A quantitative evaluation data set construction method, device and equipment

Abstract

The present invention relates to a method, device, and equipment for constructing a quantitative evaluation data set, which includes the following steps: performing XYZ to RGB conversion on an actual screen mura image pixel by pixel to obtain an RGB simulation image; performing quantitative scoring based on the RGB simulation image and obtaining a data set; and applying the data set to mura testing. The embodiment of the present invention provides a method, device, and equipment for constructing a quantitative evaluation data set. Since the actual screen mura image is converted from XYZ to RGB pixel by pixel, the obtained RGB simulation image is described based on the XYZ color space. At this time, even if the RGB image is used to drive the display of the mura simulation image, the same mura simulation image in the data set can have the same display effect on different display devices, that is, device independence. Therefore, it is unlikely that there will be differences in subjective evaluation values ​​for the same mura simulation image, nor will it be likely to affect the development of the mura quantification algorithm.

Description

Technical Field

[0001] The present invention relates to the field of quantitative evaluation of display panel Mura, and in particular to a method, device and equipment for constructing a quantitative evaluation data set. Background Art

[0002] Display screens have become ubiquitous in our daily lives, including mobile phones, televisions, electronic road signs, and outdoor advertising. During the display manufacturing process, defects are inevitably introduced, impacting display quality and ultimately product quality. Traditionally, mura detection based on AOI (automated optical inspection) and manual visual inspection has been implemented at key stages of the manufacturing process to categorize mura levels.

[0003] AOI inspection primarily uses high-resolution industrial cameras. Leveraging large datasets of real-world mura images collected from production lines and mature deep learning algorithms, mura defect detection currently achieves excellent results, with an accuracy rate of 98% and a missed detection rate of less than 0.2%. However, mura quantification and assessment still rely on manual labor, which is time-consuming, labor-intensive, and has poor repeatability. To address this issue, researchers have proposed replacing traditional manual mura assessment with mura quantification assessment algorithms. Despite some progress in mura quantification algorithms, their effectiveness has not been universally recognized within the industry, and manual visual inspection remains the primary method. Furthermore, datasets consisting of mura simulation images and their corresponding subjective scores are key factors in determining the quality of mura quantification algorithm development.

[0004] In the related art, when performing quantitative scoring, the mainstream method generally regenerates several Mura simulation images with different RGB color gamut space descriptions according to the actual screen Mura type, and then Figure 1 First, it is displayed on the screen and then quantitatively scored using the five-level image quality subjective evaluation experiment recommended by ITU-R.

[0005] However, the simulation image is expressed using the RGB color space, and the display effect of this color space is device-dependent. That is, the same mura simulation image will appear differently on different screens, leading to different subjective evaluation values ​​for the same mura simulation image, ultimately affecting the development of mura quantification algorithms.

[0006] Therefore, it is necessary to design a new method for constructing quantitative evaluation datasets to overcome the above problems. Summary of the Invention

[0007] Embodiments of the present invention provide a method, apparatus, and device for constructing a quantitative evaluation dataset to address the problem in related technologies where the same mura simulation image may display differently on different screens, leading to different subjective evaluation values ​​for the same mura simulation image, ultimately affecting the development of mura quantification algorithms.

[0008] In a first aspect, a method for constructing a quantitative evaluation dataset is provided, comprising the following steps: converting an actual screen mura image pixel by pixel from XYZ to RGB to obtain an RGB simulation image; performing quantitative scoring based on the RGB simulation image and obtaining a dataset; and applying the dataset to mura testing.

[0009] In some embodiments, the actual screen Mura map is converted from XYZ to RGB pixel by pixel to obtain an RGB simulation map, including: obtaining an XYZ simulation map based on the actual screen Mura map; and converting the XYZ simulation map to an RGB simulation map pixel by pixel based on the sub-pixel luminescence characteristics of the experimental ideal screen.

[0010] In some embodiments, obtaining an XYZ simulation image based on an actual screen Mura image includes: extracting pixel-by-pixel brightness of the actual screen Mura image and normalizing it to obtain a normalized image; and multiplying the normalized image by a set white point to obtain an XYZ simulation image.

