Image quality optimization method and device applied to optoelectronic device display and storage medium

Abstract

Embodiments of the application disclose a picture quality optimization method and device applied to optoelectronic device display and a storage medium. The method comprises the following steps: extracting a device aging attenuation feature set carrying a use cycle label and an operating environment label; acquiring a human eye visual perception change feature set in different display states and completing multi-dimensional correlation analysis to obtain a perception device linkage feature; determining an action path of device aging on visual perception based on the perception device linkage feature and deducing a physical display feature; generating a linkage optimization migration strategy in combination with a cross-device generality adaptation rule; creating a pre-judgment event recognition model by using the strategy to generate a pre-judgment basis for picture quality optimization triggering; generating a trigger signal based on the pre-judgment basis, and outputting a display parameter adaptation adjustment instruction by an optimization task decision algorithm. Embodiments of the application reconstruct picture quality optimization logic and improve the accuracy and cross-device generality of optimization.

Description

Technical Field

[0001] This application relates to the field of computer vision technology, and in particular to a method, device and storage medium for image quality optimization applied to optoelectronic device displays. Background Technology

[0002] Image quality optimization technology for optoelectronic display devices has always been a core research area in the display field. With the widespread adoption of various display devices such as OLED, LCD, and MicroLED, two mainstream technical directions have emerged in the industry regarding image quality optimization: one focuses on the human visual perception dimension, directly adjusting the output parameters of the display device by collecting subjective or objective perception data of parameters such as brightness, color, and contrast. For example, dynamically adjusting screen brightness based on the visual sensitivity of the human eye under different lighting conditions, or correcting the display color gamut based on the human eye's color preference curve; the other focuses on the performance defects of the device itself, passively repairing components that exhibit attenuation or abnormality by real-time monitoring parameters such as pixel luminous efficiency, driving circuit performance, and backlight module brightness uniformity. For example, for OLED pixel attenuation issues, local correction of pixel brightness is achieved by compensating the driving current, or for LCD backlight unevenness issues, the luminous power of the backlight zones is adjusted. In addition, some research attempts to simply combine the two types of technologies, such as adjusting the compensation magnitude based on the human eye's perception threshold after detecting pixel attenuation. However, such combinations still do not break through the logical limitations of single-dimensional optimization and are merely functional superpositions.

[0003] It is evident that existing technologies have failed to effectively uncover the deep-seated correlation between device aging and degradation and changes in human visual perception, resulting in optimization strategies that are limited by being one-dimensional and passive, and are difficult to reuse across different types of optoelectronic display devices. Summary of the Invention

[0004] This application provides a method, device, and storage medium for optimizing image quality in optoelectronic device displays.

[0005] This application provides, in one aspect, a method for optimizing image quality in optoelectronic device displays, applied to computer equipment, the method comprising:

[0006] Extract the device aging and degradation feature set of the target optoelectronic display device. Each device aging and degradation feature in the device aging and degradation feature set carries the usage period tag and working environment tag corresponding to the target optoelectronic display device. At least one type of tag is different in the usage period tag and working environment tag of different device aging and degradation features.

[0007] A set of human visual perception change features under different display states is obtained, and a multi-dimensional correlation analysis is performed on the set of human visual perception change features and the set of device aging and decay features to obtain the linkage features of the sensing device.

[0008] Based on the linkage characteristics of the sensing device, the path of the effect of device aging on visual perception is determined, and the physical display characteristics corresponding to the target optoelectronic display device are derived through the path of the effect.

[0009] Combining the physical display characteristics and the preset cross-device universality adaptation rules, a linkage optimization migration strategy is generated, which includes optimization directions and parameter adjustment variables for adapting to different device types.

[0010] The aforementioned linkage optimization migration strategy is used to create a predictive event recognition model covering different degradation scenarios. The predictive event recognition model is used to identify the correlation between device aging and decay characteristics and visual perception change characteristics under different degradation scenarios, and generate a preliminary judgment basis for triggering image quality optimization.

[0011] Based on the aforementioned pre-judgment criteria, a picture quality optimization trigger signal is generated and input into the optimization task decision algorithm. The optimization task decision algorithm then combines the real-time physical display status information of the target optoelectronic display device to output a display parameter adaptation and adjustment command.

[0012] One embodiment of this application provides a computer device, including:

[0013] A processor; a storage device having a computer program stored thereon; a network interface for providing network communication functions; when the computer program is executed by the processor, the processor enables the processor to implement any of the image quality optimization methods applied to optoelectronic device displays.

[0014] One embodiment of this application provides a readable storage medium storing a program or instructions, which, when executed by a processor, implements the steps of the image quality optimization method applied to optoelectronic device displays.

[0015] The method described in this application breaks through the rigid thinking of "single-dimensional optimization" in the field of image quality optimization of existing optoelectronic display devices. By constructing a cross-dimensional linkage logic of "device aging-visual perception", it achieves a core transformation from passive defect repair to proactive predictive adaptation. It no longer focuses solely on isolated parameter adjustments based on human eye perception, nor does it perform single passive repairs for device defects. Instead, it extracts device aging and decay feature sets carrying usage cycle and working environment tags, and combines them with human eye visual perception change feature sets under different display states to complete multi-dimensional correlation analysis. This qualitatively uncovers the deep linkage patterns between two independent types of features, thereby clarifying the path of device aging on visual perception and deriving physical display characteristics. On this basis, a linkage optimization migration strategy covering multiple types of optoelectronic display devices is constructed through cross-device universal adaptation rules, breaking the limitations of previous single-device-specific optimization and realizing the underlying refinement of cross-device universal optimization logic. Meanwhile, based on the predictive event recognition model, the correlation patterns under different degradation scenarios are identified to generate preliminary judgment criteria, and the triggering logic and decision-making process of image quality optimization are reconstructed. This fundamentally improves the accuracy, foresight and universality of image quality optimization, and solves the problems of poor adaptability, optimization lag and low cross-device reusability caused by single-dimensional optimization in existing technologies. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for optimizing image quality in optoelectronic device displays, as provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the basic structure of a computer device provided in an embodiment of this application.

[0019] Figure 3 This is a functional block diagram of an image quality optimization device provided in an embodiment of this application. Detailed Implementation

[0020] 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, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0021] Please see Figure 1 , Figure 1 This is a flowchart of a method for optimizing image quality in optoelectronic device displays provided in an embodiment of this application. The method can be executed by a computer device or by a computer device and a server. The method may include steps 110-160.

[0022] In the field of optoelectronic display image quality optimization involved in the embodiments of this application, when processing various physical parameters and features, there may be problems of inconsistent dimensions or scale differences due to the diverse sources and different physical meanings of the parameters. For example, device aging and degradation characteristics may involve time dimensions such as usage period, environmental dimensions such as temperature and humidity, and electrical or optical dimensions such as brightness decay rate; while human visual perception change characteristics may be based on psychophysical quantities such as subjective ratings, contrast sensitivity functions, or color discrimination thresholds. In the process of extracting, associating, and calculating the above parameters, if mathematical operations such as addition, comparison, or fusion are directly performed, the calculation may lack clear physical meaning due to mismatched dimensions. For example, directly summing time units and brightness units, or performing uncalibrated calculations between different scale units such as percentages and absolute values.

[0023] Therefore, those skilled in the art, when implementing the embodiments of this application, can overcome such dimensional errors through adaptive normalization techniques, thereby enhancing the feasibility of the overall solution. Normalization methods can convert parameters with different physical meanings and dimensions into a unified dimensionless scale. For example, by using minimum-maximum scaling to map various feature values ​​to a preset range, or by using Z-score standardization to adjust the data distribution based on the mean and standard deviation, calculation conflicts caused by unit differences can be eliminated. In specific steps, such as extracting multidimensional correlations from device aging and decay feature sets and human visual perception change feature sets, technicians will first preprocess the features, converting time labels, environmental labels, and perception scores into standardized feature vectors. This ensures that subsequent mapping matrix construction, feature transfer model iteration, or parameter adjustment model calculations are all performed under a consistent dimension, avoiding path derivation errors or optimization instruction deviations caused by unit inconsistencies. Furthermore, when generating linkage optimization migration strategies or predictive event recognition models, before fusing feature sets under different degradation scenarios, feature dimension adjustment and normalization weighting can be used to correlate physical display features such as brightness distribution features and color reproduction features with perceptual change features on the same scale, thereby ensuring the physical consistency of the judgment basis for image quality optimization.

[0024] By employing adaptive normalization, potential dimensional calculation problems in the embodiments of this application are pre-solved, ensuring that parameter migration across device types, real-time status information matching, and decision algorithm outputs are all based on comparable and computable data, thus enhancing the engineering applicability of the method. Adaptively normalizing parameters with different physical meanings and dimensions to overcome corresponding dimensional errors is a conventional technique implemented by those skilled in the art based on existing technologies.

[0025] In the fields of optoelectronic displays, image processing, and machine learning, normalization, as a standard step in data preprocessing, is widely used in feature engineering and model training. Existing techniques such as feature scaling, principal component analysis, or domain adaptive methods have been maturely applied to process multidimensional heterogeneous data. For example, in display device calibration, luminance nits values ​​and color coordinate values ​​are normalized through linear transformation and then input into adjustment algorithms; or in visual perception models, subjective evaluation scores and objective measurement parameters are standardized into a unified index.

[0026] The subsequent steps in the embodiments of this application, such as performing feature encoding on physical display features to obtain standardized codes, or dynamically adjusting fusion weights through feature fusion algorithms, essentially rely on normalization techniques to coordinate the scale difference between the device aging influence coefficient and the quantized value perceived by the human eye. Based on conventional knowledge, those skilled in the art will first identify the parameter type, such as continuous or categorical variables, and then select an appropriate normalization method, such as decimal scaling or logarithmic transformation, to ensure that the input historical dataset and real-time feature set are dimensionally compatible when creating a predictive event recognition model, thereby accurately outputting the correlation pattern. Simultaneously, in the optimization task decision algorithm, when comparing the quantized value of features with the trigger condition threshold, normalization ensures that the threshold setting is based on a dimensionless scale, making the judgment logic unaffected by the original units and improving the stability of the instruction output. Therefore, although the original parameters may have a risk of dimensional mismatch, through conventional normalization operations, those skilled in the art can seamlessly integrate the device's physical characteristics and visual perception data, making the image quality optimization method executable from feature extraction to parameter adjustment throughout the entire process, and the results have clear physical meaning. The aforementioned adaptive normalization is not only a standard means of solving subsequent dimensional problems, but also the foundation for achieving cross-device versatility and robustness enhancement, supporting the implementation of advanced functions such as semantic jump analysis, transfer learning, and robustness testing in the solution.

[0027] Step 110: Extract the device aging and degradation feature set of the target optoelectronic display device. Each device aging and degradation feature in the device aging and degradation feature set carries the usage period tag and working environment tag corresponding to the target optoelectronic display device. At least one type of tag is different in the usage period tag and working environment tag of different device aging and degradation features.

