Intelligent definition detection method and system for mobile phone display screen

Abstract

The invention belongs to the field of definition detection, and discloses an intelligent definition detection method and system for a mobile phone display screen, and the method comprises the steps: S1, generating a detection strategy based on the content displayed by the display screen; s2, acquiring a monitoring area based on the detection strategy, and calculating a local definition score of the monitoring area; s3, calculating a local physical loss score based on the local definition score; s4, calculating a comprehensive definition score based on the local definition score and the local physical loss score; s5, generating a definition evaluation result and a personalized optimization suggestion based on the comprehensive definition score; and S6, updating the detection strategy. According to the display screen definition detection method, intelligent and comprehensive display screen definition detection is achieved by integrating detection of dynamic content, pixel point health and screen physical state, the defects of detection singleness and precision in the prior art are overcome, and it is ensured that a definition detection result can truly reflect the actual visual experience of a user.

Description

Technical Field

[0001] The present invention relates to the field of clarity detection, and in particular to an intelligent clarity detection method and system for a mobile phone display screen. Background Art

[0002] In modern smartphones and other mobile devices, the display screen is one of the core components of the user experience, and the clarity of the screen directly affects the user's visual experience. With the popularization of high-resolution display technology and the widespread application of new display screens such as OLED and AMOLED, the detection of screen clarity has become an important part of equipment quality management. At present, the detection of the clarity of mobile phone displays mainly relies on two traditional technologies: one is the detection based on image processing algorithms, and the other is the parameter monitoring based on display hardware. The detection method based on image processing algorithms usually judges the clarity by analyzing the content on the screen (such as images, text, videos, etc.). Such methods often use image algorithms such as edge detection and sharpness detection. However, these methods perform poorly under dynamic content (such as videos) and are difficult to adapt to the complexity of content changes. Especially when the user is actually operating, the detection effect is often inconsistent with the user's subjective perception, and it is easy to over-detect or miss detection. On the other hand, the detection method based on the display hardware parameters mainly infers the change of clarity by monitoring the brightness, pixel status or screen response time of the screen. Although these methods perform well in static environments, they lack the ability to deal with external factors such as physical damage to the screen (such as scratches and stains), and it is difficult to accurately reflect the physical changes of the screen in long-term use. At the same time, existing technologies rarely consider the impact of the user's actual usage environment (such as ambient brightness, the angle and distance of the phone, etc.) on display clarity. The intensity of ambient light, scratches on the screen surface, or pixel aging problems will directly affect the user's visual experience, but these factors are often ignored in existing detection systems, resulting in a large deviation between the detection results and the actual experience.

[0003] There is still a big problem with the existing clarity detection technology, that is, its detection mechanism is often single, either focusing on visual content or relying only on the hardware parameters of the display screen, and failing to achieve a comprehensive analysis of multi-dimensional factors. For example, when there are scratches or stains on the screen surface, users may perceive a decrease in clarity, but the current detection technology based on image processing cannot accurately identify these physical losses; and when the pixels inside the screen are aging and the brightness is uneven, image processing technology is difficult to effectively perceive these hardware-level problems. This makes the detection results one-sided and it is difficult to provide users with accurate clarity feedback. In addition, under long-term use or in harsh environments (such as high brightness, low light, etc.), the physical state of the screen and changes in pixel behavior will further affect the display effect. The existing detection methods fail to achieve real-time adjustment and feedback, and lack intelligent adaptive detection mechanisms. Therefore, developing an intelligent clarity detection system that can integrate screen dynamic content, pixel state, and physical loss factors to overcome the above defects has become the key to improving user experience. Summary of the invention

[0004] The purpose of the present invention is to disclose an intelligent clarity detection method and system for a mobile phone display screen, so as to solve the technical problems raised in the background technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] In one aspect, the present invention provides an intelligent definition detection method for a mobile phone display screen, comprising:

[0007] S1, generating a detection strategy based on the content displayed on the display screen;

[0008] S2, obtains the monitoring area based on the detection strategy and calculates the local clarity score of the monitoring area;

[0009] S3, calculating a local physical loss score based on the local clarity score;

[0010] S4, calculating a comprehensive clarity score based on the local clarity score and the local physical loss score;

[0011] S5, generates clarity assessment results and personalized optimization suggestions based on the comprehensive clarity score;

[0012] S6, update detection strategy;

[0013] Among them, S1 includes:

[0014] S11, obtaining the image frame Frame of the content displayed on the mobile phone display t :

[0015] Frame t =[P uv|u∈[1,H],v∈[1,W]] (1)

[0016] t represents the time step; P uv Represents the RGB value of the pixel at the uth row and vth column; H and W represent the height and width of the display screen respectively;

[0017] S12, through the convolutional neural network uv Mapped to a high-dimensional feature vector F t ;

[0018] S13, using principal component analysis to analyze F t Perform dimensionality reduction to obtain the reduced eigenvector F′ t :

