A Method for Measuring Subjective Crosstalk in Naked-Eye 3D Display Systems Based on Image Features

Abstract

A method for measuring subjective crosstalk in a glasses-free 3D display system based on image features is presented. This method first constructs a scoring crosstalk test image library and a feature scoring test image library, designs a subjective perception scoring scale, obtains subjective scores from human viewing, and establishes a functional relationship between subjective scores and image crosstalk values ​​using function fitting. Image feature values ​​are extracted through camera capture, and combined with the calculated subjective scores to train a subjective score prediction model, ultimately achieving subjective perception crosstalk prediction based on image feature values. This invention overcomes the shortcomings of existing technologies that rely solely on objective parameter measurements and cannot reflect the viewer's subjective experience. It can quantitatively measure the subjective perception crosstalk of a specific image on a specific device, exhibiting high accuracy (fitting accuracy 97.8%, correlation 97.37%) and high efficiency (measurement accuracy 94.6%), providing reliable technical support for optimizing the performance of glasses-free 3D display systems.

Description

Technical Field

[0001] This invention relates to naked-eye 3D display testing technology, and in particular to a method for measuring subjective crosstalk in naked-eye 3D display systems based on image features. Background Technology

[0002] Vision plays a crucial role in human perception, with 80% of information acquired visually in daily life. To meet ever-growing needs, display technology has evolved. Since the 21st century, 3D display technology has become a prominent cutting-edge technology. Among these, glasses-free 3D display technology has garnered significant attention due to its advantages such as requiring no visual aids, simple structure, low cost, and superior performance, and has been widely applied in outdoor advertising screens, product exhibitions, and other fields. However, glasses-free 3D displays suffer from crosstalk issues, which can negatively impact the viewing experience.

[0003] To evaluate the display effect of glasses-free 3D display systems, researchers have proposed various methods for measuring crosstalk. However, current crosstalk testing methods mainly rely on objective parameter measurements. Crosstalk values ​​are related to display device performance and image content, failing to reflect the impact of crosstalk levels on the viewer's subjective experience. Furthermore, existing evaluation systems are complex and diverse, hindering widespread application. Therefore, a rapid evaluation method for crosstalk in glasses-free 3D display systems that incorporates viewer subjective experience is urgently needed.

[0004] It should be noted that the information disclosed in the background section above is only for understanding the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The main objective of this invention is to overcome the deficiencies in the aforementioned background technology and provide a method for measuring subjective crosstalk in naked-eye 3D display systems based on image features.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for measuring subjective crosstalk in a naked-eye 3D display system based on image features, comprising the following steps:

[0008] S1: Construct a score crosstalk test library and a feature score test library for naked-eye 3D display devices, respectively. The score crosstalk test library contains multiple sets of stereo image pairs with increasing image crosstalk rate gradients, and the feature score test library contains stereo image pairs without crosstalk and stereo image pairs with crosstalk.

[0009] S2: Design a subjective feeling rating scale, and divide the rating into multiple levels based on subjective evaluation indicators of visual perception characteristics; load the rating crosstalk test image library on a naked-eye 3D display device to obtain the subjective rating of naked-eye 3D images viewed by the human eye.

[0010] S3: Use the image crosstalk values ​​of the scoring crosstalk test library and the subjective ratings of the library obtained in step S2 to perform function fitting to establish a functional relationship between the subjective ratings of the images and the image crosstalk values.

[0011] S4: Obtain the left and right monocular views captured by the camera after the naked-eye 3D display device loads the feature scoring test image library, and fuse them into a binocular image; extract image feature values ​​and concatenate all feature values;

[0012] S5: Select several observers and load the feature rating test image library onto a naked-eye 3D display device to conduct a subjective experiment. Obtain the observers' subjective ratings of the images in the feature rating test image library as labels for training data. Use the subjective ratings obtained from the experiment and the cascaded feature value data obtained in step S4 to train a subjective rating prediction model through regression prediction or time series prediction algorithms until the error is reduced to a predetermined range. Establish a subjective feeling rating prediction model based on the measured feature values ​​obtained from the images captured by the camera. Predict the subjective feeling rating based on the image feature values ​​and determine the subjective feeling crosstalk based on the functional relationship between the rating and the function obtained in step S3.