[0028] In this embodiment, the computer device directly retrieves all historical operation records from the target optoelectronic display device's operating status database. These records cover multi-dimensional information such as the device's cumulative power-on time, cumulative display time, operating environment temperature and humidity change sequence, and pixel luminous efficiency change sequence. Each record is bound to a corresponding timestamp and environmental acquisition node identifier. Based on the sliding window feature extraction algorithm, the computer device segments the historical operation records into segments with a preset time window length, and each segment is matched with a corresponding usage period tag. Simultaneously, based on the temperature and humidity interval division rules corresponding to the environmental acquisition node identifier, a corresponding operating environment tag is matched for each segment. For the pixel luminous efficiency change sequence within each time window, statistical features such as the mean, variance, and trend slope of the sequence are extracted through feature aggregation rules. For the temperature and humidity change sequence, features such as extreme values ​​and fluctuation amplitude are extracted. All extracted features together constitute a device aging and degradation feature.

[0029] Furthermore, the computer device integrates the features and labels corresponding to all time windows into a multi-dimensional device aging and degradation feature set. This feature set is stored in the form of an array, where each element is a structured object containing a feature vector, a usage period label, and an operating environment label. At least one of the usage period label or operating environment label is different for different elements, so that the feature set covers the aging state of the target optoelectronic display device in different usage stages and environments.

[0030] Step 120: Obtain the human eye visual perception change feature set under different display states, and perform multi-dimensional correlation analysis on the human eye visual perception change feature set and the device aging and decay feature set to obtain the sensing device linkage feature.

[0031] In this embodiment of the application, the computer device retrieves human visual perception records of the target optoelectronic display device under different combinations of display brightness, color mode, and refresh rate through a connected visual perception testing system. The records include display status parameters, human eye brightness perception score, color saturation perception score, contrast perception score, and visual fatigue score for the display status, and each record is bound to a corresponding display status tag.

[0032] Based on feature mapping rules, the computer device converts each visual perception record into a fixed-dimensional feature vector of human visual perception change, and integrates all vectors into a human visual perception change feature set. After introducing a multidimensional association analysis algorithm, the computer device first aligns the dimensions of the device aging and decay feature set with the human visual perception change feature set, establishing an association mapping between the usage cycle label, working environment label, and corresponding display status label for each device aging and decay feature; then, based on the semantic matching degree and statistical correlation of each dimension of the feature vector, it calculates the association strength between each pair of device aging and decay features and human visual perception change features through feature association degree calculation rules; finally, it selects feature combinations with association strength higher than a preset threshold, sorts the above feature combinations according to association strength to form a sensing device linkage feature, which is stored in the form of an adjacency list, where each node represents a feature combination, and each edge represents the association relationship between features and the corresponding association strength value.

[0033] Step 130: Determine the effect path of device aging on visual perception based on the linkage characteristics of the sensing device, and deduce the physical display characteristics corresponding to the target optoelectronic display device through the effect path.

[0034] In this embodiment of the application, the computer device performs the following sub-steps based on the linkage characteristics of the sensing device:

[0035] Step 131: Decompose the linkage features of the sensing device into layers to obtain the first dimension features related to device aging and the second dimension features related to visual perception, and establish a mapping relationship matrix between the first dimension features and the second dimension features. Based on the mapping relationship matrix, select feature combinations with causal correlation.

[0036] Based on the feature-level decomposition rules, the computer device divides the sensing device linkage features into first-dimensional features (related to device aging) and second-dimensional features (related to visual perception) based on the feature attribute labels. The first-dimensional features include sub-features such as pixel luminous efficiency decay rate and driving circuit performance degradation rate, while the second-dimensional features include sub-features such as brightness perception deviation and color saturation perception deviation. Subsequently, the computer device constructs a mapping relationship matrix based on the correlation strength between features. The rows of the matrix correspond to the first-dimensional features, the columns correspond to the second-dimensional features, and the matrix element values ​​are the correlation strength between the corresponding features. After introducing causal association filtering rules, the computer device analyzes the temporal correlation of the element values ​​in the mapping relationship matrix to determine whether there is a cause-and-effect temporal logic between the first-dimensional features and the second-dimensional features, filtering out feature combinations that meet the causal association conditions. These combinations exclude simple statistical correlations, ensuring that the subsequent deduced action paths have physical causality.

[0037] Step 132: Generate an initial action path network using the feature combination as the target path node, and perform correlation enhancement processing on the path nodes in the initial action path network to obtain the action path of device aging on visual perception.

[0038] The computer device uses the causal relationship feature combinations selected in step 131 as target path nodes, and connects the nodes according to the causal logic order between the features to generate an initial action path network. For each node in the initial action path network, the computer device introduces a correlation enhancement algorithm. By calculating the feature transmission probability between a node and its neighboring nodes, the connection weights between nodes are adjusted. Nodes with higher feature transmission probabilities are connected with larger corresponding weight values. At the same time, the computer device removes weakly correlated nodes in the initial network whose connection weights are below a preset threshold to avoid redundant nodes interfering with the accuracy of the action path. After the correlation enhancement processing, a clear action path of device aging on visual perception is formed. This path is presented in the form of a directed acyclic graph, where each node represents an aging feature or a perception feature, and each directed edge represents the causal transmission relationship between features and the corresponding transmission weight.

[0039] Step 133: Create a feature transfer model based on the feature transfer weights of each path node in the action path, input the device aging and decay feature set and the human visual perception change feature set into the feature transfer model, and obtain intermediate feature parameters in the feature transfer process through iterative calculation of the feature transfer model.

[0040] The computer device extracts the feature transfer weights of each node in the action path and constructs a feature transfer model based on these weights. This model includes a feature input layer, a transfer calculation layer, and an intermediate output layer. The computational logic of the transfer calculation layer perfectly matches the feature transfer relationship of the action path. During model execution, the computer device synchronously inputs the device aging and degradation feature set and the human visual perception change feature set into the feature transfer model. The model performs feature calculations node by node based on preset transfer weights. At each node, the feature data is updated and adjusted according to the transfer weights. Through multiple rounds of iterative computation, the model outputs intermediate feature parameters during the feature transfer process. These parameters include the feature change amount, transfer efficiency value, and feature loss rate at each node during the transfer process, thus completely reconstructing the transfer process from aging features to perceived features.

[0041] Step 134: Based on the correspondence between the intermediate feature parameters and the basic display parameters of the target optoelectronic display device, the physical display features corresponding to the target optoelectronic display device are derived. The physical display features include brightness distribution features, color reproduction features and contrast variation features, and each feature carries a corresponding aging influence coefficient.

[0042] The computer equipment pre-retrieves the basic display parameters of the target optoelectronic display device, including the standard brightness distribution curve, standard color reproduction matrix, and standard contrast threshold information at the factory. It then matches the intermediate feature parameters output by the feature transfer model with the basic display parameters one by one. For example, it matches the intermediate parameter corresponding to the attenuation of pixel luminous efficiency with the standard brightness distribution curve, deriving the current brightness distribution characteristics through feature difference analysis; it matches the intermediate parameter corresponding to the attenuation of color filters with the standard color reproduction matrix, deriving the current color reproduction characteristics; and it matches the intermediate parameter corresponding to the performance degradation of the backlight module with the standard contrast threshold, deriving the current contrast change characteristics. During the derivation process, the computer equipment calculates the deviation rate between each physical display feature and the basic display parameters, and binds this deviation rate as an aging effect coefficient to the corresponding physical display feature. This ultimately forms a physical display feature set containing brightness distribution features, color reproduction features, and contrast change features. Each feature is a multi-dimensional vector carrying a corresponding aging effect coefficient.

[0043] Step 140: Combining the physical display characteristics and the preset cross-device universality adaptation rules, generate a linkage optimization migration strategy that includes optimization directions and parameter adjustment variables for adapting to different device types.

[0044] Step 141: Obtain the preset cross-device universality adaptation rules, parse the adaptation conditions and parameter constraint ranges of different device types contained in the cross-device universality adaptation rules, and establish a correspondence library between adaptation rules and device types.

[0045] The computer device retrieves cross-device universal adaptation rules from a pre-set rule database. These rules are constructed based on the common display principles of different types of optoelectronic display devices, covering the adaptation requirements of various common optoelectronic display devices such as OLEDs, liquid crystal displays, and micro-LEDs. Using a rule parsing algorithm, the computer device breaks down the adaptation rules into adaptation conditions and parameter constraint ranges corresponding to different device types. For example, the adaptation condition for OLEDs is the range of pixel luminous efficiency deviation, and the parameter constraint range includes the maximum brightness adjustment range and the color gamut range for color correction. The computer device then binds the decomposed adaptation conditions and parameter constraint ranges to the corresponding device types, establishing a structured database of the correspondence between adaptation rules and device types. Each record in the database contains a device type identifier, a set of adaptation conditions, and a set of parameter constraint ranges.

[0046] Step 142: Perform feature encoding processing on the physical display features to obtain standardized physical display feature codes. Match the physical display feature codes with the adaptation conditions in the corresponding relationship library one by one to determine the basic adaptation rules corresponding to the target optoelectronic display device.

[0047] The computer equipment introduces a feature encoding algorithm to standardize and encode each feature in the physical display feature set. During the encoding process, the aging impact coefficient and dimensional information of the features are preserved, ensuring that the encoded features accurately reflect the actual state of the physical display features. The standardized physical display feature codes are then matched one-to-one with the adaptation conditions in a corresponding relational database. For example, the brightness distribution feature code is matched with the brightness adaptation conditions of an organic light-emitting diode (OLED) display device to determine whether the current brightness distribution feature meets the device's adaptation requirements. Through multiple rounds of matching, the computer equipment selects adaptation conditions and parameter constraints that perfectly match the target optoelectronic display device type and the current physical display state, forming the basic adaptation rules corresponding to the target optoelectronic display device.

[0048] Step 143: Based on the parameter constraint range in the basic adaptation rules and combined with the aging effect coefficient in the physical display characteristics, determine the image quality optimization direction for different device types. The optimization direction covers brightness calibration, color correction and contrast adjustment.

[0049] The computer equipment uses the parameter constraints in the basic adaptation rules as a benchmark, combined with the aging effect coefficient in the physical display characteristics, to analyze the current display defect type and severity of the target optoelectronic display device. For example, when the aging effect coefficient of the brightness distribution characteristic exceeds the upper limit of the parameter constraint range, it is determined that optimization in the brightness calibration direction needs to be initiated; when the aging effect coefficient of the color reproduction characteristic exceeds the upper limit of the parameter constraint range, it is determined that optimization in the color correction direction needs to be initiated; when the aging effect coefficient of the contrast variation characteristic exceeds the upper limit of the parameter constraint range, it is determined that optimization in the contrast adjustment direction needs to be initiated. At the same time, the computer equipment will refer to the parameter constraint range of other device types in the corresponding relational database to analyze the optimization directions for the same type of display defect for different device types, forming a set of image quality optimization directions adapted to different device types. This set clearly marks the optimization direction and corresponding defect judgment criteria for each device type.