[0019] F′ t = PCA(F t ) (2)

[0020] F t Represents a high-dimensional feature vector extracted by a convolutional neural network; PCA represents the principal component analysis dimensionality reduction method;

[0021] S14, use the classification model to classify F′ t Perform classification and obtain the classification results of the content displayed on the display screen:

[0022]

[0023] It is a classification model based on multi-layer perceptron structure;

[0024] C t Indicates the classification result, and the value range is {1,2,3}, where 1, 2, and 3 correspond to the classification results of text, picture, and video respectively;

[0025] S15, generate detection strategy:

[0026]

[0027] Indicates the detection strategy, Freq indicates the detection frequency, Precision indicates the detection accuracy, and Area indicates the detection area.

[0028] On the other hand, the present invention provides an intelligent definition detection system for a mobile phone display screen, comprising a detection strategy generation module, a first calculation module, a second calculation module, a third calculation module, a result generation module and an update module;

[0029] The detection strategy generation module is used to generate a detection strategy based on the content displayed on the display screen;

[0030] The first calculation module is used to obtain the monitoring area based on the detection strategy and calculate the local clarity score of the monitoring area;

[0031] The second calculation module is used to calculate the local physical loss score based on the local clarity score;

[0032] The third calculation module is used to calculate the comprehensive clarity score based on the local clarity score and the local physical loss score;

[0033] The result generation module is used to generate clarity evaluation results and personalized optimization suggestions based on the comprehensive clarity score;

[0034] The update module is used to update the detection strategy;

[0035] Among them, the monitoring area is obtained based on the detection strategy, and the local clarity score of the monitoring area is calculated, including:

[0036] S11, obtaining the image frame Frame of the content displayed on the mobile phone display t :

[0037] Frame t =[P uv |u∈[1,H],v∈[1,W]] (1)

[0038] t represents the time step; P uv Represents the RGB value of the pixel at the uth row and vth column; H and W represent the height and width of the display screen respectively;

[0039] S12, through the convolutional neural network uv Mapped to a high-dimensional feature vector F t ;

[0040] S13, using principal component analysis to analyze F t Perform dimensionality reduction to obtain the reduced eigenvector F′ t :

[0041] F′ t = PCA(F t ) (2)

[0042] F t Represents a high-dimensional feature vector extracted by a convolutional neural network; PCA represents the principal component analysis dimensionality reduction method;

[0043] S14, use the classification model to classify F′ t Perform classification and obtain the classification results of the content displayed on the display screen:

[0044]

[0045] It is a classification model based on multi-layer perceptron structure;

[0046] C t Indicates the classification result, and the value range is {1,2,3}, where 1, 2, and 3 correspond to the classification results of text, picture, and video respectively;

[0047] S15, generate detection strategy:

[0048]

[0049] Indicates the detection strategy, Freq indicates the detection frequency, Precision indicates the detection accuracy, and Area indicates the detection area.

[0050] Beneficial effects:

[0051] First, in view of the shortcomings of existing image processing methods in processing dynamic content, the present invention introduces a dynamic content analysis scheme, which can identify the type of displayed content on the screen (such as text, pictures or videos) in real time and adaptively adjust the detection strategy. For example, for highly dynamic content such as videos, the present invention reduces the detection frequency to save resources, and when processing text and static pictures, it improves the detection accuracy, focuses on analyzing the areas of user concern, and avoids indiscriminate high-intensity detection of full-screen content, thereby solving the problems of low detection efficiency and waste of resources in the prior art. In this way, the present invention can provide clarity feedback that meets the user's visual experience according to the changes in the actual content, and optimizes the accuracy and real-time performance of the detection results.

[0052] Secondly, in order to solve the impact of screen hardware aging, pixel brightness attenuation, etc. on clarity, the present invention introduces a pixel behavior tracking solution to monitor the brightness and response time of each pixel on the screen in real time, and identify the problem of decreased clarity caused by pixel aging or overuse. This module can not only track the performance of pixels over a long period of time, but also predict possible display problems and provide early warnings based on changes in pixel status, making up for the lack of pixel health status detection in the prior art. Through pixel tracking, the present invention can effectively identify physical problems inside the screen and provide more refined clarity detection.

[0053] Finally, in response to the impact of physical loss on the screen surface (such as scratches, stains, etc.) on clarity, the present invention proposes a surface state monitoring solution that detects changes in the physical state of the display surface by analyzing the optical reflection characteristics of the screen. Unlike traditional image processing algorithms, this module can accurately detect subtle scratches or stains that are difficult to detect with the naked eye through micro-reflection analysis, and evaluate the actual impact of these physical problems on clarity, solving the deficiency of ignoring the problem of physical loss in the prior art. By incorporating the surface state of the screen into clarity detection, the present invention provides a comprehensive detection system that can comprehensively evaluate the clarity of the display screen from multiple dimensions including visual content, screen hardware status, and physical loss.