[0013] Furthermore, in step S1, the scoring crosstalk test image library contains multiple sets of stereo image pairs with increasing image crosstalk rate gradients, and the crosstalk rate gradient range covers the maximum objective crosstalk rate range that the human eye can perceive as the stereo effect of the image.

[0014] Furthermore, in step S2, the subjective evaluation indicators based on visual perception characteristics include the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen.

[0015] Furthermore, in step S2, the process of obtaining the subjective score of the crosstalk test library for human eye viewing of naked-eye 3D display devices includes:

[0016] The descriptions in the subjective rating scales were used as a comparative reference.

[0017] Observers view the images displayed by the naked-eye 3D display system and rate three indicators: the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen.

[0018] The subjective perception score of the image on the naked-eye 3D display system is obtained by calculating the weighted average of the scores for the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen.

[0019] Furthermore, in step S3, the functional relationship established by fitting the image crosstalk value and the subjective score includes, but is not limited to, exponential functions, logarithmic functions, linear functions, and quadratic functions.

[0020] Further, in step S3, the image crosstalk rate and the subjective feeling score of each observer are respectively fitted by a function, and then the crosstalk-subjective feeling score fitting curves of multiple observers are averaged to obtain the average crosstalk-subjective feeling score fitting function.

[0021] Further, in step S4, the acquired left and right monocular views are first preprocessed, including image cropping, removing screen background and edges, etc., and the images with removed screen background and edges are subjected to noise reduction filtering and grayscale processing; then the processed monocular images are fused into binocular images, and feature values ​​are extracted from the monocular images and binocular images.

[0022] Further, in step S4, the image feature values ​​include disparity feature values ​​between stereo image pairs and HOG feature values ​​of the fused binocular image. The feature concatenation includes merging the HOG feature values ​​of the binocular image with the global disparity difference and local disparity difference feature values ​​of the stereo image pairs into a feature matrix.

[0023] Further, in step S5, the training process of the subjective rating prediction model includes:

[0024] Model training is performed using regression prediction or time series prediction algorithms based on subjective ratings and concatenated feature value data from the feature rating test library.

[0025] By adjusting one or more parameters, such as the sampling ratio, regression parameters, and number of iterations, the error can be reduced to a predetermined range, thus completing model training.

[0026] Furthermore, in step S5, the subjective rating prediction model is trained using a regression prediction algorithm or a time series prediction algorithm, and the mean squared error (MSE) is used as an indicator to measure the model's prediction performance.

[0027] Further, in step S5, predicting the subjective perception crosstalk score based on image feature values ​​includes: loading the test image onto the display device under test, extracting the display features of the image under test using the method in step S4, inputting the display features of the image under test as test set data into the subjective score prediction model, and obtaining the subjective score output.

[0028] Furthermore, the subjective rating output is used as an independent variable and input into the rating crosstalk fitting model based on the functional relationship in step S3 to obtain the image subjective perception crosstalk output; preferably, the subjective perception crosstalk outputs of different test images are weighted and averaged to obtain the subjective perception crosstalk of the display system.

[0029] A naked-eye 3D display system for measuring subjective crosstalk includes a computer-readable storage medium and a processor. The computer-readable storage medium stores an executable program, which, when executed by the processor, implements the method for measuring subjective crosstalk in a naked-eye 3D display system.

[0030] The present invention has the following beneficial effects:

[0031] This invention provides a method for measuring subjective crosstalk in naked-eye 3D display systems based on image features. By constructing a scoring crosstalk test image library and a feature scoring test image library, designing a subjective perception scoring scale, and establishing a functional relationship between image subjective scores and crosstalk values, high-precision detection of subjective perception crosstalk in naked-eye 3D display systems is achieved. This method utilizes image feature values ​​captured by a camera and trains a subjective score prediction model through regression prediction or time-series prediction algorithms, enabling quantitative measurement of subjective perception crosstalk of specific images on specific naked-eye 3D display devices. The fitting accuracy of the subjective perception-crosstalk fitting scoring function relationship of this invention reaches 97.8%, with a correlation of 97.37%, and the measurement accuracy of subjective perception using the feature image library reaches 94.6%, significantly improving the accuracy and reliability of crosstalk measurement. Furthermore, by establishing a feature-scoring-crosstalk-based subjective perception crosstalk measurement model, this method can analyze the influence of system parameters on subjective perception crosstalk, significantly improving the crosstalk evaluation effect of naked-eye 3D display systems, and has high practical value and good application prospects.

