A color migration driven X-ray enhanced image quality assessment method and system

Abstract

The present application belongs to the field of X-ray image quality assessment, and specifically discloses a color migration driven X-ray enhanced image quality assessment method and system. Through this application, the idea of bidirectional mapping is proposed, and two enhanced images are obtained by the two methods of "enhancement first and then migration" and "migration first and then enhancement". The accuracy of the assessment is guaranteed by the complementarity of the two enhanced images; this application simulates the visual perception of the human eye based on the Retinex theory, decomposes the image into reflection components and illumination components, and simulates the visual perception mechanism of the human eye to the image by performing comprehensive matching calculation, reflection component evaluation, and illumination component evaluation on the reflection components and illumination components of the original image and the enhanced image, so that the evaluation results can be closer to the subjective perception of human beings to image quality. Therefore, this application can quantitatively evaluate the quality of X-ray enhanced images in actual production scenarios without reference images or high-quality reference images, quantitatively evaluate the effects of various image enhancement algorithms, and provide technical support for selecting more suitable image enhancement algorithms.

Description

Technical Field

[0001] The present application belongs to the field of X-ray image quality assessment, and more specifically, relates to a color migration driven X-ray enhanced image quality assessment method and system. Background Art

[0002] Image quality assessment can be categorized into subjective and objective image quality assessment based on the evaluation subject. Subjective image quality assessment involves multiple domain experts evaluating a series of images, scoring the target image quality according to predefined evaluation criteria. The weighted average of the experts' scores is then used as the subjective score for the image. Therefore, subjective image quality assessment is time-consuming and labor-intensive, and the accuracy of the scoring is heavily dependent on the individual performance of the experts. Based on the degree of reliance on reference information, objective image quality assessment algorithms can be categorized as full-reference quality assessment, no-reference quality assessment, and semi-reference quality assessment. Full-reference quality assessment requires a complete reference image as a baseline, and uses metrics such as peak signal-to-noise ratio (PSNR), structural similarity, and feature similarity to perform a detailed assessment of image quality. However, in practical applications, high-quality reference images are often unavailable, making it impractical for unsupervised scenarios without reference images or supervised scenarios with low-quality reference images. No-reference quality assessment, which requires no reference images, is typically based on neural networks (such as CNNs and Res-Nets) for quality assessment, involving both training and testing. However, it is difficult to obtain a training set suitable for quality assessment, and each image requires subjective scoring by multiple experts, which is complex and time-consuming. In addition, the quality assessment dataset is different from other digital image datasets. The dataset can be expanded through image enhancement techniques such as rotation and cropping. The reason is that any transformation will affect the subjective perceptual quality score. Semi-reference quality assessment is a compromise method, which is between full reference and no-reference quality assessment. It mainly relies on a part of important features extracted from the reference image, such as texture features, color features, etc., rather than the entire reference image. Compared with full reference quality assessment, it has lower computational overhead and more accurate calculation results than no-reference quality assessment. However, this method requires determining in advance which features are most important to image quality. Different image types or distortion types may require reselection of features, and it is not applicable to real no-reference scenarios.

[0003] The aforementioned quality assessment methods are inadequate for real-world image quality assessment scenarios, which require extensive human effort, reference images, and reliable datasets. Furthermore, most existing quality assessment algorithms are applied to natural image fusion scenarios, lacking specialized quality assessment algorithms for evaluating the quality of X-ray inspection image enhancement. X-ray inspection technology is widely used in various applications, including nondestructive testing of castings. The grayscale range of the resulting DICOM inspection images is typically 16 bits (0–65535), while common display devices and common image formats (such as JPEG and PNG) can typically only process 8-bit grayscale (0–255). Furthermore, due to the complex structure and uneven thickness of the inspected castings, DICOM inspection images have low contrast, hindering the clear display of subtle details. Therefore, in practical applications, image enhancement algorithms such as windowing, denoising, tone mapping, and wavelet transforms are often required. However, these image enhancement theories and methods have varying effectiveness. Faced with the massive amount of enhanced images, how to quickly and accurately assess their quality remains an urgent challenge.