[0050] Step 144: Construct a parameter adjustment model for each optimization direction, input the physical display characteristics and the corresponding aging influence coefficient into the parameter adjustment model, and obtain the parameter adjustment variables for each optimization direction through model calculation. The parameter adjustment variables include adjustment range, adjustment step size and adjustment period.

[0051] For each defined optimization direction, the computer equipment constructs a corresponding parameter adjustment model. The model's operational logic is designed based on the parameter constraint range and the changing trend of the aging influence coefficient in the basic adaptation rules. For example, the parameter adjustment model for the brightness calibration direction includes a deviation analysis module for brightness distribution characteristics, an adjustment amplitude calculation module, an adjustment step size adaptation module, and an adjustment cycle planning module. After inputting the physical display characteristics and the corresponding aging influence coefficient into the parameter adjustment model, the model first analyzes the degree of deviation between the current display characteristics and the standard display characteristics. Combining this with the changing trend of the aging influence coefficient, it calculates the adjustment amplitude that can bring the display characteristics back to the parameter constraint range. Then, based on the hardware response speed of the target optoelectronic display device, it determines the adjustment step size for each adjustment. Finally, based on the rate of change of the aging influence coefficient, it plans the adjustment cycle so that the parameter adjustment can dynamically adapt to the aging state of the device. After the model's calculation, it outputs the parameter adjustment variables for each optimization direction. Each variable is structured data containing the adjustment amplitude, adjustment step size, and adjustment cycle.

[0052] Step 145: Combine the optimization directions for different device types and the corresponding parameter adjustment variables to generate a structured linkage optimization migration strategy. The linkage optimization migration strategy also includes parameter migration logic between different device types and priority ranking of optimization directions.

[0053] The computer system integrates optimization directions and corresponding parameter adjustment variables adapted to different device types, categorizing and arranging them according to device type to form a structured, interconnected optimization migration strategy. During integration, the computer system introduces parameter migration rules to analyze the transferability of parameter adjustment variables between different device types. For example, the brightness adjustment range parameter for OLED displays can be appropriately converted and migrated based on the brightness response characteristics of LCD displays. Simultaneously, the computer system prioritizes the optimization directions for each device type based on their importance to image quality; for example, brightness calibration has higher priority than color correction, and color correction has higher priority than contrast adjustment. The final generated interconnected optimization migration strategy includes device type identification, corresponding optimization directions, parameter adjustment variables, parameter migration logic, and optimization direction priority ranking, providing a complete guidance scheme for image quality optimization of different types of optoelectronic display devices.

[0054] Step 150: Use the aforementioned linkage optimization migration strategy to create a predictive event recognition model covering different degradation scenarios. Use the predictive event recognition model to identify the correlation between device aging and decay characteristics and visual perception change characteristics under different degradation scenarios, and generate a preliminary judgment basis for triggering image quality optimization.

[0055] Step 151: Based on the optimization direction and parameter adjustment variables in the linkage optimization migration strategy, various device degradation scenarios are obtained, and the scenario feature identifiers corresponding to each degradation scenario are extracted.

[0056] Based on the optimization direction and parameter adjustment variables in the linkage optimization migration strategy, the computer equipment classifies the possible degradation states of the target optoelectronic display device into scenarios. For example, a state where the brightness distribution characteristic deviation exceeds the parameter constraint range and the aging influence coefficient continues to rise is classified as a rapid brightness degradation scenario; a state where the color reproduction characteristic deviation exceeds the parameter constraint range and the aging influence coefficient fluctuates greatly is classified as a color unstable degradation scenario; and a state where the contrast change characteristic deviation exceeds the parameter constraint range and the aging influence coefficient rises slowly is classified as a slow contrast degradation scenario. For each degradation scenario, the computer equipment extracts the corresponding scenario feature identifier, which includes information such as degradation type, degradation degree, and change trend, for subsequent rapid scenario identification and matching.

[0057] Step 152: Collect historical device aging and decay feature sets and historical human visual perception change feature sets under different degradation scenarios, and associate the historical device aging and decay feature sets and historical human visual perception change feature sets with the corresponding scene feature labels to generate a prediction event training dataset.

[0058] The computer equipment retrieves historical operating records under different degradation scenarios from a historical database, extracting corresponding historical device aging and degradation feature sets and historical human visual perception change feature sets. These feature sets have the same dimensions and attribute characteristics as the feature set of the current target optoelectronic display device. The computer equipment then binds the historical device aging and degradation feature sets and historical human visual perception change feature sets with corresponding scene feature identifiers. Each record contains a subset of historical device aging and degradation features, a subset of historical human visual perception change features, and a corresponding scene feature identifier. All bound records are integrated into a prediction event training dataset, stored as an array, ensuring that each data sample corresponds to a specific degradation scenario.

[0059] Step 153: Based on the parameter adjustment logic in the linkage optimization migration strategy, build a model network for the predictive event recognition model. The model network includes a feature input layer, a feature fusion layer, and a pattern output layer.

[0060] The computer equipment uses the parameter adjustment logic in the linkage optimization migration strategy as its core to build a model network for predictive event recognition. The feature input layer receives subsets of historical device aging and decay features and subsets of historical human visual perception change features, and performs dimensionality verification and format conversion on the input features to ensure they meet the processing requirements of subsequent layers. The feature fusion layer, based on the parameter association logic in the linkage optimization migration strategy, fuses the two types of input features, adjusting the fusion weights based on the correlation strength between the features during the fusion process. The pattern output layer converts the fused features into association patterns under deterioration scenarios, and the output patterns include the collaborative relationships between feature changes, triggering conditions, and the degree of influence. Data transfer between the layers of the model network is achieved through fully connected layers, enabling efficient feature flow and processing within the network.

[0061] Step 154: Input the predicted event training dataset into the predicted event recognition model for training. By adjusting the network parameters of the predicted event recognition model, the deviation between the correlation rules output by the predicted event recognition model and the correlation rules in the actual scene is within a preset range, thus completing the creation of the predicted event recognition model.

[0062] The computer device divides the predicted event training dataset into training, validation, and test sets, allocating data samples according to a preset ratio. During the training phase, the computer device inputs the training set into the predicted event recognition model. The model processes features based on the fusion logic of the feature fusion layer and outputs association patterns through the pattern output layer. The association patterns output by the model are compared with the association patterns recorded in the actual scene, and the deviation value between the two is calculated. When the deviation value exceeds a preset range, the computer device adjusts the model's network parameters, including the fusion weights of the feature fusion layer and the connection weights of the fully connected layer, using a backpropagation algorithm. This training process is iterated repeatedly until the deviation between the association patterns output by the model on the validation set and the actual association patterns is within a preset range. Then, the model's generalization ability is validated using a test set. Once validation is successful, the predicted event recognition model is created.

[0063] Step 155: Input the real-time device aging and decay feature set and the real-time human visual perception change feature set under different degradation scenarios into the trained prediction event recognition model. The feature fusion layer performs fusion processing on the real-time device aging and decay feature set and the real-time human visual perception change feature set. The correlation rule between device aging and decay features and visual perception change features under different degradation scenarios is output through the rule output layer.

[0064] The computer equipment inputs the real-time device aging and degradation feature set of the target optoelectronic display device and the real-time human visual perception change feature set into the trained predictive event recognition model, and executes the following sub-steps:

[0065] Step 1551: After receiving the real-time device aging and decay feature set and the real-time human visual perception change feature set through the feature fusion layer, the feature dimensions of the real-time device aging and decay feature set are adjusted so that the dimensions of each feature in the real-time device aging and decay feature set match the dimensions of the corresponding features in the real-time human visual perception change feature set, resulting in two feature sets with unified dimensions.

[0066] After receiving the real-time device aging and decay feature set and the real-time human visual perception change feature set, the feature fusion layer first verifies the dimensions of the two feature sets. When it is found that the dimensions of some features in the real-time device aging and decay feature set do not match the dimensions of the corresponding features in the real-time human visual perception change feature set, the feature dimension adjustment algorithm is used to expand or compress the dimensions of the real-time device aging and decay feature set. For example, the feature interpolation method is used to expand the dimensions of low-dimensional features, and the feature dimensionality reduction method is used to compress the dimensions of high-dimensional features, so that the dimensions of each feature in the adjusted real-time device aging and decay feature set are completely consistent with the dimensions of the corresponding features in the real-time human visual perception change feature set, forming two feature sets with unified dimensions.

[0067] Step 1552: Based on the preset feature association rules in the predicted event recognition model, perform feature association mapping on the two types of feature sets with uniform dimensions, and establish a one-to-one relationship between the aging and decay features of each real-time device and the corresponding real-time human visual perception change features.

[0068] Based on the pre-defined feature association rules in the predictive event recognition model, the computer device performs feature association mapping on two sets of features with uniform dimensions. These feature association rules are formulated based on the physical causal relationship between aging features and perceived features; for example, pixel luminous efficiency decay features correspond to brightness perception deviation features, and color filter decay features correspond to color saturation perception deviation features. Through feature association mapping, the computer device matches the aging decay features of each real-time device with corresponding real-time human visual perception change features, establishing a one-to-one association. This allows subsequent fusion processing to accurately target causally related feature combinations.

[0069] Step 1553: The feature fusion algorithm built into the feature fusion layer is used to fuse features with one-to-one relationships to generate a fused feature set. The feature fusion algorithm dynamically adjusts the fusion weights based on the correlation strength between the two types of features. Feature combinations with high correlation strength correspond to higher fusion weights.

[0070] The feature fusion layer's built-in feature fusion algorithm dynamically adjusts fusion weights based on the strength of the correlation between features; feature combinations with higher correlation strength receive larger fusion weights. The computer device determines the correlation strength of each feature combination by calculating the semantic matching degree and statistical correlation between features with one-to-one relationships. Then, based on the correlation strength, fusion weights are assigned to each feature combination. The real-time device aging and degradation features and real-time human visual perception change features in each feature combination are concatenated according to the fusion weights to generate a fused feature vector. All fused feature vectors are integrated into a fused feature set, which accurately reflects the collaborative change relationship between aging features and perception features.

[0071] Step 1554: Perform feature denoising processing on the fused feature set to obtain a purified fused feature set. Pass the purified fused feature set to the regularity output layer. The regularity output layer parses the feature attributes and feature associations of each fused feature in the purified fused feature set. Based on the feature attributes and feature associations, multiple feature association clusters are obtained. Each feature association cluster corresponds to a feature association type under a deterioration scenario.