[0032] Other beneficial effects of the embodiments of the present invention will be further described below. Attached Figure Description

[0033] Figure 1 This is a flowchart of the overall scheme of the subjective perception crosstalk test system according to an embodiment of the present invention.

[0034] Figure 2 This is a flowchart of the subjective scoring experiment according to an embodiment of the present invention.

[0035] Figure 3 This is a flowchart of the scoring crosstalk function fitting process according to an embodiment of the present invention.

[0036] Figure 4 This is a flowchart of image preprocessing according to an embodiment of the present invention.

[0037] Figure 5 This is a flowchart of the feature scoring prediction model algorithm according to an embodiment of the present invention.

[0038] Figure 6 This is a flowchart of a method for measuring subjective crosstalk in a naked-eye 3D display system according to an embodiment of the present invention. Detailed Implementation

[0039] The embodiments of the present invention will be described in detail below. It should be emphasized that the following description is merely exemplary and is not intended to limit the scope and application of the present invention.

[0040] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of the present invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0041] See Figures 1 to 6 This invention provides a method for measuring subjective crosstalk in a naked-eye 3D display system based on image features, comprising the following steps:

[0042] Step S1: Construct a naked-eye 3D display device scoring crosstalk test library and a feature scoring test library respectively. The scoring crosstalk test library contains multiple sets of stereo image pairs with increasing image crosstalk rate gradients, and the feature scoring test library contains stereo image pairs without crosstalk and stereo image pairs with crosstalk.

[0043] In a preferred embodiment, in step S1, the scoring crosstalk test image library contains multiple sets of stereo image pairs with increasing image crosstalk rate gradients, wherein the crosstalk rate gradient range covers the maximum objective crosstalk rate range that the human eye can perceive as a stereo effect.

[0044] As an example, the crosstalk test image library for naked-eye 3D display devices can contain five sets of images. The crosstalk rate of each set of stereo image pairs is divided into 20 gradients from 0% to 50%. The crosstalk of the first 16 stereo image pairs increases by 2% from 0% to 30%, and the crosstalk of the 17th to 20th stereo image pairs increases by 5% from 30% to 50%. The feature scoring test image library for naked-eye 3D display devices can contain 235 stereo image pairs without crosstalk and 100 stereo image pairs with crosstalk.

[0045] Step S2: Design a subjective perception rating scale, and divide the rating into multiple levels based on subjective evaluation indicators of visual perception characteristics; load the rating crosstalk test image library on the naked-eye 3D display device to obtain the subjective rating of the naked-eye 3D image viewed by the human eye.

[0046] In a preferred embodiment, in step S2, the subjective evaluation indicators based on visual perception characteristics include the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen. As an example, the rating can be divided into 5 levels, where 5 points represents the best image quality and 1 point represents the worst image quality.

[0047] In a preferred embodiment, step S2, the process of obtaining the subjective score of the crosstalk test library for human eye viewing of naked-eye 3D display devices, includes:

[0048] The descriptions in the subjective rating scales were used as a comparative reference.

[0049] Observers view the images displayed by the naked-eye 3D display system and rate three indicators: the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen.

[0050] The subjective perception score of the image on the naked-eye 3D display system is obtained by calculating the weighted average of the scores for the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen.

[0051] For example, the weighted average score of the three metrics—ghosting, blur, and in / out depth—is calculated as follows:

[0052] Subjective perception score = ghosting score × 35% + blur × 35% + in-screen / out-of-screen depth × 30%.

[0053] Subjective feeling scores can be obtained based on the measurement results of multiple observers viewing the content displayed by the naked-eye 3D display system; preferably, subjective feeling scores are obtained based on the measurement results of each observer viewing the content displayed by the naked-eye 3D display system in different image sequences.