[0054] In summary, the present invention realizes an intelligent and comprehensive display screen clarity detection system by integrating the detection of dynamic content, pixel health and screen physical state, overcomes the shortcomings of the existing technology in detection uniformity and accuracy, and ensures that the clarity detection results can truly reflect the user's actual visual experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0056] Figure 1 The present invention is a schematic diagram of an intelligent clarity detection method for a mobile phone display screen.

[0057] Figure 2 The present invention is a schematic diagram of an intelligent clarity detection system for a mobile phone display screen. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0059] The present invention proposes an intelligent display screen clarity detection system based on dynamic content analysis, pixel behavior tracking and surface state monitoring, aiming to solve several key problems in current technology.

[0060] like Figure 1 In one embodiment shown, the present invention provides an intelligent definition detection method for a mobile phone display screen, comprising:

[0061] S1, generating a detection strategy based on the content displayed on the display screen;

[0062] S2, obtains the monitoring area based on the detection strategy and calculates the local clarity score of the monitoring area;

[0063] S3, calculating a local physical loss score based on the local clarity score;

[0064] S4, calculating a comprehensive clarity score based on the local clarity score and the local physical loss score;

[0065] S5, generates clarity assessment results and personalized optimization suggestions based on the comprehensive clarity score;

[0066] S6, update detection strategy;

[0067] Among them, S1 includes:

[0068] S11, obtaining the image frame Frame of the content displayed on the mobile phone display t :

[0069] Frame t =[P uv |u∈[1,H],v∈[1,W]] (1)

[0070] t represents the time step; P uv represents the RGB value of the pixel at the uth row and the vth column; H and W represent the height and width of the display screen respectively; the content displayed on the display screen is provided by the mobile phone screen capture module in the form of image frames;

[0071] S12, through the convolutional neural network uv Mapped to a high-dimensional feature vector F t ;

[0072] In order to avoid full-size operations on the entire image, we first need to t Extract effective features for classification. Based on the complexity of the screen content, the above mapping is performed;

[0073] S13, using principal component analysis to analyze F t Perform dimensionality reduction to obtain the reduced eigenvector F′ t :

[0074]

[0075] F t Represents a high-dimensional feature vector extracted by a convolutional neural network; PCA represents the principal component analysis dimensionality reduction method;

[0076] S13 can reduce the computational complexity;

[0077] S14, use the classification model to classify F′ t Perform classification and obtain the classification results of the content displayed on the display screen:

[0078]

[0079] It is a classification model based on multi-layer perceptron structure;

[0080] C t Indicates the classification result, and the value range is {1,2,3}, where 1, 2, and 3 correspond to the classification results of text, picture, and video respectively;

[0081] In view of the diversity of screen content (such as text, pictures, videos, etc.), classification is needed to generate adaptive detection strategies for different content types.

[0082] S15, generate detection strategy:

[0083]

[0084] Indicates the detection strategy, Freq indicates the detection frequency, Precision indicates the detection accuracy, and Area indicates the detection area.

[0085] Detection frequency Freq: For dynamic content (such as video), the detection frequency should be reduced, while for static content (such as text, pictures), the frequency should be appropriately increased.

[0086] Detection accuracy: Text requires high-precision detection, while videos and pictures have relatively low detection accuracy requirements.

[0087] Detection area: For text or pictures, the focus is on detecting the area where the user is likely to look, while video content is evenly distributed.

[0088] Freq with classification label C t Changes are determined by the empirical formula, Freq = α·exp(-C t ), α is an adjustment coefficient used to match the dynamics of the content (e.g., lower the frequency for video and keep high frequency for text).

[0089] Precision is also based on C t , using a similar formula Precision = β·C t , β is the classification adjustment coefficient.

[0090] Area is based on C t , determine the key detection area for different types. For example, for text, you can select the key area in the middle of the screen for detection, while for pictures and videos, you can evenly detect the entire screen.

[0091] The generation of detection strategies effectively ensures the adaptability and targeting of clarity detection.

[0092] Preferably, S2 includes:

[0093] S21, initialize the monitoring area of ​​the pixel:

[0094] P Area ={(i,j)|(i,j)∈Area} (5)

[0095] P Area Represents the set of coordinates of the pixel points in the Area area. Area is composed of The detection area provided can be the area on the screen that displays text, images or dynamic content, rather than the entire screen.

[0096] (i,j) represents the coordinates of the pixel.

[0097] P Area The area corresponding to the coordinates in is considered to be the screen area that users pay most attention to and is also the focus of screen clarity detection.

[0098] The purpose of this step is to avoid the waste of computing resources caused by full-screen detection by locally monitoring the pixels in the key area, and to ensure that the clarity detection accurately focuses on the user's visual area.