[0072] The computer device employs a feature denoising algorithm to denoise the fused feature set, removing noisy features that may be generated by data acquisition errors or random interference, affecting the accuracy of pattern recognition. After obtaining the purified fused feature set, the computer device passes it to the pattern output layer. The pattern output layer first analyzes the feature attributes of each fused feature in the purified feature set, including the aging type, perception type, and change trend of the feature; then it analyzes the feature correlations between the fused features, including correlation direction, correlation strength, and synergistic change. Based on the feature attributes and correlations, the computer device divides the fused features into multiple feature correlation clusters. Each feature correlation cluster corresponds to a feature correlation type under a certain degradation scenario; for example, a rapid brightness degradation scenario corresponds to one feature correlation cluster, and a color unstable degradation scenario corresponds to another feature correlation cluster.

[0073] Step 1555: Perform association pattern recognition processing on each feature association cluster, determine the dynamic association trend between real-time device aging and decay characteristics and real-time human visual perception change characteristics within the same feature association cluster, and bind the dynamic association trend to the scene in combination with the corresponding scene feature identifier to obtain the association pattern fragment of the bound scene information.

[0074] For each feature association cluster, the computer equipment introduces an association pattern recognition algorithm to analyze the dynamic changing trends of real-time device aging and degradation characteristics and real-time human visual perception changes within the cluster, including the order of feature changes, the proportional relationship of the change magnitude, and the triggering conditions of the changes. The identified dynamic association trends are bound to the corresponding scene feature identifiers to clarify the degradation scene to which the dynamic association trend belongs, forming association pattern fragments bound to scene information. Each fragment contains scene feature identifiers, dynamic association trends, triggering conditions, and other information.

[0075] Step 1556: Integrate the association pattern fragments of all bound scene information to obtain association patterns covering different deterioration scenarios. The association patterns include the initiation conditions of feature association, feature transmission paths, and feature change synergy relationships under each deterioration scenario.

[0076] The computer device integrates all the associated pattern fragments of bound scene information, classifies and arranges the fragments according to the type of degradation scene, forming association patterns covering different degradation scenes. During the integration process, the computer device performs deduplication and merging of the associated pattern fragments to avoid duplicate pattern information affecting subsequent use. The final association patterns include the initiation conditions of feature association under each degradation scene (such as the aging effect coefficient reaching a certain threshold), feature transmission paths (such as the transmission path from pixel luminous efficiency decay to brightness perception deviation), and the synergistic relationship of feature changes (such as the proportional relationship between the increase of aging features and the synchronous decrease of perception features), which can provide an accurate basis for triggering image quality optimization.

[0077] Step 156: Set the trigger condition threshold for image quality optimization based on the correlation pattern, and generate the pre-judgment basis for triggering image quality optimization based on the trigger condition threshold and the corresponding correlation pattern features.

[0078] Based on the triggering conditions in the correlation rules, computer equipment, combined with the hardware performance of the target optoelectronic display device and the user's requirements for image quality, sets trigger thresholds for image quality optimization. For example, when the aging influence coefficient of the brightness distribution feature reaches a certain threshold, optimization in the brightness calibration direction is triggered; when the aging influence coefficient of the color reproduction feature reaches a certain threshold, optimization in the color correction direction is triggered. Binding the trigger thresholds to the corresponding correlation rule features forms a preliminary judgment criterion for image quality optimization triggering. This criterion includes the degradation scene type, trigger threshold, corresponding correlation rule features, and optimization direction hints, providing clear judgment standards for subsequent image quality optimization triggering.

[0079] Step 160: Generate an image quality optimization trigger signal based on the aforementioned pre-judgment criteria and input it into the optimization task decision algorithm. The optimization task decision algorithm then outputs a display parameter adaptation and adjustment command in conjunction with the real-time physical display status information of the target optoelectronic display device.

[0080] Step 161: Analyze the trigger condition threshold and correlation feature in the pre-judgment criteria, quantize the correlation feature to obtain the feature quantization value, and compare the feature quantization value with the trigger condition threshold.

[0081] The computer device analyzes the trigger condition thresholds and correlation features in the pre-judgment criteria. It then quantifies these correlation features, converting semantic feature descriptions into calculable values. For example, it converts the description of "severe deviation" in brightness perception into a corresponding quantized value, and the description of "moderate deviation" in color saturation perception into a corresponding quantized value. The quantized feature values ​​are then compared with the trigger condition thresholds to determine whether the display state of the current target optoelectronic display device meets the trigger conditions for image quality optimization.

[0082] Step 162: When the feature quantization value reaches the trigger condition threshold, a corresponding image quality optimization trigger signal is generated. The image quality optimization trigger signal carries a trigger level identifier and the corresponding degraded scene features.

[0083] When the feature quantization value reaches the trigger condition threshold, the computer device generates a corresponding image quality optimization trigger signal. The trigger signal carries a trigger level identifier, which is divided into three levels—mild, moderate, and severe—based on the degree to which the feature quantization value exceeds the trigger condition threshold. Different levels correspond to different optimization urgency. At the same time, the trigger signal carries corresponding degradation scene features, including degradation type, degradation degree, and change trend information, for accurate decision-making in subsequent optimization tasks.

[0084] Step 163: Input the image quality optimization trigger signal into the preset optimization task decision algorithm, and call the real-time physical display status information of the target optoelectronic display device through the optimization task decision algorithm. Match and analyze the real-time physical display status information with the degradation scene features in the image quality optimization trigger signal to determine the optimization task type corresponding to the current device display status.

[0085] The computer device inputs the image quality optimization trigger signal into a preset optimization task decision algorithm. The algorithm first retrieves the real-time physical display state information of the target optoelectronic display device, including current brightness distribution characteristics, color reproduction characteristics, contrast change characteristics, and the corresponding aging effect coefficient. It then performs matching analysis between the real-time physical display state information and the degradation scene characteristics in the image quality optimization trigger signal. For example, it matches the aging effect coefficient of the current brightness distribution characteristics with the characteristics of a rapid brightness degradation scene to determine whether the current device display state belongs to a rapid brightness degradation scene. Through multi-dimensional matching analysis, it determines the optimization task type corresponding to the current device display state, such as a brightness calibration task, a color correction task, or a contrast adjustment task.

[0086] Step 164: Based on the optimization task type, retrieve the corresponding optimization direction and parameter adjustment variables in the linkage optimization migration strategy, input the real-time physical display status information into the calculation model corresponding to the parameter adjustment variables, and obtain the real-time parameter adjustment requirements.

[0087] Step 1641: Determine the mapping knowledge graph between the task type and the optimization direction based on the optimization task type. Retrieve the optimization direction and corresponding parameter adjustment variables that match the current optimization task type from the linkage optimization migration strategy through the mapping knowledge graph. The parameter adjustment variables carry parameter attribute identifiers that are adapted to the optimization direction.

[0088] Based on a pre-defined knowledge graph mapping task types and optimization directions, the computer device retrieves the optimization direction and corresponding parameter adjustment variables that match the current optimization task type from the linked optimization migration strategy. The knowledge graph contains a one-to-one correspondence between task types and optimization directions; for example, a brightness calibration task corresponds to a brightness calibration direction, and a color correction task corresponds to a color correction direction. The retrieved parameter adjustment variables carry parameter attribute identifiers that are compatible with the optimization direction. For example, parameter adjustment variables for the brightness calibration direction carry a "brightness" attribute identifier, and parameter adjustment variables for the color correction direction carry a "color" attribute identifier, enabling subsequent processing to accurately match the corresponding features.

[0089] Step 1642: Perform feature analysis on the real-time physical display state information of the target optoelectronic display device, extract the state features related to the retrieved optimization direction from the real-time physical display state information to obtain the target state feature set, the target state feature set includes state features corresponding to the brightness calibration direction, color correction direction or contrast adjustment direction.

[0090] The computer equipment performs feature analysis on the real-time physical display state information of the target optoelectronic display device. Based on the parameter attribute identifiers corresponding to the retrieved optimization direction, it extracts state features related to that optimization direction from the real-time physical display state information. For example, when the optimization direction is brightness calibration, it extracts brightness distribution features and the corresponding aging effect coefficient; when the optimization direction is color correction, it extracts color reproduction features and the corresponding aging effect coefficient; and when the optimization direction is contrast adjustment, it extracts contrast change features and the corresponding aging effect coefficient. The extracted state features are integrated into a target state feature set, which accurately reflects the parts of the current device display state that need optimization.

[0091] Step 1643: Based on the parameter attribute identifier of the parameter adjustment variable, establish the attribute matching relationship between the target state feature set and the parameter adjustment variable, so that each target state feature corresponds to a parameter adjustment variable with the same attribute identifier.

[0092] The computer device matches each feature in the target state feature set with the corresponding parameter adjustment variable based on the parameter attribute identifier of the parameter adjustment variable. For example, it matches the brightness distribution feature with the parameter adjustment variable carrying the "brightness" attribute identifier, and the color reproduction feature with the parameter adjustment variable carrying the "color" attribute identifier, thus establishing an attribute matching relationship. This ensures that the state features and the attributes of the parameter adjustment variables are consistent during subsequent calculations, avoiding calculation errors caused by attribute mismatches.

[0093] Step 1644: Input the target state feature set and parameter adjustment variable after establishing the attribute matching relationship into the preset operation model. The operation model performs feature difference analysis based on the current feature value of the target state feature and the preset standard state feature value in the linkage optimization migration strategy to obtain feature difference information. Based on the feature difference information, adjust the initial value of the parameter adjustment variable so that the adjusted parameter adjustment variable is adapted to the feature difference information.

[0094] The computer equipment inputs the target state feature set (after establishing attribute matching relationships) and parameter adjustment variables into a preset computational model. The model first compares the current feature values ​​of the target state features with the preset standard state feature values ​​in the linkage optimization migration strategy, analyzing the differences between the two to obtain feature difference information, including the direction, magnitude, and trend of the differences. Based on this feature difference information, the computational model adjusts the initial values ​​of the parameter adjustment variables. For example, when the current feature value of the brightness distribution feature is lower than the standard state feature value, the adjustment magnitude of the brightness calibration parameter in the parameter adjustment variables is increased; when the current feature value of the color reproduction feature deviates from the standard state feature value in a reddish direction, the parameter value of the color correction parameter in the parameter adjustment variables is adjusted so that the adjusted parameter adjustment variables can accurately adapt to the current feature differences.

[0095] Step 1645: Through iterative calculation of the computational model, the adjusted parameter adjustment variables are dynamically adapted to the target state features so that the values ​​of the parameter adjustment variables match the differences in the target state features.

[0096] The computational model initiates an iterative computation process. In each iteration, the adjusted parameter variable is applied to the target state feature to simulate the displayed state after parameter adjustment. The simulated displayed state is then compared with the standard state feature value to calculate new feature difference information. Based on the new feature difference information, the value of the parameter variable is adjusted again, and this process is repeated until the value of the parameter variable ensures that the difference between the target state feature value and the standard state feature value is within a preset range, completing the dynamic adaptation process. At this point, the parameter variable accurately matches the difference requirements of the target state feature.