[0054] Step S3: Use the image crosstalk values ​​of the scoring crosstalk test library and the subjective ratings of the library obtained in step S2 to perform function fitting, and establish a functional relationship between the subjective ratings of the images and the image crosstalk values.

[0055] In a preferred embodiment, in step S3, the functional relationship established by fitting the image crosstalk value and the subjective score includes, but is not limited to, exponential functions, logarithmic functions, linear functions, and quadratic functions.

[0056] Preferably, the image crosstalk rate is respectively converted into a functional relationship with the subjective feeling score of each observer. Number fitting, then The average value is obtained by averaging the crosstalk-subjective perception score fitting curves of multiple observers. Crosstalk-subjective perception rating fitting function.

[0057] Step S4: Obtain the left and right monocular views of the naked-eye 3D display device after loading the feature scoring test image library, and fuse them into a binocular image; extract the image feature values, and perform feature concatenation on all feature values.

[0058] The camera can be a DSLR camera or a binocular CCD. When using a DSLR camera, left and right monocular images can be captured from two shooting positions respectively, based on the visual characteristics of the human eye. The camera is built based on the visual characteristics of the human eye, wherein the effective acquisition distance and lens angle are determined according to the optimal viewing distance and viewing angle range of the naked-eye 3D display device, and the camera spacing and height are set according to the average interocular distance and average height of the observer; and camera calibration and distortion correction are performed.

[0059] In a preferred embodiment, in step S4, the acquired left and right monocular views are first preprocessed, including image cropping, removing screen background and edges, etc., and the images with removed screen background and edges are subjected to noise reduction filtering and grayscale processing; then the processed monocular images are fused into binocular images, and feature values ​​are extracted from the monocular images and binocular images.

[0060] In a preferred embodiment, in step S4, the image feature values ​​include disparity feature values ​​between stereo image pairs and HOG feature values ​​of the fused binocular image. The feature concatenation includes merging the HOG feature values ​​of the binocular image with the global disparity difference and local disparity difference feature values ​​of the stereo image pairs into a feature matrix.

[0061] Step S5: Select several observers and load the feature rating test image library onto a naked-eye 3D display device to conduct a subjective experiment. Obtain the observers' subjective ratings of the images in the feature rating test image library as labels for training data. Use the subjective ratings obtained from the experiment and the cascaded feature value data obtained in step S4 to train a subjective rating prediction model through regression prediction or time series prediction algorithms until the error is reduced to a predetermined range. Establish a subjective feeling rating prediction model based on the measured feature values ​​obtained from the images captured by the camera. Predict the subjective feeling rating based on the image feature values ​​and determine the subjective feeling crosstalk based on the functional relationship between the rating and the function described in step S3.

[0062] In a preferred embodiment, step S5, the training process of the subjective rating prediction model includes:

[0063] Using regression prediction or time series prediction algorithms, the model is trained based on subjective ratings and concatenated feature value data from the feature rating test image library; a portion can be randomly selected as the training set to build the model, and the remaining portion as the test set.

[0064] By adjusting one or more parameters, such as the sampling ratio, regression parameters, and number of iterations, the error can be reduced to a predetermined range, thus completing model training.

[0065] In a preferred embodiment, in step S5, the subjective rating prediction model is trained using the SVR (Support Vector Regression) algorithm, and the mean squared error (MSE) is used as an indicator to measure the subjective feeling rating measurement results predicted by the model. Preferably, the SVR algorithm uses a grid search method with five-fold cross-validation as the optimization parameter method to obtain the best result.

[0066] In some embodiments, the process of measuring the subjective crosstalk of a naked-eye 3D display device using the obtained subjective crosstalk prediction model includes: loading a test image onto the display device under test, and extracting the display features of the test image using the method in step S4. The display features of the test image are input as test set data into the subjective rating prediction model to obtain a subjective rating output. Further, the subjective rating output is input as an independent variable into the rating crosstalk fitting model of the functional relationship in step S3 to obtain the image subjective crosstalk output. Preferably, the subjective crosstalk outputs of different test images are weighted and averaged to obtain the subjective crosstalk of the display system.