[0004] Therefore, there is an urgent need for efficient, low-cost, and high-quality quality assessment algorithms specifically for X-ray inspection images. Summary of the Invention

[0005] In response to the defects of the existing technology, the purpose of this application is to provide a color migration driven X-ray enhanced image quality assessment method and system, aiming to solve the problem that the existing quality assessment algorithms are all aimed at natural images and lack a dedicated quality assessment algorithm for evaluating the quality of X-ray detection image enhancement.

[0006] A first aspect of the present application relates to a color migration driven X-ray enhanced image quality assessment method, the image quality assessment method comprising:

[0007] S1. Divide the X-ray image into sub-images of equal size and perform the following processing on each sub-image:

[0008] Obtain a natural image with similar texture features to the sub-image, and perform grayscale mapping on the sub-image and the natural image respectively;

[0009] Using the enhancement algorithm to be evaluated to enhance the grayscale mapped sub-image, and migrating the color of the enhanced sub-image to the grayscale mapped natural image to obtain a first enhanced image;

[0010] Migrating the color of the grayscale mapped sub-image to the grayscale mapped natural image, and then enhancing the color-migrated natural image using the enhancement algorithm to be evaluated to obtain a second enhanced image;

[0011] Calculate the texture eigenvalues, visual perception eigenvalues and grayscale eigenvalues of the two enhanced images respectively, and calculate the mean;

[0012] The mean value of texture features, the mean value of visual perception features and the mean value of grayscale features are mapped into scores and then weighted to obtain the evaluation score of the sub-image;

[0013] S2. The average of the evaluation scores of all sub-images is used as the quality evaluation result of the enhancement algorithm to be evaluated for X-ray image enhancement.

[0014] In some embodiments, the grayscale mapping is a nonlinear grayscale mapping, and the grayscale distribution range of each sub-image is adjusted to 0-255.

[0015] In some embodiments, the color migration specifically includes:

[0016] Convert input and output images to CIELab color space;

[0017] Calculate the mean and variance of the three color channels of the output image and the input image in Lab space respectively;

[0018] Color migration is achieved through the following formula:

[0019]

[0020] in, For output image Channel value, , For the input image Channel value, For the target image Channel value, For the input image The standard deviation of the channel, For the target image The standard deviation of the channel, For the input image The mean of the channel, For the target image The mean of the channel.

[0021] In some embodiments, the calculation formula of the texture feature value is as follows:

[0022] Calculate the contrast of an image , correlation and energy :

[0023]

[0024]

[0025]

[0026] in, Grayscale value and grayscale values The possibility of appearing in a certain spatial relationship in the image, , , is the grayscale level of the image;

[0027] The contrast, correlation and energy of the image are calculated by weighted method to obtain the image texture feature value.

[0028] In some embodiments, the visual perception feature value The calculation formula is as follows:

[0029]

[0030]

[0031]

[0032]

[0033] in, It is used to measure the comprehensive matching degree between the decomposed reflection and illumination components of the enhanced image and the original image. Used to measure the difference between the enhanced image and the original image reflection component, Used to measure the difference between the illumination components of the enhanced image and the original image. , , , are weighted coefficients, To enhance the reflective component of the image, To enhance the illumination component of the image, To enhance the pixel value of the image; is the reflection component of the original image, is the lighting component of the original image, is the pixel value of the original image; is the matrix element-wise multiplication, is the L1 norm, is the number of image pixels, To enhance image pixels Light components The gradient, is a weighting factor based on the enhanced image pixels Reflective component gradient.

[0034] In some embodiments, the image is decomposed into a reflection component and an illumination component based on the Retinex theory to simulate human visual perception.

[0035] In some embodiments, the grayscale feature value is the information entropy of the image.

[0036] The second aspect of the present application relates to a color migration-driven X-ray enhanced image quality assessment system, which includes at least one processor and at least one memory; the at least one memory is used to store computer instructions; and the at least one processor is used to execute at least part of the computer instructions to implement the image quality assessment method described in any embodiment of the present application.