[0097] Step 1646: Based on the values ​​of the parameter adjustment variables after iterative calculation and the corresponding feature difference information, generate real-time parameter adjustment requirements. The real-time parameter adjustment requirements include the parameter adjustment type corresponding to the optimization direction, the range boundary of parameter adjustment, and the dynamic adaptation rules of parameter adjustment. The dynamic adaptation rules are used to guide the adaptive adjustment based on the changes in the real-time physical display state during the parameter adjustment process.

[0098] The computer device generates real-time parameter adjustment requirements based on the values ​​of parameter adjustment variables after iterative calculations and the corresponding feature differences. The requirements clearly define the parameter adjustment type corresponding to the optimization direction; for example, the parameter adjustment type for brightness calibration is backlight module current adjustment, and the parameter adjustment type for color correction is color filter voltage adjustment. They also clearly define the boundaries of the parameter adjustment range, including the minimum and maximum adjustment amplitudes and step size limits. Furthermore, they include dynamic adaptation rules for parameter adjustment. These rules guide the real-time adjustment of parameter adjustment variables based on changes in the real-time physical display state during the parameter adjustment process, ensuring that the parameter adjustment always matches the changing needs of the device's display state.

[0099] Step 165: Combining the real-time parameter adjustment requirements and the decision logic in the optimization task decision algorithm, derive the adjustment scheme for the display parameters. The adjustment scheme includes the adjustment parameter type, adjustment value, and adjustment timing.

[0100] Step 1651: Analyze the real-time parameter adjustment requirements and the decision logic in the optimization task decision algorithm, extract the parameter adjustment types, parameter adjustment range boundaries, and dynamic adaptation rules contained in the real-time parameter adjustment requirements, and decompose the task priority determination rules, parameter adjustment conflict resolution rules, and adjustment scheme testing rules covered in the decision logic. Establish a correlation derivation model between the core elements of the real-time parameter adjustment requirements and the rule elements of the decision logic, so that each parameter adjustment type corresponds to a rule entry with the same task attribute in the decision logic.

[0101] The computer equipment analyzes the decision logic in the real-time parameter adjustment requirements and optimization task decision-making algorithm, extracting the parameter adjustment types, parameter adjustment range boundaries, and dynamic adaptation rules from the real-time parameter adjustment requirements. It also breaks down the task priority determination rules, parameter adjustment conflict resolution rules, and adjustment scheme testing rules from the decision logic. A correlation derivation model is established between the core elements of the real-time parameter adjustment requirements and the rule elements of the decision logic. For example, the parameter adjustment type for brightness calibration is mapped to the brightness optimization-related items in the task priority determination rules, and the parameter adjustment type for color correction is mapped to the color optimization-related items in the task priority determination rules, ensuring that each parameter adjustment type can be matched with a corresponding decision rule.

[0102] Step 1652: Based on the established correlation derivation model, retrieve the optimization direction priority ranking and parameter migration logic corresponding to the current optimization task type in the linkage optimization migration strategy, compare the parameter adjustment range boundary in the real-time parameter adjustment requirement with the preset standard parameter range in the linkage optimization migration strategy, determine the operable range of parameter adjustment, and prioritize the multiple parameter adjustment types involved in the real-time parameter adjustment requirement according to the task priority determination rule in the decision logic to obtain the parameter adjustment execution sequence.

[0103] Based on the correlation derivation model, the computer equipment retrieves the optimization direction priority ranking and parameter migration logic corresponding to the current optimization task type from the linkage optimization migration strategy. It compares the parameter adjustment range boundary in the real-time parameter adjustment requirement with the preset standard parameter range in the linkage optimization migration strategy to determine the operable range for parameter adjustment. This range is the safe and effective range for parameter adjustment; adjustments exceeding this range may damage the device or fail to achieve the expected optimization effect. According to the task priority determination rules in the decision logic, the multiple parameter adjustment types involved in the real-time parameter adjustment requirement are prioritized. For example, brightness calibration has a higher priority than color correction, and color correction has a higher priority than contrast adjustment, resulting in the parameter adjustment execution sequence, ensuring that optimization tasks are executed sequentially according to their urgency.

[0104] Step 1653: Continuously input the real-time physical display status information of the target optoelectronic display device into the associated derivation model. Through the parameter adjustment conflict resolution rules in the decision logic, perform conflict coordination processing on the adjustment types with parameter adjustment conflicts in the execution sequence. Based on the dynamic adaptation rules and the changing trend of the real-time physical display status information, dynamically correct the execution sequence of parameter adjustment and the corresponding adjustment range boundary, so that the corrected execution sequence is adapted to the changing requirements of the real-time physical display status.

[0105] The computer continuously inputs real-time physical display status information of the target optoelectronic display device into the correlation derivation model. Through parameter adjustment conflict resolution rules in the decision logic, it coordinates and handles conflicting adjustment types within the execution sequence. For example, when brightness calibration adjustment parameters and color correction adjustment parameters influence each other, the conflict is eliminated by adjusting the adjustment timing or magnitude. Simultaneously, based on dynamic adaptation rules and the changing trends of real-time physical display status information, the execution sequence of parameter adjustments and the corresponding adjustment range boundaries are dynamically corrected. For instance, when the device's aging rate exceeds expectations, the magnitude of parameter adjustments is increased, and higher-priority optimization tasks are executed earlier, ensuring that the corrected execution sequence always adapts to the changing needs of the real-time physical display status.

[0106] Step 1654: Based on the corrected parameter adjustment execution sequence and adjustment range boundary, construct the initial task description of the adjustment scheme. Analyze the adjustment parameter types, adjustment values, and adjustment timing in the initial task description using the adjustment scheme test rules of the decision logic. If the initial task description fails the test, return to the parameter adjustment execution sequence correction step, and revise the execution sequence and adjustment range boundary again based on the evaluation feedback information until the initial task description passes the test.

[0107] The computer equipment adjusts the execution sequence and adjustment range boundaries based on the corrected parameters, constructing an initial task description of the adjustment scheme. This description clearly defines the types, values, and timing of the adjusted parameters. The initial task description is analyzed using adjustment scheme test rules based on decision logic. These rules include tests for the validity of the adjusted parameters, the safety of the adjusted values, and the rationality of the adjustment timing. If the initial task description fails the tests—for example, if the adjusted values ​​exceed the device's safety range or the adjustment timing results in poor optimization—the computer equipment returns to the parameter adjustment execution sequence correction step. It then re-corrects the execution sequence and adjustment range boundaries based on evaluation feedback, reconstructs the initial task description, and re-tests it until the initial task description passes all tests.

[0108] Step 1655: Using the adjustment values, adjustment timing, and execution priority corresponding to each parameter adjustment type, perform a structured integration operation on the initial task description that has passed the test. During the integration process, synchronously associate the identification information of the target optoelectronic display device, so that the standardized adjustment entries formed after integration correspond one-to-one with the specific display parameters of the target optoelectronic display device; based on the standardized adjustment entries, derive a display parameter adjustment scheme containing adjustment logic and execution details. The display parameter adjustment scheme is used to adapt to the conversion requirements of display parameter adaptation adjustment instructions.

[0109] The computer equipment uses the adjustment values, timing, and execution priority corresponding to each parameter adjustment type to perform a structured integration operation on the initial task description that has passed the test. During the integration process, the identification information of the target optoelectronic display device is synchronously associated, so that each standardized adjustment item can be mapped to the specific display parameters of the target optoelectronic display device. For example, the adjustment item for brightness calibration is mapped to the backlight module current parameter of the device, and the adjustment item for color correction is mapped to the color filter voltage parameter of the device. Based on the standardized adjustment items, the computer equipment derives a display parameter adjustment scheme that includes adjustment logic and execution details. The scheme clearly defines the adjustment basis, adjustment process, and expected effect of each adjustment item, and can directly adapt to the conversion requirements of display parameter adaptation adjustment instructions.

[0110] Step 166: Convert the adjustment scheme into a display parameter adaptation adjustment instruction to complete the output of the display parameter adaptation adjustment instruction. The display parameter adaptation adjustment instruction carries the identification information of the target optoelectronic display device and the execution priority of the adjustment scheme.

[0111] The computer equipment converts the display parameter adjustment scheme into display parameter adaptation and adjustment instructions. These instructions use a format recognizable by the target optoelectronic display device and include information such as the adjustment parameter type, adjustment value, and adjustment timing. The instructions carry the identification information of the target optoelectronic display device, ensuring accurate transmission; they also carry the execution priority of the adjustment scheme, guiding the target device to execute the adjustment operations according to priority, thus completing the output of the display parameter adaptation and adjustment instructions.

[0112] Optionally, based on steps 110-160 above, after step 160, the method further includes:

[0113] Step 210: Associate and map the display parameter adaptation and adjustment instructions with the effect path of device aging on visual perception, extract the parameter adjustment features corresponding to the effect path nodes in the display parameter adaptation and adjustment instructions, and start the semantic jump analysis of the effect path with the parameter adjustment features as semantic anchors.

[0114] The computer device maps display parameter adaptation and adjustment commands to the path of device aging's impact on visual perception, extracting parameter adjustment features corresponding to nodes in the path. For example, it associates the parameter features of backlight module current adjustment with pixel luminous efficiency attenuation nodes in the path, and the parameter features of color filter voltage adjustment with color filter attenuation nodes in the path. Using these parameter adjustment features as semantic anchors, the computer device initiates semantic jump analysis of the path, analyzing the semantic associations and extensions of the parameter adjustment features within the path.

[0115] Step 220: Semantically extend the feature transmission relationship of adjacent nodes in the action path through semantic jump rules to generate a multi-dimensional semantic jump path. The semantic jump rules are constructed based on the semantic correlation of feature attributes to realize the semantic extension from parameter adjustment features to deep features of the action path.

[0116] Based on semantic jump rules constructed using semantic associations of feature attributes, the computer device semantically extends the feature transmission relationships between adjacent nodes in the action path. For example, it jumps from a pixel luminous efficiency attenuation node to a driving circuit performance degradation node, analyzing the semantic association and feature transmission logic between the two; similarly, it jumps from a color filter attenuation node to a color reproduction feature deviation node, analyzing the semantic association and feature transmission logic between the two. Through multiple rounds of semantic extension, the computer device generates multi-dimensional semantic jump paths. These paths enable semantic extension from parameter adjustment features to deeper features of the action path, uncovering the potential impact of parameter adjustment on the action path.

[0117] Step 230: Perform convergence analysis on the generated multi-dimensional semantic jump paths, select the target semantic jump paths that match the current display parameter adjustment requirements, and construct a semantic association topology network of the action paths based on the target semantic jump paths.