[0067] This invention also provides a naked-eye 3D display system for measuring subjective crosstalk, including a computer-readable storage medium and a processor. The computer-readable storage medium stores an executable program, which, when executed by the processor, implements the naked-eye 3D display system for measuring subjective crosstalk in any of the foregoing embodiments.

[0068] This invention establishes a method for measuring subjective crosstalk in naked-eye 3D display systems based on image features, achieving high-precision detection of crosstalk in naked-eye 3D display systems. This overcomes the shortcomings of existing technologies that rely solely on objective parameter measurements and cannot reflect the viewer's subjective experience. The method constructs a scoring crosstalk test image library and a feature scoring test image library, designs a subjective experience scoring scale, obtains subjective scores by combining human visual perception characteristics, and establishes a functional relationship between subjective scores and image crosstalk values ​​using function fitting. Furthermore, it extracts image feature values ​​through camera capture, trains a subjective score prediction model, and ultimately achieves subjective experience crosstalk prediction based on image feature values. This invention can not only quantitatively measure the subjective experience crosstalk of specific images on specific naked-eye 3D display devices, but also directly evaluate the crosstalk performance of display devices through model prediction. It features high precision (the fitting accuracy of the subjective experience-crosstalk scoring function relationship is 97.8%, and the correlation is 97.37%) and high efficiency (the subjective experience measurement accuracy of the feature image library is 94.6%), significantly improving the crosstalk evaluation effect of naked-eye 3D display systems and providing reliable technical support for optimizing display device performance.

[0069] The following further describes specific embodiments and experimental verifications of the present invention.

[0070] See Figures 1 to 6A method for measuring perceived crosstalk in a glasses-free 3D display system includes establishing a scoring crosstalk function relationship and a feature scoring prediction model for the glasses-free 3D display system. The main processes include:

[0071] 1. Based on the definition and principle of crosstalk, establish a scoring crosstalk test library and a feature scoring test library;

[0072] 2. Combine the characteristics of human visual perception with the features of 3D image libraries to obtain subjective evaluation scores for the image libraries;

[0073] 3. Based on the crosstalk values ​​added to the image library, the relationship between the score and the crosstalk is obtained through function fitting;

[0074] 4. Based on geometrical optics and diffractive optics theories, a binocular perception model is established. A series of image processing methods are used to obtain parallax and binocular perception feature values. Specifically, the following steps are included:

[0075] We extracted features of global disparity difference and local disparity difference from the preprocessed left and right images respectively, and obtained 4 feature values ​​for each pair of images;

[0076] The feature scoring test image library was extracted sequentially, resulting in 335 images. × A matrix of 4, stored in Excel;

[0077] Gabor texture features were extracted from the left and right images respectively to obtain the energy response maps of the images;

[0078] Use the SIFT matching operator to obtain the spacing between corresponding points in the left and right images;

[0079] By utilizing binocular fusion theory, the energy response maps of the left and right images are fused into a binocular image, which facilitates HOG feature extraction.

[0080] Gradient features of the binocular image are extracted using the HOG directional gradient algorithm;

[0081] Principal component analysis was used to reduce the dimensionality of HOG features, thereby improving the speed of subsequent model execution.

[0082] The feature scoring test image library was extracted sequentially, resulting in 335 images. × A matrix of 30 is stored in Excel;

[0083] Concatenating the disparity features and HOG features yields 335. × The feature matrix of 34 is stored in Excel.

[0084] 5. Compare the obtained perceptual crosstalk feature values ​​with the subjective scores to analyze the factors that affect the deviation between the subjective scores and the feature values.

[0085] Preferably, in step 1, five images that meet the experimental requirements are selected, and crosstalk is added at a gradient of 2%. When the crosstalk reaches 30%, crosstalk is added at a gradient of 5% to 50%.

[0086] Preferably, in step 2, 20 observers are selected to view the images individually in turn, with each observer's continuous viewing time not exceeding 25 minutes and a rest period of not less than 8 minutes. The process of obtaining subjective ratings for the image library includes the following steps, such as... Figure 2 As shown:

[0087] Preferably, a subjective preliminary experiment is conducted, including:

[0088] Select 3D images from the preliminary experiment;

[0089] Observers then proceeded to provide subjective ratings and evaluations using the given rating scale in turn.