[0011] In some embodiments, the conversion of the XYZ simulation image to the RGB simulation image is performed pixel by pixel based on the sub-pixel luminescence characteristics of the experimental ideal screen, including: measuring the XYZ values ​​and gamma values ​​of the RGB sub-pixels of the experimental ideal screen at the G0 grayscale, wherein the G0 grayscale is the grayscale that is closest to the actual measured screen brightness and the preset brightness; using the XYZ values ​​and gamma values ​​of the RGB sub-pixels of the experimental ideal screen at the G0 grayscale, the conversion of the XYZ simulation image to the RGB simulation image is performed pixel by pixel.

[0012] In some embodiments, performing quantitative scoring based on the RGB simulation image and obtaining a data set includes: performing double stimulation scoring based on a limit sample on the RGB simulation image to obtain an RGB simulation image score.

[0013] In some embodiments, the RGB simulation image is subjected to a dual-stimulation scoring based on a limit sample to obtain an RGB simulation image score, including: displaying the RGB simulation image and the limit sample simulation image on a screen; finding the limit sample that is closest to the visual parameters of the RGB simulation image, and scoring the RGB simulation image with reference to the quantized value of the limit sample.

[0014] In some embodiments, before converting the actual screen mura image pixel by pixel from XYZ to RGB to obtain an RGB simulation image, the process further includes: using AOI to perform mura identification on the actual screen mura image to obtain a mura type and a binary mask image.

[0015] In some embodiments, the data set includes an XYZ simulation image obtained based on an actual screen Mura image, a score corresponding to the simulation image, a binary mask image for describing the position of the simulation image, and a text file describing the simulation image information.

[0016] In a second aspect, a device for constructing a quantitative evaluation dataset is provided, which includes: a conversion module for converting the actual screen mura image from XYZ to RGB pixel by pixel to obtain an RGB simulation image; a data construction module for performing quantitative scoring based on the RGB simulation image and obtaining a dataset; and a testing module for applying the dataset to mura testing.

[0017] In a third aspect, a quantitative evaluation dataset construction device is provided, which includes a processor, a memory, and a quantitative evaluation dataset construction program stored in the memory and executable by the processor, wherein when the quantitative evaluation dataset construction program is executed by the processor, the steps of the above-mentioned quantitative evaluation dataset construction method are implemented.

[0018] The beneficial effects brought about by the technical solution provided by the present invention include:

[0019] Embodiments of the present invention provide a method, apparatus, and device for constructing a quantitative evaluation dataset. By performing a pixel-by-pixel XYZ-to-RGB conversion on the actual screen mura image, the resulting RGB simulation image is described based on the XYZ color space. This allows the same mura simulation image in the dataset to appear identical on different display devices, even when the RGB image is used as a driver for display. This ensures device independence. Consequently, subjective evaluations of the same mura simulation image are less likely to differ, minimizing the impact on the development of mura quantification algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A flowchart of a method for constructing a quantitative evaluation data set provided by an embodiment of the present invention;

[0022] Figure 2 A schematic diagram of a Mura simulation diagram provided by an embodiment of the present invention;

[0023] Figure 3 A schematic diagram of a binary mask image provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0025] Embodiments of the present invention provide a method for constructing a quantitative evaluation dataset, which can address the problem in related technologies that the same mura simulation image may display differently on different screens, leading to different subjective evaluation values ​​for the same mura simulation image, ultimately affecting the development of mura quantification algorithms.

[0026] See also Figure 1 As shown, a method for constructing a quantitative evaluation dataset provided by an embodiment of the present invention may include the following steps:

[0027] S1: Convert the actual screen mura image from XYZ to RGB pixel by pixel to obtain an RGB simulation image. In this embodiment, an imaging device can be used to capture the actual screen mura image.

[0028] S2: Perform quantitative scoring based on the RGB simulation image and obtain a dataset. The dataset consists of individual data items, each of which may include: an XYZ-based mura simulation image, the score corresponding to each simulation image, a binary mask image used to describe the location of the mura simulation image, and a text file describing the mura simulation image information. The mura simulation image information description text file may include information such as mura type, quantity, size, and location.

[0029] S3: Applying the dataset to mura testing. The dataset can be applied to mura quantification algorithm development and testing, and the development and testing are all for the purpose of mura testing.

[0030] In this embodiment, the actual screen mura image is converted pixel by pixel from XYZ to RGB. The resulting RGB simulation image is described based on the XYZ color space. Therefore, even if the RGB image is used to drive the display of the mura simulation image, the same mura simulation image in the dataset can be displayed identically on different display devices, i.e., device-independent. Therefore, it is unlikely that subjective evaluation values ​​for the same mura simulation image will differ, nor will it affect the development of the mura quantification algorithm.