[0099] S22, generates calculation function for brightness and response time:

[0100] B ij (t) = B ij (0)·exp(-γ ij ·t)(6)

[0101] R ij (t) = R ij (0)+δ ij ·t(7)

[0102] B ij (0) and R ij (0) represents the brightness and response time of the pixel with coordinates (i, j) in the initial state; γij is the brightness attenuation coefficient; as time goes by, the brightness of the pixel will gradually decrease. It is related to the aging speed of the pixel. The larger the γ ij Indicates that the pixels age quickly.

[0103] δ ij It is the rate at which the response time increases; the longer the pixel response time, the more obvious the delay in displaying the content, and the more it affects the clarity.

[0104] In order to evaluate the health of screen pixels, we focus on monitoring the changes in pixel brightness and response time. Over time, pixels may become less bright or have longer response times due to factors such as aging and usage frequency, which will directly affect the clarity of the screen. Therefore, a calculation function for brightness and response time is generated.

[0105] Through these two calculation functions, the present invention can capture the changes of pixel points in real time, especially the problems of aging, uneven brightness, etc., thereby providing a basis for subsequent clarity detection.

[0106] S23, calculate the pixel health status value:

[0107]

[0108] H ij (t) represents the health status value of the pixel with coordinates (i, j); α2 and β2 are the influence weights of brightness and response time on health status respectively; λ2 is The influence weight of is used to penalize pixels with large brightness fluctuations; Indicates the fluctuation of pixel brightness; ΔB ij Represents the brightness change value of the pixel with coordinates (i, j); compared with the general health status assessment method, the present invention innovatively introduces dynamic brightness change, which is particularly suitable for switching between different scenes on the display screen. If the brightness of the pixel fluctuates greatly, it means that the clarity of the area may be unstable.

[0109] H ij (t) can quantify the health of the pixel. ij When (t)>1, the pixel P ij The state is considered healthy; when H ij When (t)<1, the pixel may have brightness attenuation or response delay problems, which may affect the clarity.

[0110] S24, calculate the local clarity score Clarity local :

[0111]

[0112] Wij is the weight of the pixel with coordinates (i, j), indicating the importance of the pixel with coordinates (i, j) in the monitoring area. The present invention assigns different weights to pixels based on user sight hotspots (such as text display areas) and content types to ensure that the clarity of the area that the user is concerned about is prioritized.

[0113] For the entire monitoring area P Area Based on the health status of each pixel, the present invention defines a local clarity score Clarity local This score is a weighted average of the health of all pixels, used to quantify the overall display clarity of the area.

[0114] A higher local clarity score indicates better clarity within the area, while a low score indicates blurriness, uneven brightness, and other issues in the area.

[0115] Preferably, S3 includes:

[0116] S31, determine the key area A for physical loss detection damage :

[0117] A damage ={(i,j)∈P Area ∣Clarity local <θ} (10)

[0118] A damage It is the key area for physical loss detection; θ is the clarity threshold (usually set to 0.85≤θ≤0.95, which can be adjusted through experiments). Areas below this threshold usually have physical losses.

[0119] S31 is able to locate low-definition areas. Usually, physical wear such as scratches, stains, etc. can significantly affect the clarity of certain areas, especially when the clarity is below a certain threshold θ.

[0120] By automatically detecting areas with poor clarity, the present invention ensures that it only detects potential physical problem areas, thereby optimizing detection efficiency.

[0121] S32, respectively calculating the reflected light deviation ΔR of each pixel point ij (t):

[0122]

[0123] R ij (t) is the actual reflected light intensity of the pixel with coordinates (i, j) at time t; is the ideal reflected light intensity of the pixel point with coordinates (i, j) in an undamaged state (the standard reflected light intensity when there is no physical loss); the reflected light intensity of the screen surface is collected by an optical sensor; by comparing the current reflected light intensity with the standard reflected light intensity, the present invention can calculate the reflected light deviation ΔR of each pixel point ij (t), thereby identifying potential physical loss issues.

[0124] ΔR ij The larger the (t), the more serious the physical damage, and there may be problems such as scratches and stains.

[0125] This reflected light intensity deviation detection is the basis for evaluating the physical condition of the screen surface. The greater the deviation, the more obvious the impact of physical loss on clarity.

[0126] S33, calculate the physical loss score of the pixel

[0127]

[0128] η is the weight coefficient of the reflected light deviation, μ is the weight coefficient of the pixel health status value, and λ3 is the regularization term The weight coefficient is used to measure the impact of pixel brightness fluctuation on clarity. Indicates the relative change in pixel brightness;

[0129] S33 not only combines ΔR ij (t), and also incorporates the pixel health status H ij (t), through the combination of these two dimensions, the degree of physical loss is comprehensively evaluated, and the impact of physical loss of each pixel on screen clarity can be more accurately quantified. While detecting surface damage, it is possible to more flexibly handle brightness changes on the display, especially when video is playing or dynamic content is displayed, the dynamic changes in reflected light intensity and brightness may affect each other. It is able to innovatively combine information from multiple dimensions to comprehensively evaluate the impact of physical loss. Compared with traditional methods that rely only on reflected light deviation or pixel status, this model can more accurately identify potential problems that affect clarity.