[0090] Table 1

[0091]

[0092] Record each observer's subjective rating of the image;

[0093] Using statistical principles, 95% confidence intervals were calculated for different images, and candidates for the formal experiment were selected based on the results falling within the confidence intervals.

[0094] The image library was randomly shuffled for use in the formal experiment;

[0095] Following the principles of longest viewing time and shortest rest time, the selected observers were formally rated in the experiment.

[0096] Each observer is positioned at the optimal viewing distance for the glasses-free 3D display device;

[0097] Subjectively rate each image in the gallery using a rating scale;

[0098] When subjectively rating the scoring crosstalk test image library, observers can compare before and after images to improve the reliability of the rating.

[0099] Record each observer's subjective rating and perform weighted processing to obtain the subjective rating of the images in the image library.

[0100] Preferably, in step 3, a suitable function form is selected to fit the score value and crosstalk, including the following steps, such as... Figure 3 As shown:

[0101] A preliminary experiment on the goodness of fit of the obtained scoring data and crosstalk data was conducted. The nonlinear regression method of iterative least squares estimation was used to fit the crosstalk and the score into various function models. The fitting function models included linear models, multinomial models, logarithmic models, exponential models and other forms.

[0102] Compare the fitting coefficients R2 Based on the root mean square error (RMSE), select the optimal fit.

[0103] Using the best-fit relation, the best-fit relation was fitted to the data of each formal experimental observer, resulting in multiple fitting curves, fitting relation formulas and root mean square error (RMSE).

[0104] The fitted equation is averaged to obtain the final fitted curve and the root mean square error (RMSE).

[0105] Preferably, in step 4, based on the constructed feature scoring test image library, a binocular camera is used to acquire stereo image pairs. Images from the image library are displayed on a naked-eye 3D display screen, and then photographed. Image acquisition and preprocessing are required before feature extraction, such as... Figure 4 As shown, it specifically includes:

[0106] To capture images, an image acquisition device was built based on human visual characteristics. First, the effective acquisition distance was determined to be 5 meters, and the lens angle of view to be 60 degrees. Second, a binocular camera was configured according to the interpupillary distance and average human height.

[0107] Preferably, in step 4, after preprocessing the captured image, parallax features and HOG features are extracted, such as... Figure 5 The feature extraction unit, as shown, specifically includes the following steps:

[0108] Feature extraction of global disparity and local disparity is performed on the preprocessed left and right images respectively, and four feature values ​​are obtained for each pair of images;

[0109] The feature scoring test image library was extracted sequentially, resulting in 335 images. × A matrix of 4, stored in Excel;

[0110] Gabor texture features were extracted from the left and right images respectively to obtain the energy response maps of the images;

[0111] Use the SIFT matching operator to obtain the spacing between corresponding points in the left and right images;

[0112] By utilizing binocular fusion theory, the energy response maps of the left and right images are fused into a binocular image, which facilitates HOG feature extraction.

[0113] Gradient features of the binocular image are extracted using the HOG directional gradient algorithm;

[0114] Principal component analysis was used to reduce the dimensionality of HOG features, thereby improving the speed of subsequent model execution.

[0115] The feature scoring test image library was extracted sequentially, resulting in 335 images. × A matrix of 30 is stored in Excel;

[0116] Concatenating the disparity features and HOG features yields 335. × The feature matrix of 34 is stored in Excel.

[0117] Preferably, in step 5, the concatenated feature values ​​are used as input, and the subjective rating values ​​corresponding to the image are used as labels to train the regression prediction model. In this embodiment, SVR (Support Vector Regression) is used to evaluate the deviation between the subjective rating and the feature values. The specific steps are as follows:

[0118] Taking the feature scoring test image library as an example, the image library contains 335 image pairs as samples, and random sampling is performed on the 335 samples at different proportions.

[0119] A portion of the samples is extracted as a training set (feature data is the input to the training set, and subjective rating values ​​are the labels of the training set) to obtain a subjective rating prediction model with feature values ​​as input.