[0031] further, Figure 2 and Figure 3 The following table shows the Mura simulation image and the binary mask image. The subjective score is a floating point number between 0 and 5, where 5 represents the strongest visual experience and 0 represents an ideal uniform image without Mura. Figure 3 It can be seen that Figure 3 The mura area in the binary mask image is Figure 2 The mura locations in the mura simulation images are exactly the same, marking the exact location and range of the mura.

[0032] Furthermore, this embodiment preferably uses the CIE 1931XYZ color space to describe the mura simulation diagram. The CIE uses monochromatic light with wavelengths of 700nm, 546.1nm, and 435.8nm as the three primary colors of RGB for color matching experiments. The matching results are then transformed to eliminate negative values, resulting in the CIE 1931XYZ color space. According to the additive color model, if any color and another color mixed with different components of the three primary colors appear the same to humans, then the components of these three primary colors are called the tristimulus values ​​of that color, namely X, Y, and Z in the CIE 1931 color space. The CIE XYZ color space is based on direct measurements of human color vision and serves as the definition basis for many other color spaces. Therefore, describing mura using the CIE 1931XYZ color space is device-independent, and the mura simulation diagram described in this color space can theoretically fully reproduce the visual effect of actual mura.

[0033] In the related art, a number of Mura simulation images with different RGB color gamut descriptions are generated based on the actual screen Mura type. This Mura simulation image is too idealized and the generated simulation is too Figure 1A Gaussian distribution is generally used to describe the brightness transition of mura from the center to the background, but in reality, brightness transitions vary, and the contours of the simulation image are ideal shapes, such as ideal geometric shapes such as circles, ellipses, or horizontal lines. In this embodiment, before converting the actual screen mura image pixel by pixel from XYZ to RGB to obtain the RGB simulation image, the following process may also be included: using AOI (Automated Optical Inspection) to identify mura on the actual screen mura image to obtain the mura type and binary mask image. In this embodiment, the mura type, brightness distribution, shape, size, etc. of the mura simulation image can be generated using the actual mura obtained in the AOI inspection as a sample, thereby improving the authenticity of the mura simulation image.

[0034] In some embodiments, in step S1, converting the actual screen mura map from XYZ to RGB pixel by pixel to obtain an RGB simulation map may include: obtaining an XYZ simulation map based on the actual screen mura map; and converting the XYZ simulation map to an RGB simulation map pixel by pixel based on the sub-pixel luminescence characteristics of an experimental ideal screen. Since display panels generally have RGB primary color sub-pixels, each sub-pixel's brightness is controlled by 8-bit grayscale data. The relationship between brightness and grayscale is generally a power function, where the exponent value is called the gamma value. Different display panels generally have certain differences in brightness and color gamut. A qualified display panel has consistent sub-pixel luminescence characteristics at different locations, ensuring that they are difficult for the human eye to distinguish. In this embodiment, based on this feature, the mura simulation map in the CIE 1931 XYZ color space can be converted to the RGB space of each screen and displayed, achieving the same mura simulation display effect on different screens.

[0035] Of course, in other embodiments, the manufacturer of the ideal experimental screen may provide a luminous characteristic index document, and the document parameters may be directly looked up in a table for determination, or the conversion from the XYZ simulation image to the RGB simulation image may be performed pixel by pixel.

[0036] In some optional embodiments, obtaining an XYZ simulation image based on an actual screen mura image may include: extracting pixel-by-pixel brightness from the actual screen mura image and normalizing it to obtain a normalized image; and multiplying the normalized image by a set white point to obtain the XYZ simulation image. In this embodiment, a white point value must be specified. The white point can be a commonly used value such as D65 or D50. For database specifications and standards, the white point is generally not subsequently changed after it is determined. In this embodiment, the white point is preferably D65 (X=95.047, Y=100, Z=108.883).

[0037] Preferably, after obtaining the XYZ simulation image, the size of the simulation image can be unified to a specific pixel by filling and cropping, such as 2400×1080 selected in this embodiment, of course, other values ​​can also be used, and the XYZ value of the filled pixel is equal to the XYZ value of the D65 white point.