[0130] S34, calculate the local physical loss score Clarity damage :

[0131]

[0132] W ij is the weight of the pixel with coordinates (i, j).

[0133] Clarity damageIndicates the impact of physical loss on the clarity of the entire loss area. damage The smaller the score, the greater the impact of physical loss on the clarity of the area. This score is used to quantify the negative impact of physical loss on the overall screen clarity. In this way, the present invention can identify the different impacts of physical loss on different areas and provide a basis for subsequent clarity repair or optimization.

[0134] Preferably, S4 includes:

[0135] S41, calculating the initial value of the comprehensive clarity score

[0136]

[0137] α4, β4 and γ4 represent Clarity local and (1-Clarity damage ) can be tuned through experimental data to ensure that these parameters have good adaptability to different screen types and environments; It is the average health status value of pixels in the entire screen;

[0138] The factors of the three dimensions are integrated in a weighted manner. Specifically, the model can be divided into three main parts:

[0139] Pixel health status: H ij (t) directly reflects the internal health of the pixel.

[0140] Local clarity: Clarity local Reflects the clarity of the area based on the overall health of the pixels.

[0141] Physical loss: Clarity damage Assess the negative impact of physical surface loss on screen display.

[0142] Through this weighted combination, It will be able to balance different factors (such as internal health and physical wear and tear) to ensure a more accurate assessment of the screen's overall clarity.

[0143] S42, establish the clarity regularization term λ reg :

[0144]

[0145] Is an indicator function, indicating that if Clarity local<θ, the value is 1, otherwise the value is 0. If the local clarity is lower than the set threshold θ, the pixel contributes to the regularization term; W ij is the weight of the pixel with coordinates (i, j);

[0146] λ reg Used to penalize situations where local losses are severe but the overall health is good, S42 can improve the robustness of the present invention and cope with extreme situations (such as partial screen damage but not affecting most of the display).

[0147] Through this regularization term, the model can prevent the clarity score from being overly sensitive to some areas with severe loss but small area, ensuring that the overall score is more representative.

[0148] S43, calculate the final value of the comprehensive clarity score:

[0149]

[0150] Represents the final value of the comprehensive clarity score.

[0151] Is the final comprehensive clarity score. Higher values ​​indicate better clarity, while lower values ​​indicate a greater probability of pixel health issues or physical wear and tear.

[0152] Preferably, S5 includes:

[0153] like The clarity evaluation result is: the screen display quality is excellent, with no significant problems;

[0154] like The clarity assessment result is: there are some problems with the screen display, and it is recommended to check the local area;

[0155] like The clarity assessment result is: there are obvious problems with the screen display, which may affect daily use.

[0156] According to the local clarity local and physical loss score Clarity damage , the present invention highlights the local problem area in the user interface. For each problem area, the present invention will display the specific problem type and location. For example, if a certain area A damage Clarity damage If the display effect is lower, the present invention will mark the area as "existing physical damage (such as scratches, stains)" and point out the area that affects the display effect.

[0157] Personalized optimization suggestions include physical cleaning suggestions, equipment setting suggestions, and user behavior suggestions;

[0158] like If the threshold κ is exceeded, a physical cleaning suggestion is generated to prompt the user to clean a specific area, such as “Please clean the lower left corner area, stains may cause blurry display”.

[0159] Recommended intensity values ​​for physical cleaning The formula for generating is:

[0160]

[0161] is the indicator function, W ij is the weight of the pixel with coordinates (i, j);

[0162] like A higher value indicates that the dynamic characteristics of pixels in the loss area are more abnormal, and a strong physical cleaning recommendation should be generated, prompting key areas to be cleaned.

[0163] like If the value is low, the physical cleaning recommendation may be just a reminder.

[0164] If H in some areas ij (t) or the brightness fluctuation is greater than 50%, a device setting suggestion will be generated, suggesting that the user adjust the device brightness, for example: "It is recommended to lower the screen brightness to reduce the impact of brightness fluctuations on the display."

[0165] Recommended intensity values ​​for device settings The formula for generating is:

[0166]

[0167] θ B Indicates the brightness fluctuation threshold, which indicates the fluctuation range of brightness relative to the initial brightness (such as 50%). If it exceeds this value, it is considered to be a large fluctuation;

[0168] A higher value indicates that there are more areas with abnormal brightness fluctuations or a higher weight, and the user is advised to adjust the device brightness;

[0169] Lower values ​​indicate less abnormal brightness fluctuations and device setting recommendations may be minor adjustments or no recommendations may be generated.