[0120] The remaining samples are used as a test set to evaluate and test the performance of the model obtained from the training set.

[0121] By adjusting relevant parameters (such as sampling ratio, regression parameters, cross-fold number, etc.), the mean squared error (MSE) can be minimized as much as possible. When the MSE is sufficiently small, the prediction result of the feature rating prediction model can be considered to be close enough to the subjective rating value.

[0122] Therefore, it can be divided into the following steps, such as Figure 5 The rating prediction unit is shown below:

[0123] Input the concatenated feature value data, run the SVR algorithm, and use deep learning methods to obtain the score prediction value.

[0124] Specifically, during the repeated running of the algorithm, the MSE and RMSE of the test set are used to measure the quality of the predicted scores. In order to obtain the smallest possible MSE and RMSE values, an algorithm for optimizing parameters should be selected for iteration.

[0125] Preferably, in this embodiment, a grid search algorithm is used to optimize the parameters of the SVR algorithm, calculate the optimal hyperparameters c and g, and optimize the parameters to avoid getting trapped in local optima.

[0126] Preferably, in this embodiment, five-fold cross-validation is used simultaneously to improve the randomness and reliability of the training and test sets;

[0127] Application Example 1

[0128] This invention is applied to predict crosstalk with high accuracy. The measurement system performs feature scoring prediction and scoring crosstalk function fitting on the constructed image database. The accuracy and MSE of the two models are shown in Table 2.

[0129] Table 2

[0130]

[0131] The feature scoring prediction model error comparison is based on subjective scoring, with a scoring scale of 5 points. The MSE is 0.2695, and the accuracy is 94.6%.

[0132] The correlation of the scoring crosstalk function fitting function was 97.37%, the error comparison was the crosstalk rate as a percentage, the RMSE was 2.1631%, and the accuracy reached 97.8%.

[0133] This invention also provides a storage medium for storing a computer program, which, when executed, performs at least the methods described above.

[0134] This invention also provides a control device, including a processor and a storage medium for storing a computer program; wherein the processor executes the computer program by performing at least the method described above.

[0135] This invention also provides a processor that executes a computer program, at least performing the methods described above.

[0136] The storage medium can be implemented by any type of non-volatile storage device, or a combination thereof. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); the magnetic surface memory can be a disk drive or magnetic tape drive. The storage media described in the embodiments of this invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0137] In the several embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and 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. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0138] The units described above as separate components may or may not be physically separate. 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 to achieve the purpose of this embodiment according to actual needs.

[0139] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0140] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0141] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0142] The methods disclosed in the several method embodiments provided by this invention can be arbitrarily combined without conflict to obtain new method embodiments.

[0143] The features disclosed in the several product embodiments provided by this invention can be arbitrarily combined without conflict to obtain new product embodiments.

[0144] The features disclosed in the several method or device embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0145] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various equivalent substitutions or obvious modifications can be made without departing from the concept of the present invention, and all such modifications, achieving the same performance or application, should be considered within the scope of protection of the present invention.

Claims

1. A method for measuring subjective crosstalk in a naked-eye 3D display system based on image features, characterized in that, Includes the following steps: S1: Construct a score crosstalk test library and a feature score test library for naked-eye 3D display devices, respectively. The score crosstalk test library contains multiple sets of stereo image pairs with increasing image crosstalk rate gradients, and the feature score test library contains stereo image pairs without crosstalk and stereo image pairs with crosstalk. S2: Design a subjective perception rating scale, and divide the rating into multiple levels based on subjective evaluation indicators of visual perception characteristics; load the rating crosstalk test image library on a naked-eye 3D display device to obtain the subjective rating of naked-eye 3D images viewed by the human eye. S3: Use the image crosstalk values ​​of the scoring crosstalk test library and the subjective ratings of the library obtained in step S2 to perform function fitting to establish a functional relationship between the subjective ratings of the images and the image crosstalk values. S4: Obtain the left and right monocular views captured by the camera after the naked-eye 3D display device loads the feature scoring test image library, and fuse them into a binocular image; extract image feature values ​​and concatenate all feature values; S5: Select several observers and load the feature rating test image library onto a naked-eye 3D display device to conduct a subjective experiment. Obtain the observers' subjective ratings of the images in the feature rating test image library as labels for training data. Use the subjective ratings obtained from the experiment and the cascaded feature value data obtained in step S4 to train a subjective rating prediction model through regression prediction or time series prediction algorithms until the error is reduced to a predetermined range. Establish a subjective feeling rating prediction model based on the measured feature values ​​obtained from the images captured by the camera. Predict the subjective feeling rating based on the image feature values ​​and determine the subjective feeling crosstalk based on the functional relationship between the rating and the function described in step S3.

2. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 1, characterized in that, In step S1, the scoring crosstalk test image library contains multiple sets of stereo image pairs with increasing image crosstalk rate gradients, and the crosstalk rate gradient range covers the maximum objective crosstalk rate range that the human eye can perceive as a stereo effect.

3. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 1 or 2, characterized in that, In step S2, the subjective evaluation indicators based on visual perception characteristics include the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen.

4. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 3, characterized in that, Step S2, the process of obtaining the subjective score of the crosstalk test library for human eye viewing of naked-eye 3D display devices, includes: The descriptions in the subjective rating scales were used as a comparative reference. Observers view the images displayed by the naked-eye 3D display system and rate three indicators: the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen. The subjective perception score of the image on the naked-eye 3D display system is obtained by calculating the weighted average of the scores for the severity of ghosting, the degree of image blurring, and the depth of image entry and exit from the screen.

5. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 1 or 2, characterized in that, In step S3, the functional relationship established by fitting the image crosstalk value and the subjective score is an exponential function, a logarithmic function, a linear function, or a quadratic function.

6. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 5, characterized in that, In step S3, the image crosstalk rate and the subjective feeling score of each observer are fitted with functions respectively, and then the crosstalk-subjective feeling score fitting curves of multiple observers are averaged to obtain the average crosstalk-subjective feeling score fitting function.

7. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 1 or 2, characterized in that, In step S4, the acquired left and right monocular views are preprocessed, including image cropping, removing screen background and edges, and performing noise reduction filtering and grayscale processing on the images with screen background and edges removed. The processed monocular images are then fused into a binocular image, and feature values ​​are extracted from the monocular and binocular images.

8. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 1 or 2, characterized in that, In step S4, the image feature values ​​include disparity feature values ​​between stereo image pairs and HOG feature values ​​of the fused binocular image. The feature concatenation includes merging the HOG feature values ​​of the binocular image with the global disparity difference and local disparity difference feature values ​​of the stereo image pairs into a feature matrix.

9. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 1 or 2, characterized in that, In step S5, the training process of the subjective rating prediction model includes: Using regression prediction or time series prediction algorithms, the model is trained based on subjective ratings and concatenated feature value data from the feature rating test image library; a portion can be randomly selected as the training set to build the model, and the remaining portion as the test set. By adjusting one or more parameters, such as the sampling ratio, regression parameters, and number of iterations, the error can be reduced to a predetermined range, thus completing model training.

10. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 9, characterized in that, In step S5, the subjective rating prediction model is trained using a regression prediction algorithm or a time series prediction algorithm, and the mean squared error (MSE) is used as an indicator to measure the model's prediction performance.

11. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 10, characterized in that, In step S5, the subjective rating prediction model is optimized using an optimization algorithm, including cross-validation, grid search algorithm, sparrow optimization algorithm, or genetic optimization algorithm.

12. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 1 or 2, characterized in that, In step S5, predicting the subjective perception crosstalk score based on image feature values ​​includes: loading the test image onto the display device under test, extracting the display features of the image under test using the method in step S4, inputting the display features of the image under test as test set data into the subjective score prediction model, and obtaining the subjective score output. Furthermore, the subjective rating output is used as an independent variable and input into the rating crosstalk fitting model based on the functional relationship in step S3 to obtain the image subjective perception crosstalk output.

13. The method for measuring subjective crosstalk in a naked-eye 3D display system as described in claim 12, characterized in that, The subjective perceived crosstalk outputs of different test images in the image library are weighted and averaged to obtain the subjective perceived crosstalk of the display system.

Images