[0038] Of course, in other embodiments, an area array colorimeter can also be used to obtain an XYZ simulation image of the actual screen Mura, and the image can be processed using conventional computer graphics processing methods (such as rotation, scaling, contrast stretching, shape transformation, etc.), and the background area value is set to equal the set white point value to obtain the XYZ simulation image.

[0039] In some embodiments, the conversion of the XYZ simulation image to the RGB simulation image pixel by pixel based on the sub-pixel luminescence characteristics of the experimental ideal screen can include: measuring the XYZ values ​​and gamma values ​​of the RGB sub-pixels of the experimental ideal screen at the G0 grayscale, wherein the G0 grayscale value is first determined. The G0 grayscale can be the grayscale at which the actually measured screen brightness is closest to the preset brightness, for example, the G0 grayscale can be the grayscale at which the W screen brightness of 32, 64, 96, 128, 160, 192, 224, or 255 is closest to 100 nit (of course, it can also be the grayscale at which the screen brightness of other values ​​is closest to 100 nit, and 100 nit can also be selected from other values). A colorimeter can be used to measure the XYZ values ​​and gamma values ​​of the RGB sub-pixels at the G0 grayscale, and the pixel pitch P can also be measured or extracted. Then, the XYZ simulation image to the RGB simulation image can be converted pixel by pixel using the XYZ values ​​and gamma values ​​of the RGB sub-pixels of the experimental ideal screen at the G0 grayscale.

[0040] Of course, in other embodiments, conversion can also be achieved using other grayscales, but the conversion error will be relatively large. The fundamental reason is that this conversion process relies on the screen brightness and grayscale satisfying an ideal exponential model, but the actual screen cannot be ideal and there will always be deviations. The XYZ measured values ​​of the G0 grayscale are close to the target values, and the error will be relatively small.

[0041] In this embodiment, the conversion from the XYZ image to the RGB image is because when the ideal screen displays the RGB image in the experiment, the XYZ is measured with a colorimeter, and the measured XYZ is consistent with the XYZ simulation image.

[0042] The conversion formula is as follows:

[0043]

[0044]

[0045] Where, X R 、Y R and ZR They are the XYZ values ​​of the R sub-pixel at G0 grayscale, X G 、Y G and Z G They are the XYZ values ​​of the G sub-pixel at G0 grayscale, X B 、Y B and Z B are the XYZ values ​​of the B sub-pixel at grayscale G0, and gammaR, gammaG, and gammaB are the gamma values ​​of the RGB sub-pixels respectively.

[0046] The above conversion formula can be used to convert the XYZ analog image to the RGB analog image.

[0047] Most subjective scoring experiments in related technologies use a single stimulus, or conduct dual-stimulus subjective quantitative evaluation experiments of image quality by comparing with a mura-free image. This method works well for video images with texture features, but has limited ability to resolve mura patterns in images lacking detailed images.

[0048] In some embodiments of the present application, in step 2, performing a quantitative score based on the RGB simulation image and obtaining a dataset may include performing a dual-stimulus score on the RGB simulation image based on a limit sample to obtain an RGB simulation image score. In this embodiment, the quantified value of the mura simulation image is obtained using a dual-stimulus method, i.e., the mura simulation image to be evaluated is compared with a mura limit sample with a given score. The limit sample score with the highest matching degree can be used as the subjective evaluation value. This can greatly reduce observer subjective bias and increase the objectivity and stability of the subjective score.

[0049] The limit sample is a standard mura simulation image. Its subjective scoring process is largely the same as the subjective scoring workflow for the mura simulation image sample, but the scoring is determined by the transmittance of the ND filter. The corresponding relationship is shown in the table below.

[0050] Table 1 Subjective ratings based on ND filters

[0051] Subjective rating condition 0 No mura is visible without ND filters 1 Mura is visible without ND filter, but invisible through ND8 2 Mura is visible through an ND8 filter, but not through an ND16 filter. 3 Mura is visible through an ND16 filter, but not through an ND32 filter. 4 Mura is visible through the ND32 filter, but not through the ND64 filter. 5 Mura visible through ND64

[0052] The limit samples in this embodiment are obtained using ND filters with different transmittances, which is consistent with actual usage on the production line. The observer only needs to make a binary judgment on whether mura is visible to complete the experimental process. The experimental process is simple and the experimental results are reliable.