[0118] The computer device performs convergence analysis on the generated multi-dimensional semantic jump paths, filtering out target semantic jump paths that match the current display parameter adjustment requirements. For example, when the adjustment requirement is brightness calibration, the device filters out semantic jump paths from backlight module current adjustment features to pixel luminous efficiency attenuation nodes, and then to brightness perception deviation nodes; when the adjustment requirement is color correction, the device filters out semantic jump paths from color filter voltage adjustment features to color filter attenuation nodes, and then to color saturation perception deviation nodes. Based on the target semantic jump paths, the computer device constructs a semantic association topology network of action paths. This topology network uses path nodes as vertices and semantic jump relationships as edges, clearly presenting the semantic association between parameter adjustment features and deep features of the action paths.

[0119] Step 240: By inferring the potential feature transmission links that have not been triggered in the action path through semantic association topology network, and combining the execution feedback information of the display parameter adaptation adjustment command, identify the influence trend of potential feature transmission links on image quality optimization effect.

[0120] Computer devices use semantic association topology networks to deduce potential feature transmission links in the action path that have not yet been triggered. For example, a link from a node where the performance of the driving circuit deteriorates to a node where the luminous efficiency of a pixel decays may not be triggered by the current parameter adjustment, but it may be triggered later during the device aging process. Combining the execution feedback information of display parameter adaptation adjustment instructions, such as the image quality improvement effect after parameter adjustment and changes in device operating status, the computer device identifies the trend of the impact of potential feature transmission links on the image quality optimization effect. For example, the triggering of such a link may cause the brightness perception deviation to increase again, affecting the long-term effect of image quality optimization.

[0121] Step 250: Based on the influence trend, generate the semantic inference result of the action path, and input the semantic inference result in reverse to the optimization task decision algorithm to realize the reverse adaptation of the action path and parameter adjustment, so that the output display parameter adaptation adjustment command adapts to the dynamic semantic changes of the action path.

[0122] The computer device generates semantic inference results of the action path based on the impact trend of the latent feature transmission chain on the image quality optimization effect. The results include the triggering conditions, impact degree, and corresponding suggestions for the latent feature transmission chain. The semantic inference results are then fed back into the optimization task decision algorithm. The algorithm adjusts its decision logic based on the inference results, such as adjusting the magnitude or timing of parameter adjustments in advance, to avoid the triggering of the latent feature transmission chain from negatively affecting the image quality optimization effect. This achieves reverse adaptation between the action path and parameter adjustment, ensuring that the output display parameter adaptation adjustment instructions can always adapt to the dynamic semantic changes of the action path.

[0123] Optionally, based on steps 110-160 above, after step 160, the method further includes:

[0124] Step 310: Extract the identification information of the target optoelectronic display device and the corresponding optimization direction features carried in the display parameter adaptation and adjustment instruction, and use the identification information and optimization direction features as source domain feature identifiers for transfer learning.

[0125] The computer device extracts the identification information of the target optoelectronic display device and the corresponding optimization direction features carried in the display parameter adaptation and adjustment instruction. The above information is used as the source domain feature identifier for transfer learning. The identifier includes information such as device type, current optimization direction, and aging degree, which is used for matching source domain data and target domain data in the subsequent transfer learning process.

[0126] Step 320: Collect historical display parameter adaptation and adjustment instructions and historical degradation scenario data corresponding to different types of optoelectronic display devices. Based on the source domain feature identifier, perform feature screening on the historical data to obtain a migration data set with similar optimization requirements to the target optoelectronic display device.

[0127] The computer equipment collects historical display parameter adaptation and adjustment instructions and historical degradation scenario data corresponding to different types of optoelectronic display devices from a historical database. Based on source domain feature identification, the historical data is feature-filtered to select historical data with similar optimization needs to the target optoelectronic display device, such as historical data of the same type, aging degree, and optimization direction as the target device. The selected historical data is integrated into a transfer dataset, which can provide effective data support for the transfer learning of the predictive event recognition model.

[0128] Step 330: Perform feature alignment processing on the transfer data set and the training dataset of the predicted event recognition model to ensure that the feature dimensions of the transfer data set are consistent with those of the training dataset, thereby generating a standardized transfer dataset.

[0129] The computer equipment performs feature alignment processing on the transfer dataset and the training dataset of the predictive event recognition model. This adjusts the feature dimensions of the transfer dataset to match those of the training dataset, for example, through feature interpolation or dimensionality reduction. Missing or redundant features in the transfer dataset are then removed, allowing the aligned data to be directly input into the predictive event recognition model. The aligned transfer dataset is then integrated into a standardized transfer dataset, with a format identical to the training dataset, enabling direct use for transfer learning training of the model.

[0130] Step 340: Based on the transfer learning task, input the standardized transfer dataset into the predicted event recognition model, and adjust the feature fusion layer parameters of the predicted event recognition model through the domain adaptation algorithm to reduce the distribution difference between the source domain features and the target domain features.

[0131] Computer equipment, based on transfer learning tasks, inputs standardized transfer datasets into a predictive event recognition model, introduces a domain adaptation algorithm, and adjusts the model's feature fusion layer parameters. By calculating the distribution difference between source domain features (features of the target optoelectronic display device) and target domain features (features in the transfer dataset), the fusion weights of the feature fusion layer are adjusted, enabling the model to better adapt to the feature distribution of the target domain data, reduce the distribution difference between source domain features and target domain features, and improve the model's ability to identify degradation scenarios of different types of devices.

[0132] Step 350: By analyzing the model's accuracy in identifying degraded scenarios in the standardized transfer dataset, the feature extraction capability of the prediction event recognition model after transfer learning is tested to determine the transfer adaptation effect.

[0133] The computer equipment analyzes the model's accuracy in identifying degradation scenarios in a standardized transfer dataset. It then tests the feature extraction capability of the prediction event recognition model after transfer learning, evaluating its accuracy, speed, and stability across different degradation scenarios. Based on the test results, the transfer learning effect is determined. If the recognition accuracy reaches a preset threshold, the transfer learning is considered effective, and the model can accurately identify degradation scenarios for different types of devices. If the recognition accuracy does not reach the preset threshold, the transfer learning effect is poor, requiring further adjustment of model parameters or an increase in the amount of transfer data.

[0134] Step 360: Generate model parameter adjustment instructions based on the migration adaptation effect, and integrate the model parameter adjustment instructions into the predicted event recognition model.

[0135] The computer device generates model parameter adjustment instructions based on the transfer adaptation effect. If the transfer adaptation effect is good, the instructions include fixed instructions for the model parameters, enabling the model to maintain good recognition ability during subsequent runs. If the transfer adaptation effect is poor, the instructions include suggestions for adjusting the feature fusion layer parameters, such as adjusting the calculation logic of the fusion weights or increasing the number of feature fusion layers. Integrating the model parameter adjustment instructions into the predictive event recognition model completes iterative optimization of the model, improving its generalization ability and recognition accuracy.

[0136] Optionally, based on steps 110-160 above, after step 160, the method further includes:

[0137] Step 410: Analyze the execution process data of the display parameter adaptation and adjustment instruction, and extract the decision logic elements related to the optimization task decision algorithm during the execution process. The decision logic elements include parameter adjustment priority sorting logic, feature matching logic, and instruction generation logic.

[0138] The computer equipment analyzes and displays the execution process data of parameter adaptation and adjustment instructions, including the instruction generation process, the parameter calculation process, and the execution timing determination process. It extracts the decision logic elements related to the optimization task decision algorithm, including parameter adjustment priority sorting logic (priority determination rules for different optimization tasks), feature matching logic (matching rules between real-time display status and deterioration scenarios), and instruction generation logic (conversion rules from adjustment scheme to instruction). These elements can reflect the core decision logic of the optimization task decision algorithm.

[0139] Step 420: Determine the robustness test task based on the decision logic elements, and introduce different types of interference feature sets into the robustness test task. The interference feature sets include device state fluctuation features, environmental interference features, and data transmission delay features, and the dimensions of the interference features match the feature dimensions of the real-time physical display state information.

[0140] Based on the extracted decision logic elements, the computer equipment determines a robustness test task. This task includes robustness testing of parameter adjustment priority ranking logic, feature matching logic, and instruction generation logic. Different types of interference feature sets are introduced into the test task, including device state fluctuation features (such as random fluctuations in device luminous efficiency), environmental interference features (such as sudden changes in operating temperature and humidity), and data transmission delay features (such as the transmission delay of real-time display status data). All interference features are dimensioned to match the feature dimensions of the real-time physical display status information, ensuring that the interference realistically simulates interference conditions in actual scenarios.

[0141] Step 430: Inject the interference feature set into the input of the optimization task decision algorithm to simulate the decision-making process of the algorithm under different interference scenarios, and record the deviation information between the display parameter adaptation adjustment command output by the algorithm and the command under the interference-free scenario.

[0142] The computer equipment injects interference feature sets into the input of the optimization task decision algorithm to simulate the algorithm's decision-making process under different interference scenarios. For example, it injects device state fluctuation features into real-time physical display state information to simulate random fluctuations in device operating status; it injects environmental interference features into working environment data to simulate sudden environmental changes; and it injects data transmission delay features into the data transmission process to simulate data transmission delays. It records the deviation information between the display parameter adaptation adjustment instructions output by the algorithm under interference scenarios and the instructions under non-interference scenarios, including deviations in adjustment parameters, adjustment timing, and trigger levels.

[0143] Step 440: Hierarchically classify the deviation information and locate the decision logic vulnerabilities that cause the deviation. Determine the algorithm optimization objectives based on the vulnerabilities. The algorithm optimization objectives include adaptive filtering of interference features, dynamic adjustment of decision logic, and improvement of the stability of instruction output.

[0144] The computer equipment categorizes deviation information into levels based on severity: mild, moderate, and severe deviations, with different processing priorities corresponding to different levels. By analyzing the correlation between deviation information and decision logic elements, the vulnerability of the decision logic causing the deviation is identified. For example, if the parameter adjustment priority ranking logic deviates under interference scenarios, it indicates that the logic's anti-interference capability is weak, making it a vulnerability. Based on these vulnerabilities, algorithm optimization objectives are determined, including adaptive filtering of interference features (automatically identifying and filtering interference features in the input data), dynamic adjustment of decision logic (adjusting decision rules in real time according to interference conditions), and improvement of instruction output stability (reducing instruction deviation under interference scenarios).

[0145] Step 450: Generate a robustness enhancement strategy based on the algorithm optimization objective, integrate the robustness enhancement strategy into the decision logic module of the optimization task decision algorithm, and optimize iteratively to ensure that the output deviation of the optimization task decision algorithm after injecting interference features is within a preset deviation range.

[0146] The computer equipment generates a robustness enhancement strategy based on the algorithm optimization objective. This strategy includes an adaptive filtering algorithm for interference features, dynamic adjustment rules for the decision logic, and a stability verification mechanism for instruction output. The robustness enhancement strategy is integrated into the decision logic module of the optimization task decision algorithm. For example, an interference feature filtering module is added before the feature matching logic, dynamic adjustment rules are added to the parameter adjustment priority ranking logic, and a stability verification mechanism is added after the instruction generation logic. Through iterative optimization, the algorithm's output deviation under interference scenarios is repeatedly tested, and the parameters of the robustness enhancement strategy are adjusted until the output deviation of the optimization task decision algorithm after the injection of interference features is within a preset deviation range, thus completing the robustness enhancement of the algorithm.

