives the pupil change mechanism installed inside the eyeball to achieve pupil change. A pupil size and position determination mechanism is connected to the pupil change mechanism, which also houses a pupil light-sensing mechanism. The eyeball is connected to the base via a support plate. This invention integrates the blinking and pupil change functions into a single integrated system, resulting in a simple and compact structure and reliable transmission. By adding a control system to control the motor rotation, it can simulate both blinking at different frequencies and pupil size changes due to internal and external factors. The above technology is used to study the correspondence between the display brightness of a large screen and the changes in the viewer's pupils. The above technology is used to study the changing relationship between the changes in the pupils of the human eyeball and the intensity of external light (that is, the display brightness of the large screen here). The display brightness of the large screen is adjusted according to the changes in the pupil diameter of the eyeball, and experimental statistical calculations are performed to obtain the data collected by the data collection unit.

[0038] In one embodiment, when using a preset deep learning model to train a large amount of sample data based on pupil diameter data and large-screen display brightness, the validation step requires calculating the model's diagnostic accuracy, recall rate, and other metrics on the validation set. Therefore, the model training submodule also includes an optimization unit, which is responsible for optimizing the parameters and structure of the preset machine learning model based on the evaluation results and feedback. In this case, the optimization unit first optimizes the model parameters using the Adagrad optimization algorithm, adaptively adjusting the learning rate based on the gradient changes of each parameter during training, enabling the model to converge to a more optimal solution more quickly.

[0039] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as above in terms of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can, without departing from the scope of the technical solution of the present invention, make some changes or modifications to equivalent embodiments using the technical contents disclosed above. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A large-screen display system based on machine learning, characterized by: It includes an environment perception module, a user perception module, a calculation module, a compensation module and a display control module; The environment perception module is used to obtain environmental parameters, which include ambient light intensity; The user perception module is used to obtain the viewer's pupil diameter parameter and positioning coordinate parameter; The user perception module and the user perception module respectively transmit the acquired data to the calculation module, and the calculation module calculates the initial brightness adjustment coefficient according to the environmental parameters; The compensation module calculates the distance and viewing angle according to the positioning coordinate parameters, and when the distance and viewing angle exceed the preset distance threshold and viewing angle threshold, the compensation module compensates the initial brightness adjustment coefficient according to the distance, viewing angle and pupil diameter parameters to generate a modified brightness adjustment coefficient; The display control module dynamically adjusts the screen brightness according to the modified brightness adjustment coefficient.

2. The large-screen display system based on machine learning according to claim 1, characterized in that: The compensation module compensates the initial brightness parameter including the following steps: Predict the viewer’s current visual comfort level V using a pre-trained machine learning model based on pupil diameter parameters; Based on the positioning coordinate parameters, the viewer's viewing angle is calculated and the ambient light impact factor R is evaluated; Compensating the initial brightness adjustment coefficient based on the viewer's current visual comfort level and viewing angle to generate a modified brightness adjustment coefficient; Here, compensating the initial brightness adjustment coefficient includes the following steps: Ccorrected = Cinitial ; Here, Ccorrect is the corrected brightness adjustment coefficient, Cinitial is the initial brightness adjustment coefficient, V is the visual quantization, and R is the ambient light intensity influence factor. is the weight coefficient.

3. The large-screen display system based on machine learning according to claim 2, characterized in that: The compensation module includes a model training submodule, and the model training submodule includes a data collection unit, a data preprocessing unit, a data feature unit and a data training unit; The data collection unit is used to obtain historical pupil diameter parameters and ambient light intensity parameters; The data preprocessing unit is used to perform a primary filtering on the historical pupil diameter parameter and the ambient light intensity parameter using the wavelet transform method through the machine learning model, and then perform a secondary filtering on the data after the primary filtering using a Kalman filter, and then perform noise reduction on the data after the secondary filtering, and use mean filling and KNN interpolation to supplement the missing data, and then use a linear transformation method to normalize the supplemented data; The data feature unit is used to determine the original features, use the principal component analysis method to reduce the dimension of the original features, and then classify and encode the original features into numerical features; The data training unit is used to train the neural network model to obtain a preset machine learning model, and the neural network model includes one of a convolutional neural network, a recurrent neural network and a graph neural network.

4. A large-screen display system based on machine learning according to any one of claims 1 to 3, characterized in that: The model training submodule also includes a data evaluation unit, which is used to evaluate the preset machine learning model using test data, and the evaluation indicators include at least one of mean absolute variance, mean square error, root mean square error and R variance.

5. The large-screen display system based on machine learning according to claim 4, characterized in that: The test data is iteratively generated using a generative adversarial network, where the training set of the generative adversarial network is historical pupil diameter parameters and ambient light intensity parameters.

6. The large-screen display system based on machine learning according to claim 5, characterized in that: It also includes a display feedback adjustment module, which is used to receive the feedback result of the current user after adjusting the brightness of the large screen.

7. The large-screen display system based on machine learning according to claim 5, characterized in that: The model training submodule also includes an optimization unit, which is used to perform parameter optimization and structure optimization on the preset machine learning model according to the evaluation results and feedback results.

8. The large-screen display system based on machine learning according to claim 4, characterized in that: It also includes a display recognition module, which is electrically connected to the display control module. The display recognition module transmits recognition content to the display control module. The display content includes: dynamic and non-dynamic. The display control module compensates the brightness adjustment coefficient according to the display content.

9. The large-screen display system based on machine learning according to claim 8, characterized in that: The compensation is that when it is determined that the screen content is in a dynamic state, the display control module increases the display brightness of the display screen.

Images