[0053] In some optional embodiments, in addition to using the aforementioned "dual stimulus test + limited sample" method to obtain a score, the method described in Table 1 (corresponding to the previously described method for generating a limited sample score) can also be used to directly quantitatively evaluate the mura simulation image to obtain a subjective score. In other words, in this embodiment, the subjective score of the mura simulation image can be directly obtained using an ND filter, without the need for the "dual stimulus test + limited sample" method.

[0054] Furthermore, the RGB simulation image is subjected to a dual-stimulation scoring based on a limit sample to obtain an RGB simulation image score, which may include: displaying the RGB simulation image and the limit sample simulation image on a screen, wherein the screen may be split or combined; finding the limit sample closest to the visual parameters of the RGB simulation image, and using the quantized value of the limit sample as a reference to score the RGB simulation image. In this embodiment, in actual operation, the experimenter may observe the RGB simulation image and the limit sample simulation image at a set distance in a dark room, wherein the set distance is preferably a distance D = 2750 × P, wherein P is the pixel pitch P measured or extracted above, and at this distance, the spatial frequency of one pixel of the screen in the human eye is exactly 48 cpd; then the experimenter may switch the limit sample simulation image and find the limit sample closest to the visual perception of the RGB simulation image, and use the quantized value of the limit sample as a reference to score the RGB simulation image, wherein the RGB simulation image also describes the Mura simulation image, so the RGB simulation image score is also the Mura simulation image score. Of course, in actual operation, other machines or devices can also be used to automatically find the limit sample that is closest to the visual parameters of the RGB simulation image, and use the quantitative value of the limit sample as a reference to score the RGB simulation image.

[0055] In other embodiments, when obtaining the RGB simulation image score based on the limit sample, the limit sample closest to the RGB simulation image can also be found according to a commonly used image similarity evaluation function (open source mature algorithms can be found) to replace manual viewing here.

[0056] In this embodiment, human visual sensitivity is related to factors such as the distance, size, and contrast of an object. Because the focal length of the human lens is constant, and the number and density of visual cells on the retina are also fixed, the human eye's perception of object size is based on spatial angle. Extensive CSF research has shown that the human eye's limit for spatial angle observation is approximately 48 cpd; details above this spatial frequency are imperceptible to human vision. Therefore, a single pixel in the mura simulation image represents a spatial frequency of 48 cpd, which fully describes the entire image information perceived by the human eye.

[0057] Furthermore, after obtaining the scores, the scoring results can be statistically screened, with the average value used as the subjective score. Finally, the XYZ simulation image (preferably a simulation image with uniform pixels), the subjective score, the binary mask image used to describe the position of the RGB simulation image (also known as the mura simulation image), and the mura simulation image information description text file (which can include information such as mura type, quantity, size, and location) can be combined into a single piece of data and added to the dataset.

[0058] The dataset constructed using the embodiments of the present invention can solve the problems of lack of unified standards and poor practicality in the traditional dataset construction process, and can be used as a standard dataset for the development and testing of mura quantification algorithms.

[0059] This paper proposes a method for constructing a quantitative assessment dataset. Compared to traditional methods, this method produces simulated mura images that more closely resemble actual mura. It also supports parallel dataset expansion, facilitating the construction of large-scale datasets. This dataset can be used for the development, improvement, and testing of mura quantification algorithms, supporting their replacement for manual assessment on production lines.

[0060] Embodiments of the present invention also provide a device for constructing a quantitative evaluation dataset, which may include: a conversion module for converting an actual screen mura image pixel by pixel from XYZ to RGB to obtain an RGB simulation image; a data construction module for performing quantitative scoring based on the RGB simulation image and obtaining a dataset; and a testing module for applying the dataset to mura testing. The conversion module and data construction module can implement the steps of the corresponding embodiments of the above-mentioned method and are not further described here.

[0061] An embodiment of the present invention also provides a quantitative evaluation dataset construction device, characterized in that the quantitative evaluation dataset construction device includes a processor, a memory, and a quantitative evaluation dataset construction program stored in the memory and executable by the processor, wherein when the quantitative evaluation dataset construction program is executed by the processor, the steps of the quantitative evaluation dataset construction method provided in any of the above embodiments are implemented.

[0062] The descriptions of the processes corresponding to the above figures have different focuses. For parts that are not described in detail in a certain process, please refer to the relevant descriptions of other processes.