[0147] In an exemplary application scenario, a commercial display equipment manufacturer deploys the image quality optimization scheme of this invention on its OLED-8K organic light-emitting diode display device. Initially, a computer retrieves historical operating records from the OLED-8K's operating status database, including cumulative power-on time, pixel luminous efficiency change sequences, and ambient temperature and humidity fluctuation data. Using a 90-day sliding window, features such as luminous efficiency decay rate and driving circuit performance degradation rate are extracted for each period. Each feature is labeled with tags such as "usage period 0-90 days" and "ambient temperature and humidity range TH1," integrating them to form a device aging and decay feature set. Simultaneously, the computer obtains human visual perception data for the device under 12 display states, including "high brightness color gamut mode" and "low power eye protection mode," through a visual perception testing system. This data is converted into features such as brightness perception deviation and color saturation deviation, forming a visual perception change feature set. Multi-dimensional correlation analysis then yields sensory device linkage features such as "pixel luminous efficiency decay corresponding to brightness perception deviation."

[0148] Based on the interconnected characteristics of sensing devices, the computer device derives the action path of "degradation of driving circuit performance - attenuation of pixel luminous efficiency - increase in brightness perception deviation," thereby obtaining brightness distribution characteristics and color reproduction characteristics carrying aging impact coefficients. Combining cross-device universal adaptation rules, the computer device determines the optimization direction for brightness calibration and color correction for OLED-8K, generating parameter adjustment variables including "brightness adjustment step size D1 and adjustment period T1," and linking them with the adaptation rules of liquid crystal display devices and micro-LED display devices to form a linked optimization migration strategy covering multiple device types. Subsequently, the computer device builds a predictive event recognition model based on this strategy, inputting historical data from six scenarios such as "rapid brightness degradation" and "unstable color degradation" for training. The model identifies the correlation between aging characteristics and perception characteristics under different scenarios, generating a pre-judgment criterion of "triggering optimization when the brightness aging impact coefficient reaches X1."

[0149] When the real-time brightness aging impact coefficient of the OLED-8K reaches the trigger threshold, the computer generates a picture quality optimization trigger signal carrying a "severe trigger" flag, which is then input into the optimization task decision algorithm. The algorithm calls upon the device's real-time physical display status information, matches and determines the optimization task type of "brightness calibration + color correction," and generates real-time adjustment requirements by linking parameter adjustment variables in the migration strategy. Through decision logic deduction, it obtains an adjustment scheme including "backlight module current adjusted to Y1, color filter voltage adjusted to Y2," which is then converted into display parameter adaptation adjustment instructions and sent to the device. Simultaneously, the computer also uses a semantic association topology network to deduce potential "drive circuit aging cascading impact chains," inputting the deduction results back into the optimized instructions. Furthermore, it adapts to degradation scenarios of other display devices of the same brand through transfer learning, and withstands interference from sudden changes in environmental temperature and humidity, data transmission delays, etc., through robustness testing, ensuring the long-term effectiveness and stability of picture quality optimization.

[0150] The method described in this application breaks through the rigid thinking of "single-dimensional optimization" in the field of image quality optimization of existing optoelectronic display devices. By constructing a cross-dimensional linkage logic of "device aging-visual perception", it achieves a core transformation from passive defect repair to proactive predictive adaptation. It no longer focuses solely on isolated parameter adjustments based on human eye perception, nor does it perform single passive repairs for device defects. Instead, it extracts device aging and decay feature sets carrying usage cycle and working environment tags, and combines them with human eye visual perception change feature sets under different display states to complete multi-dimensional correlation analysis. This qualitatively uncovers the deep linkage patterns between two independent types of features, thereby clarifying the path of device aging on visual perception and deriving physical display characteristics. On this basis, a linkage optimization migration strategy covering multiple types of optoelectronic display devices is constructed through cross-device universal adaptation rules, breaking the limitations of previous single-device-specific optimization and realizing the underlying refinement of cross-device universal optimization logic. Meanwhile, based on the predictive event recognition model, the correlation patterns under different degradation scenarios are identified to generate preliminary judgment criteria, and the triggering logic and decision-making process of image quality optimization are reconstructed. This fundamentally improves the accuracy, foresight and universality of image quality optimization, and solves the problems of poor adaptability, optimization lag and low cross-device reusability caused by single-dimensional optimization in existing technologies.

[0151] Please see Figure 2 The figure is a schematic diagram of the basic structure of a computer device 200 provided in an embodiment of this application. The computer device 200 includes: a processor 201; a storage device 202 on which a computer program 2020 is stored; and a network interface 203 for providing network communication functions. When the computer program 2020 is executed by the processor 201, the processor 201 implements any of the image quality optimization methods applied to optoelectronic device displays.

[0152] Please see Figure 3 This application provides a functional block diagram of an image quality optimization device, which includes:

[0153] The device feature extraction module is used to extract the device aging and degradation feature set of the target optoelectronic display device. Each device aging and degradation feature in the device aging and degradation feature set carries the usage period label and working environment label corresponding to the target optoelectronic display device. At least one type of label is different in the usage period label and working environment label of different device aging and degradation features.

[0154] The multidimensional correlation analysis module is used to obtain the human eye visual perception change feature set under different display states, and to perform multidimensional correlation analysis on the human eye visual perception change feature set and the device aging and decay feature set to obtain the sensor device linkage feature.

[0155] The display feature derivation module is used to determine the effect path of device aging on visual perception based on the linkage characteristics of the sensing device, and derive the physical display features corresponding to the target optoelectronic display device through the effect path.

[0156] The optimization strategy generation module is used to combine the physical display characteristics and the preset cross-device universal adaptation rules to generate a linkage optimization migration strategy that includes optimization directions and parameter adjustment variables adapted to different device types.

[0157] The judgment basis generation module is used to create a predictive event recognition model covering different degradation scenarios using the linkage optimization migration strategy. The predictive event recognition model is used to identify the correlation between device aging and decay characteristics and visual perception change characteristics under different degradation scenarios, and generate the pre-judgment basis for triggering image quality optimization.

[0158] The adjustment instruction output module is used to generate an image quality optimization trigger signal based on the pre-judgment criteria and input it into the optimization task decision algorithm. The optimization task decision algorithm then outputs a display parameter adaptation adjustment instruction in combination with the real-time physical display status information of the target optoelectronic display device.

[0159] Based on the above, a readable storage medium is provided, on which a program or instructions are stored, and when the program or instructions are executed by a processor, the steps of the above method are implemented.

[0160] Furthermore, it should be noted that this application also provides a computer program product, which may include a computer program that can be stored in a computer-readable storage medium. The processor of a computer device reads the computer program from the computer-readable storage medium, and the processor can execute the computer program, causing the computer device to perform the aforementioned... Figure 1 The methods described in the corresponding embodiments are already known, and therefore will not be repeated here. Furthermore, the beneficial effects of using the same method will also not be repeated. For technical details not disclosed in the computer program product embodiments related to this application, please refer to the description of the method embodiments of this application.

[0161] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems or apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and relevant parts can be referred to the method section.

Claims

1. A method for optimizing image quality in optoelectronic device displays, characterized in that, The method includes: Extract the device aging and degradation feature set of the target optoelectronic display device. Each device aging and degradation feature in the device aging and degradation feature set carries the usage period tag and working environment tag corresponding to the target optoelectronic display device. At least one type of tag is different in the usage period tag and working environment tag of different device aging and degradation features. A set of human visual perception change features under different display states is obtained, and a multi-dimensional correlation analysis is performed on the set of human visual perception change features and the set of device aging and decay features to obtain the linkage features of the sensing device. Based on the linkage characteristics of the sensing device, the path of the effect of device aging on visual perception is determined, and the physical display characteristics corresponding to the target optoelectronic display device are derived through the path of the effect. Combining the physical display characteristics and the preset cross-device universality adaptation rules, a linkage optimization migration strategy is generated, which includes optimization directions and parameter adjustment variables for adapting to different device types. The aforementioned linkage optimization migration strategy is used to create a predictive event recognition model covering different degradation scenarios. The predictive event recognition model is used to identify the correlation between device aging and decay characteristics and visual perception change characteristics under different degradation scenarios, and generate a preliminary judgment basis for triggering image quality optimization. Based on the aforementioned pre-judgment criteria, an image quality optimization trigger signal is generated and input into the optimization task decision algorithm. The optimization task decision algorithm, combined with the real-time physical display status information of the target optoelectronic display device, outputs a display parameter adaptation and adjustment command. The process combines the physical display characteristics and preset cross-device universality adaptation rules to generate a linked optimization migration strategy that includes optimization directions and parameter adjustment variables adapted to different device types, including: Obtain preset cross-device universal adaptation rules, parse the adaptation conditions and parameter constraint ranges of different device types contained in the cross-device universal adaptation rules, and establish a correspondence library between adaptation rules and device types; The physical display features are processed by feature encoding to obtain standardized physical display feature codes. The physical display feature codes are matched one by one with the adaptation conditions in the corresponding relational database to determine the basic adaptation rules corresponding to the target optoelectronic display device. Based on the parameter constraint range in the basic adaptation rules and combined with the aging effect coefficient in the physical display characteristics, the image quality optimization direction for different device types is determined. The optimization direction covers brightness calibration, color correction and contrast adjustment. For each optimization direction, a parameter adjustment model is constructed. The physical display characteristics and the corresponding aging influence coefficient are input into the parameter adjustment model. The parameter adjustment variables under each optimization direction are obtained through model calculation. The parameter adjustment variables include adjustment range, adjustment step size and adjustment period. By combining the optimization directions for different device types and the corresponding parameter adjustment variables, a structured linkage optimization migration strategy is generated. The linkage optimization migration strategy also includes parameter migration logic between different device types and priority ranking of optimization directions.

2. The method as described in claim 1, characterized in that, The step of determining the effect path of device aging on visual perception based on the linkage characteristics of the sensing device, and deriving the physical display characteristics corresponding to the target optoelectronic display device through the effect path, includes: The linkage features of the sensing device are decomposed into layers to obtain a first dimension feature related to device aging and a second dimension feature related to visual perception. A mapping relationship matrix between the first dimension feature and the second dimension feature is established, and feature combinations with causal relationship are selected based on the mapping relationship matrix. An initial action path network is generated using the feature combination as the target path node, and the path nodes in the initial action path network are subjected to correlation enhancement processing to obtain the action path of device aging on visual perception. A feature transfer model is created based on the feature transfer weights of each path node in the action path. The device aging and decay feature set and the human visual perception change feature set are input into the feature transfer model. Intermediate feature parameters in the feature transfer process are obtained through iterative calculation of the feature transfer model. Based on the correspondence between the intermediate feature parameters and the basic display parameters of the target optoelectronic display device, the physical display features corresponding to the target optoelectronic display device are derived. The physical display features include brightness distribution features, color reproduction features, and contrast variation features, and each feature carries a corresponding aging influence coefficient.