[0170] If it is detected that the pixels in some areas meet H ij (t)<θ H , indicating that the aging is serious, a user behavior suggestion is generated, suggesting that the user adjust the usage behavior, for example: "Please avoid using the device in a high-brightness environment for a long time to reduce the risk of pixel aging."

[0171] The proposed method generates a pixel-based health status value H ij (t):

[0172]

[0173] θ H is the threshold of the pixel's health status value.

[0174] The present invention integrates and displays all generated optimization suggestions through the user interface. Users can see:

[0175] Overall clarity rating: Display And use concise descriptions to help users understand the screen status.

[0176] Local problem marking and explanation: The present invention marks the problem area with different colors and provides detailed explanations about the problem type (such as physical loss, pixel aging, brightness fluctuation).

[0177] Specific optimization suggestions: Users can view and implement the operations suggested by the present invention, such as adjusting device settings, cleaning the screen, etc.

[0178] The generated user reports and optimization suggestions will serve as user operation guidance, and users can adjust device settings through one-click operations, such as reducing screen brightness, cleaning designated areas, adjusting screen refresh rate, etc. These operations are designed to improve screen clarity and extend screen life.

[0179] Preferably, S6 includes:

[0180] S61, update detection frequency:

[0181]

[0182] Freq new is the updated detection frequency, α6 is the adjustment factor used to control the frequency increase. is the indicator function, when the comprehensive clarity score When it is below the threshold θ, the detection frequency is increased;

[0183] Based on the feedback of the operation performed by the user and the final clarity evaluation of the screen, the present invention dynamically adjusts the detection model Ensure that the next detection can be more accurate and save computing resources.

[0184] If some areas were found to have major problems during previous inspections (e.g. or Clarity local The present invention will increase the detection frequency of this area to ensure more frequent monitoring.

[0185] S62, adjust the detection accuracy:

[0186]

[0187] Precision new is the updated detection accuracy, β6 is the accuracy adjustment coefficient, Used to indicate that the clarity score exceeds the set threshold θ up Reduce detection accuracy;

[0188] If the clarity score of certain areas is significantly improved after user adjustments (for example, cleaning or adjusting the brightness), the detection accuracy of the area can be appropriately reduced to save computing resources.

[0189] S63, update detection area:

[0190]

[0191] Area new Indicates the updated detection area;

[0192] In addition to adjusting the detection frequency and accuracy, the present invention also dynamically adjusts the detection area according to changes in specific problem areas. local If the problem is improved), the monitoring of this area can be reduced in the next inspection. Conversely, areas where the problem is still serious will continue to be inspected with high intensity.

[0193] When the local clarity local Significantly lower than the set threshold θ down When the problem area is improved, the detection range of the area is expanded; when the clarity of the problem area is improved, the detection density of the area is reduced.

[0194] This process ensures that the present invention can dynamically adapt to avoid over-detection of clear areas while maintaining a constant focus on problem areas.

[0195] S64, update the clarity regularization term:

[0196] The update formula of the clarity regularization term is:

[0197]

[0198] λ6 reg is the updated clarity regularization term, γ6 is the regularization coefficient, which is used to control the adaptive adjustment amplitude; local (new) and Clarity local (old) are the local clarity scores of the latest and last detection respectively.

[0199] In order to improve the stability of the detection model in long-term use, the present invention introduces a clarity regularization term to prevent over- or under-detection caused by frequent model adjustments. The regularization term is automatically adjusted according to user feedback results to ensure that the adaptive adjustment of the present invention does not overreact.

[0200] The present invention will compare the user feedback results with the updated detection model after each user operation. Iterative training is performed to optimize the accuracy of the detection strategy. The detection strategy after each model update will be adjusted based on the user's actual operation to ensure that the present invention can adapt to changes caused by long-term use and screen aging.

[0201] During each feedback iteration, The frequency, accuracy and area of ​​the adjustment will be optimized so that the resource utilization rate of the present invention is higher and the detection result is more accurate during each detection.

[0202] As user habits and device status change, the present invention continuously optimizes the detection process through continuous iterative learning, making the detection strategy more intelligent and personalized.

[0203] like Figure 2 In one embodiment shown, the present invention provides an intelligent definition detection system for a mobile phone display screen, comprising a detection strategy generation module, a first calculation module, a second calculation module, a third calculation module, a result generation module and an update module;

[0204] The detection strategy generation module is used to generate a detection strategy based on the content displayed on the display screen;

[0205] The first calculation module is used to obtain the monitoring area based on the detection strategy and calculate the local clarity score of the monitoring area;

[0206] The second calculation module is used to calculate the local physical loss score based on the local clarity score;

[0207] The third calculation module is used to calculate the comprehensive clarity score based on the local clarity score and the local physical loss score;

[0208] The result generation module is used to generate clarity evaluation results and personalized optimization suggestions based on the comprehensive clarity score;

[0209] The update module is used to update the detection strategy;

[0210] Among them, the monitoring area is obtained based on the detection strategy, and the local clarity score of the monitoring area is calculated, including:

[0211] S11, obtaining the image frame Frame of the content displayed on the mobile phone display t :

[0212] Frame t =[P uv |u∈[1,H],v∈[1,W]] (1)

[0213] t represents the time step; P uv Represents the RGB value of the pixel at the uth row and vth column; H and W represent the height and width of the display screen respectively;

[0214] S12, through the convolutional neural network uv Mapped to a high-dimensional feature vector F t ;

[0215] S13, using principal component analysis to analyze F t Perform dimensionality reduction to obtain the reduced eigenvector F′ t :

[0216] F′ t = PCA(F t ) (2)

[0217] F t Represents a high-dimensional feature vector extracted by a convolutional neural network; PCA represents the principal component analysis dimensionality reduction method;

[0218] S14, use the classification model to classify F′ t Perform classification and obtain the classification results of the content displayed on the display screen:

[0219]

[0220] It is a classification model based on multi-layer perceptron structure;

[0221] C t Indicates the classification result, and the value range is {1,2,3}, where 1, 2, and 3 correspond to the classification results of text, picture, and video respectively;

[0222] S15, generate detection strategy:

[0223]

[0224] Indicates the detection strategy, Freq indicates the detection frequency, Precision indicates the detection accuracy, and Area indicates the detection area.

[0225] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for detecting the clarity of a mobile phone display screen, characterized in that: include: S1, generating a detection strategy based on the content displayed on the display screen; S2, obtains the monitoring area based on the detection strategy and calculates the local clarity score of the monitoring area; S3, calculating a local physical loss score based on the local clarity score; S4, calculating a comprehensive clarity score based on the local clarity score and the local physical loss score; S5, generates clarity evaluation results and personalized optimization suggestions based on the comprehensive clarity score; S6, update detection strategy; Among them, S1 includes: S11, obtaining the image frame Frame of the content displayed on the mobile phone display t : Frame t =[P uv ∣u∈[1,H],v∈[1,W]] (1) t represents the time step; P uv Represents the RGB value of the pixel at the uth row and vth column; H and W represent the height and width of the display screen respectively; S12, through the convolutional neural network uv Mapped to a high-dimensional feature vector F t ; S13, using principal component analysis to analyze F t Perform dimensionality reduction to obtain the reduced eigenvector F′ t : F′ t =PCA(F t ) (2) F t Represents a high-dimensional feature vector extracted by a convolutional neural network; PCA represents the principal component analysis dimensionality reduction method; S14, use the classification model to classify F′ t Perform classification and obtain the classification results of the content displayed on the display screen: It is a classification model based on multi-layer perceptron structure; C t Indicates the classification result, and the value range is {1,2,3}, where 1, 2, and 3 correspond to the classification results of text, picture, and video respectively; S15, generate detection strategy: Indicates the detection strategy, Freq indicates the detection frequency, Precision indicates the detection accuracy, and Area indicates the detection area.

2. The intelligent definition detection method for a mobile phone display screen according to claim 1, characterized in that S2 include: S21, initialize the monitoring area of ​​the pixel: P Area ={(i,j)∣(i,j)∈Area} (5) P Area Represents the set of coordinates of the pixel points in the Area area. Area is composed of The detection area provided; (i, j) represents the coordinates of the pixel point; S22, generates calculation function for brightness and response time: B ij (t)=B ij (0)·exp(-γ ij ·t) (6) R ij (t)=R ij (0)+δ ij ·t (7) B ij (0) and R ij (0) represents the brightness and response time of the pixel with coordinates (i, j) in the initial state; γ ij is the brightness attenuation coefficient; δ ij is the rate at which the response time increases; S23, calculate the pixel health status value: H ij (t) represents the health status value of the pixel with coordinates (i, j); α2 and β2 are the influence weights of brightness and response time on health status respectively; λ2 is The influence weight of is used to penalize pixels with large brightness fluctuations; Indicates the fluctuation of pixel brightness; ΔB ij Indicates the brightness change value of the pixel with coordinates (i, j); S24, calculate the local clarity score Clarity local : W ij It is the weight of the pixel with coordinates (i, j), indicating the importance of the pixel with coordinates (i, j) in the monitoring area.

3. The intelligent definition detection method for a mobile phone display screen according to claim 2, characterized in that S3 include: S31, determine the key area A for physical loss detection damage : A damage ={(i,j)∈P Area ∣Clarity local <θ} (10) A damage It is the key area for physical loss detection; θ is the clarity threshold; S32, respectively calculating the reflected light deviation ΔR of each pixel point ij (t): R ij (t) is the actual reflected light intensity of the pixel with coordinates (i, j) at time t; is the ideal reflected light intensity of the pixel with coordinates (i, j) in an undamaged state; S33, calculate the physical loss score of the pixel η is the weight coefficient of the reflected light deviation, μ is the weight coefficient of the pixel health status value, and λ3 is the regularization term The weight coefficient is used to measure the impact of pixel brightness fluctuation on clarity. Indicates the relative change in pixel brightness; S34, calculate the local physical loss score Clarity damage : W ij is the weight of the pixel with coordinates (i, j).