[0063] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The above-mentioned computer program product includes one or more computer instructions. When the above-mentioned computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The above-mentioned computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above-mentioned computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the above-mentioned computer instructions can be transmitted from a website, computer, server or data center to another website, computer, server or data center by wired (such as coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.) mode. The above-mentioned computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrations. The above-mentioned available medium can be a magnetic medium (such as a floppy disk, a storage disk, a tape), an optical medium (such as a DVD), or a semiconductor medium (such as an SSD), etc. In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0064] In the several embodiments provided in this application, it should be understood that the disclosed devices can also be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the indirect coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0065] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment of the present application.

[0066] In addition, each functional unit in each embodiment of the present application may be integrated into a processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The above-mentioned integrated units may be implemented in the form of hardware or software functional units.

[0067] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium may include, for example: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk.

[0068] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is intended to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for constructing a quantitative evaluation dataset, characterized in that: It includes the following steps: Convert the actual screen Mura image from XYZ to RGB pixel by pixel to obtain an RGB simulation image; Perform quantitative scoring based on the RGB simulation image and obtain a dataset; the dataset includes an XYZ simulation image obtained based on the actual screen mura image, the score corresponding to the simulation image, a binary mask image used to describe the location of the simulation image, and a text file describing the simulation image information; The dataset was applied to mura testing.

2. The method for constructing a quantitative evaluation data set according to claim 1, wherein: The actual screen Mura image is converted from XYZ to RGB pixel by pixel to obtain an RGB simulation image, including: Get an XYZ simulation image based on the actual screen Mura image; According to the sub-pixel luminous characteristics of the ideal screen in the experiment, the XYZ simulation image is converted to the RGB simulation image pixel by pixel.

3. The method for constructing a quantitative evaluation data set according to claim 2, wherein: The XYZ simulation image is obtained based on the actual screen Mura image, including: Extract the brightness of the actual screen Mura image pixel by pixel and normalize it to obtain a normalized image; Multiply the normalized image by the set white point to obtain the XYZ simulation image.

4. The method for constructing a quantitative evaluation data set according to claim 2 or 3, wherein: The conversion of the XYZ simulation image to the RGB simulation image is performed pixel by pixel based on the sub-pixel luminous characteristics of the ideal screen in the experiment, including: Measure the XYZ values ​​and gamma values ​​of the RGB sub-pixels of the ideal screen in the experiment at grayscale G0, where grayscale G0 is the grayscale at which the actual measured screen brightness is closest to the preset brightness; Using the XYZ values ​​and gamma values ​​of the RGB sub-pixels of the ideal experimental screen at the G0 grayscale, the XYZ simulation image is converted to the RGB simulation image pixel by pixel.

5. The method for constructing a quantitative evaluation data set according to claim 1, wherein: The quantitative scoring based on the RGB simulation image and obtaining the data set include: The RGB analog image is subjected to double stimulation scoring based on the limit sample to obtain the RGB analog image score.

6. The method for constructing a quantitative evaluation data set according to claim 5, wherein: The step of performing a double stimulation score on the RGB simulation image based on a limit sample to obtain an RGB simulation image score comprises: Display the RGB simulation image and the limit sample simulation image on the screen; Find the limit sample that is closest to the visual parameters of the RGB simulation image, and use the quantitative value of the limit sample as a reference to score the RGB simulation image.

7. The method for constructing a quantitative evaluation data set according to claim 1, wherein: Before converting the actual screen Mura image pixel by pixel from XYZ to RGB to obtain an RGB simulation image, the method further includes: Use AOI to identify the actual screen mura image and obtain the mura type and binary mask image.

8. A device for constructing a quantitative evaluation data set, characterized in that: It includes: The conversion module is used to convert the actual screen mura image from XYZ to RGB pixel by pixel to obtain an RGB simulation image; A data construction module is used to perform quantitative scoring based on the RGB simulation image and obtain a data set; the data set includes an XYZ simulation image obtained based on the actual screen mura image, the score corresponding to the simulation image, a binary mask image used to describe the location of the simulation image, and a text file describing the simulation image information; A testing module is configured to apply the data set to mura testing.

9. A quantitative evaluation data set construction device, characterized in that: The quantitative evaluation dataset construction device includes a processor, a memory, and a quantitative evaluation dataset construction program stored in the memory and executable by the processor, wherein when the quantitative evaluation dataset construction program is executed by the processor, the steps of the quantitative evaluation dataset construction method according to any one of claims 1 to 7 are implemented.

Images