3. The method as described in claim 1, characterized in that, The aforementioned linkage optimization migration strategy is used to create a predictive event recognition model covering different degradation scenarios. This predictive event recognition model identifies the correlation between device aging and degradation characteristics and visual perception changes under different degradation scenarios, generating pre-judgment criteria for triggering image quality optimization, including: Based on the optimization direction and parameter adjustment variables in the aforementioned linkage optimization migration strategy, various device degradation scenarios are obtained, and the scenario feature identifier corresponding to each degradation scenario is extracted. Collect historical device aging and decay feature sets and historical human visual perception change feature sets under different degradation scenarios, and associate the historical device aging and decay feature sets and historical human visual perception change feature sets with corresponding scene feature identifiers to generate a judgment event training dataset. Based on the parameter adjustment logic in the aforementioned linkage optimization migration strategy, a model network for predictive event recognition is constructed. The model network includes a feature input layer, a feature fusion layer, and a pattern output layer. The predicted event training dataset is input into the predicted event recognition model for training. By adjusting the network parameters of the predicted event recognition model, the deviation between the correlation pattern output by the predicted event recognition model and the correlation pattern in the actual scene is within a preset range, thus completing the creation of the predicted event recognition model. The real-time device aging and decay feature sets and the real-time human visual perception change feature sets under different degradation scenarios are input into the trained prediction event recognition model. The real-time device aging and decay feature sets and the real-time human visual perception change feature sets are fused through the feature fusion layer. The correlation rules between device aging and decay features and visual perception change features under different degradation scenarios are output through the rule output layer. Based on the aforementioned correlation patterns, a threshold for triggering image quality optimization is set. Based on the threshold for triggering the threshold and the corresponding correlation pattern features, a preliminary judgment basis for triggering image quality optimization is generated.

4. The method as described in claim 3, characterized in that, The process involves fusing the real-time device aging degradation feature set and the real-time human visual perception change feature set through the feature fusion layer, and outputting the correlation rules between device aging degradation features and visual perception change features under different degradation scenarios through the rule output layer, including: After receiving the real-time device aging and decay feature set and the real-time human visual perception change feature set through the feature fusion layer, the feature dimension of the real-time device aging and decay feature set is adjusted so that the dimension of each feature in the real-time device aging and decay feature set matches the dimension of the corresponding feature in the real-time human visual perception change feature set, thus obtaining two feature sets with unified dimensions. Based on the preset feature association rules in the predicted event recognition model, feature association mapping is performed on two types of feature sets with uniform dimensions to establish a one-to-one relationship between the aging and decay characteristics of each real-time device and the corresponding real-time human visual perception change characteristics. The feature fusion algorithm built into the feature fusion layer is used to fuse features with one-to-one relationships to generate a fused feature set. The feature fusion algorithm dynamically adjusts the fusion weights based on the correlation strength between the two types of features. Feature combinations with high correlation strength correspond to higher fusion weights. The fused feature set is subjected to feature denoising processing to obtain a purified fused feature set. The purified fused feature set is then passed to the regularity output layer. The regularity output layer analyzes the feature attributes and feature correlations of each fused feature in the purified fused feature set. Based on the feature attributes and feature correlations, multiple feature correlation clusters are obtained, and each feature correlation cluster corresponds to a feature correlation type under a deterioration scenario. For each feature association cluster, the association pattern identification process is performed to determine the dynamic association trend between the real-time device aging and decay characteristics and the real-time human visual perception change characteristics within the same feature association cluster. The dynamic association trend is then bound to the scene in combination with the corresponding scene feature identifier to obtain the association pattern fragment of the bound scene information. By integrating the association pattern fragments of all bound scenario information, association patterns covering different degradation scenarios are obtained. The association patterns include the initiation conditions of feature association, feature transmission paths, and collaborative relationships of feature changes under each degradation scenario.

5. The method as described in claim 1, characterized in that, The process of generating a picture quality optimization trigger signal based on the aforementioned pre-judgment criteria and inputting it into the optimization task decision algorithm, followed by the optimization task decision algorithm combining the real-time physical display status information of the target optoelectronic display device to output a display parameter adaptation and adjustment command, includes: The trigger condition threshold and correlation pattern features in the pre-judgment criteria are analyzed, the correlation pattern features are quantized to obtain feature quantization values, and the feature quantization values ​​are compared with the trigger condition threshold. When the feature quantization value reaches the trigger condition threshold, a corresponding image quality optimization trigger signal is generated. The image quality optimization trigger signal carries a trigger level identifier and the corresponding degraded scene features. The image quality optimization trigger signal is input to a preset optimization task decision algorithm. The optimization task decision algorithm calls the real-time physical display status information of the target optoelectronic display device and performs matching analysis between the real-time physical display status information and the degradation scene features in the image quality optimization trigger signal to determine the optimization task type corresponding to the current device display status. Based on the optimization task type, the corresponding optimization direction and parameter adjustment variables in the linkage optimization migration strategy are retrieved, and the real-time physical display status information is input into the calculation model corresponding to the parameter adjustment variables to obtain the real-time parameter adjustment requirements. Based on the real-time parameter adjustment requirements and the decision logic in the optimized task decision algorithm, an adjustment scheme for the display parameters is derived, which includes the adjustment parameter type, adjustment value, and adjustment timing. The adjustment scheme is converted into a display parameter adaptation adjustment instruction to complete the output of the display parameter adaptation adjustment instruction. The display parameter adaptation adjustment instruction carries the identification information of the target optoelectronic display device and the execution priority of the adjustment scheme.

6. The method as described in claim 5, characterized in that, Based on the optimization task type, the corresponding optimization direction and parameter adjustment variables in the linkage optimization migration strategy are retrieved, and the real-time physical display status information is input into the computational model corresponding to the parameter adjustment variables to obtain the real-time parameter adjustment requirements, including: Based on the optimized task type, a mapping knowledge graph between the task type and the optimization direction is determined. The optimization direction and corresponding parameter adjustment variables that match the current optimized task type are retrieved from the linkage optimization migration strategy through the mapping knowledge graph. The parameter adjustment variables carry parameter attribute identifiers that are adapted to the optimization direction. The real-time physical display state information of the target optoelectronic display device is analyzed for features, and the state features related to the retrieved optimization direction are extracted from the real-time physical display state information to obtain the target state feature set. The target state feature set includes state features corresponding to the brightness calibration direction, color correction direction, or contrast adjustment direction. Based on the parameter attribute identifier of the parameter adjustment variable, establish the attribute matching relationship between the target state feature set and the parameter adjustment variable, so that each target state feature corresponds to a parameter adjustment variable with the same attribute identifier; The target state feature set and parameter adjustment variables after establishing attribute matching relationship are input into the preset operation model. The operation model performs feature difference analysis based on the current feature value of the target state feature and the preset standard state feature value in the linkage optimization migration strategy to obtain feature difference information. Based on the feature difference information, the initial value of the parameter adjustment variable is adjusted so that the adjusted parameter adjustment variable is adapted to the feature difference information. Through iterative computation of the computational model, the adjusted parameter adjustment variables are dynamically adapted to the target state features so that the values ​​of the parameter adjustment variables match the differences in the target state features. Based on the values ​​of the parameter adjustment variables after iterative calculation and the corresponding feature difference information, a real-time parameter adjustment requirement is generated. The real-time parameter adjustment requirement includes the parameter adjustment type corresponding to the optimization direction, the range boundary of the parameter adjustment, and the dynamic adaptation rules of the parameter adjustment. The dynamic adaptation rules are used to guide the parameter adjustment process to make adaptive adjustments according to the changes in the real-time physical display state.

7. The method as described in claim 5, characterized in that, The adjustment scheme for the display parameters is derived by combining the real-time parameter adjustment requirements and the decision logic in the optimized task decision algorithm, including: The real-time parameter adjustment requirements and the decision logic in the optimization task decision algorithm are analyzed. The parameter adjustment types, parameter adjustment range boundaries and dynamic adaptation rules contained in the real-time parameter adjustment requirements are extracted. The task priority determination rules, parameter adjustment conflict resolution rules and adjustment scheme testing rules covered in the decision logic are decomposed. A correlation derivation model between the core elements of the real-time parameter adjustment requirements and the rule elements of the decision logic is established so that each parameter adjustment type corresponds to a rule entry with the same task attribute in the decision logic. Based on the established correlation derivation model, the optimization direction priority ranking and parameter migration logic corresponding to the current optimization task type in the linkage optimization migration strategy are retrieved. The parameter adjustment range boundary in the real-time parameter adjustment requirement is compared with the preset standard parameter range in the linkage optimization migration strategy to determine the operable range of parameter adjustment. According to the task priority judgment rule in the decision logic, the multiple parameter adjustment types involved in the real-time parameter adjustment requirement are prioritized to obtain the parameter adjustment execution sequence. The real-time physical display status information of the target optoelectronic display device is continuously input into the correlation derivation model. The parameter adjustment conflict resolution rules in the decision logic are used to coordinate the adjustment types with parameter adjustment conflicts in the execution sequence. Based on the dynamic adaptation rules and the changing trend of the real-time physical display status information, the execution sequence of parameter adjustment and the corresponding adjustment range boundary are dynamically corrected so that the corrected execution sequence is adapted to the changing requirements of the real-time physical display status. Based on the revised parameter adjustment execution sequence and adjustment range boundary, an initial task description of the adjustment scheme is constructed. The adjustment parameter type, adjustment value and adjustment sequence in the initial task description are analyzed by the adjustment scheme test rules of the decision logic. If the initial task description fails the test, the process returns to the correction step of the parameter adjustment execution sequence. The execution sequence and adjustment range boundary are revised again based on the evaluation feedback information until the initial task description passes the test. By utilizing the adjustment values, timing, and execution priority corresponding to each parameter adjustment type, a structured integration operation is performed on the initial task description that has passed the test. During the integration process, the identification information of the target optoelectronic display device is synchronously associated, so that the standardized adjustment entries formed after integration correspond one-to-one with the specific display parameters of the target optoelectronic display device. Based on the standardized adjustment entries, a display parameter adjustment scheme containing adjustment logic and execution details is derived. The display parameter adjustment scheme is used to adapt to the conversion requirements of display parameter adaptation adjustment instructions.

8. A computer device, characterized in that, include: A processor; a storage device having a computer program stored thereon; a network interface for providing network communication functions; when the computer program is executed by the processor, the processor enables the processor to implement the image quality optimization method for optoelectronic device displays as described in any one of claims 1-7.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the image quality optimization method for optoelectronic device displays as described in any one of claims 1-7.

Images