4. The intelligent definition detection method for a mobile phone display screen according to claim 3, characterized in that S4 include: S41, calculating the initial value of the comprehensive clarity score α4, β4 and γ4 represent Clarity local and (1-Clarity damage ), It is the average health status value of pixels in the entire screen; S42, establish the clarity regularization term λ reg : Is an indicator function, indicating that if Clarity local <θ, the value is 1, otherwise the value is 0. If the local clarity is lower than the set threshold θ, the pixel contributes to the regularization term; W ij is the weight of the pixel with coordinates (i, j; S43, calculate the final value of the comprehensive clarity score: Represents the final value of the comprehensive clarity score.

5. The intelligent definition detection method for a mobile phone display screen according to claim 4, characterized in that S5 include: like The clarity evaluation result is: the screen display quality is excellent, with no significant problems; like The clarity assessment result is: there are some problems with the screen display, and it is recommended to check the local area; like The clarity assessment result is: there are obvious problems with the screen display, which may affect daily use. Personalized optimization suggestions include physical cleaning suggestions, equipment setting suggestions, and user behavior suggestions; like If the threshold value κ is exceeded, a physical cleaning suggestion is generated to prompt the user to clean a specific area. Recommended intensity values ​​for physical cleaning The formula for generating is: is the indicator function, W ij is the weight of the pixel with coordinates (i, j); If H in some areas ij (t) or the brightness fluctuation is greater than 50%, a device setting suggestion will be generated, suggesting that the user adjust the device brightness. Recommended intensity values ​​for device settings The formula for generating is: θ B represents the brightness fluctuation threshold, If it is detected that the pixels in some areas meet H ij (t)<θ H , indicating that the aging is serious, a user behavior suggestion is generated to suggest that the user adjust the usage behavior. The suggestion is generated based on the health status H of the pixel point. ij (t): θ H is the threshold of the pixel health status value.

6. The intelligent definition detection method for a mobile phone display screen according to claim 5, characterized in that S6 include: S61, update detection frequency: Freq new is the updated detection frequency, α6 is the adjustment factor used to control the frequency increase. is the indicator function, when the comprehensive clarity score When it is below the threshold θ, the detection frequency is increased; S62, adjust the detection accuracy: Precision new is the updated detection accuracy, β6 is the accuracy adjustment coefficient, Used to indicate that the clarity score exceeds the set threshold θ up Reduce detection accuracy; S63, update detection area: Area new Indicates the updated detection area; S64, update the clarity regularization term: The update formula of the clarity regularization term is: λ6 reg is the updated clarity regularization term, γ6 is the regularization coefficient, which is used to control the adaptive adjustment amplitude; local (new) and Clarity local (old) are the local clarity scores of the latest and last detection respectively.

7. An intelligent definition detection system for a mobile phone display screen, characterized in that: It includes a detection strategy generation module, a first calculation module, a second calculation module, a third calculation module, a result generation module and an update module; The detection strategy generation module is used to generate a detection strategy based on the content displayed on the display screen; The first calculation module is used to obtain the monitoring area based on the detection strategy and calculate the local clarity score of the monitoring area; The second calculation module is used to calculate the local physical loss score based on the local clarity score; The third calculation module is used to calculate the comprehensive clarity score based on the local clarity score and the local physical loss score; The result generation module is used to generate clarity evaluation results and personalized optimization suggestions based on the comprehensive clarity score; The update module is used to update the detection strategy; Among them, the monitoring area is obtained based on the detection strategy, and the local clarity score of the monitoring area is calculated, including: S11, obtaining the image frame Frame of the content displayed on the mobile phone display t : Frame t =[P uv ∣u∈[1,H],v∈[1,W]] (1) t represents the time step; P uv Represents the RGB value of the pixel at the uth row and vth column; H and W represent the height and width of the display screen respectively; S12, through the convolutional neural network uv Mapped to a high-dimensional feature vector F t ; S13, using principal component analysis to analyze F t Perform dimensionality reduction to obtain the reduced eigenvector F′ t : F′ t =PCA(F t ) (2) F t Represents a high-dimensional feature vector extracted by a convolutional neural network; PCA represents the principal component analysis dimensionality reduction method; S14, use the classification model to classify F′ t Perform classification and obtain the classification results of the content displayed on the display screen: It is a classification model based on multi-layer perceptron structure; C t Indicates the classification result, and the value range is {1,2,3}, where 1, 2, and 3 correspond to the classification results of text, picture, and video respectively; S15, generate detection strategy: Indicates the detection strategy, Freq indicates the detection frequency, Precision indicates the detection accuracy, and Area indicates the detection area.

Images