[0037] In general, the above technical solutions conceived by this application have the following beneficial effects compared with the existing technologies:

[0038] This application proposes a color migration-driven X-ray enhanced image quality assessment method and system, and proposes a bidirectional mapping idea. Two enhanced images are obtained by the two methods of "enhancement first, then migration" and "migration first, then enhancement". The accuracy of the assessment is guaranteed by the complementarity of the two enhanced images. This application simulates the visual perception of the human eye based on the Retinex theory, decomposes the image into reflection components and illumination components, and simulates the visual perception mechanism of the human eye to the image by performing comprehensive matching calculation, reflection component evaluation, and illumination component evaluation on the reflection components and illumination components of the original image and the enhanced image, so that the evaluation results can be closer to the subjective perception of human beings to image quality. Therefore, this application can quantitatively evaluate the quality of X-ray enhanced images in actual production scenarios without reference images or high-quality reference images, quantitatively evaluate the effects of various image enhancement algorithms, and provide technical support for selecting more suitable image enhancement algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a flow chart of a color migration driven X-ray enhanced image quality assessment method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0041] The term "and / or" as used herein describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. The symbol " / " as used herein indicates that the related objects are in an "or" relationship, for example, A / B means either A or B.

[0042] The terms "first" and "second" in this specification and claims are used to distinguish different objects rather than to describe a specific order of objects. For example, "first response message" and "second response message" are used to distinguish different response messages rather than to describe a specific order of response messages.

[0043] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0044] In the description of the embodiments of the present application, unless otherwise specified, "multiple" means two or more, for example, multiple processing units means two or more processing units, etc.; multiple elements means two or more elements, etc.

[0045] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application.

[0046] like Figure 1 As shown, the present application proposes a color migration driven X-ray enhanced image quality assessment method, which first obtains the X-ray detection image to be enhanced, and pre-processes its segmentation and grayscale range; secondly, through a public natural image dataset, obtains a natural image with texture features similar to the image to be evaluated, and performs grayscale mapping; then, based on the color migration method, adopts the idea of bidirectional mapping, and obtains two enhanced images by "enhancement first and then migration" and "migration first and then enhancement"; finally, designs a comprehensive evaluation method based on texture features, grayscale features and human visual perception to evaluate the image quality.

[0047] Obtain the X-ray detection image to be enhanced, and perform segmentation and grayscale range preprocessing as follows:

[0048] The X-ray image to be enhanced is in DICOM format, and is segmented to obtain a plurality of sub-images of the same size. In an embodiment shown, the sub-images are uniformly sized at 256*256.

[0049] Histogram equalization is used to perform nonlinear grayscale mapping on each sub-image, and the grayscale distribution range of each sub-image is adjusted to 0~255 to retain more details.

[0050] The specific steps of the histogram equalization are as follows:

[0051] 1) Count the number of pixels at each grayscale level in the original image (i.e., sub-image) to obtain the grayscale histogram of the original image

[0052] 2) Calculate the cumulative distribution function (CDF). CDF is the cumulative distribution of the grayscale histogram, which represents the proportion of pixels less than or equal to a certain grayscale value. The formula is as follows:

[0053]

[0054] in, Grayscale value The probability density function of the gray level The frequency of divided by the total number of pixels, is the current grayscale value, Indicates the number of all pixels.

[0055] 3) Normalize the cumulative distribution function and normalize the CDF to the range [0,1] to adapt it to the grayscale value range of 0 to 255. The formula is as follows:

[0056]

[0057] in, is the normalized CDF, is the minimum value of CDF, is the total number of pixels in the image.

[0058] 4) Map each grayscale value of the original image to a new grayscale value. The new grayscale value is obtained by multiplying the normalized CDF by 255:

[0059]

[0060] in, is the grayscale value after nonlinear grayscale mapping.

[0061] 5) Generate an equalized image: Using the above mapping relationship, convert the grayscale value of each pixel in the original image to a new grayscale value to obtain an image after grayscale mapping.

[0062] Using a public natural image dataset, obtain natural images with texture features similar to the image to be evaluated and perform grayscale mapping as follows:

[0063] The quality of X-ray images primarily depends on their grayscale distribution and texture. Therefore, natural images with textures similar to those of the image being evaluated are selected to preserve these features as much as possible. Grayscale mapping of natural color images, such as using histogram equalization, is performed to approximate the properties of X-ray images.

[0064] The calculation and representation of texture features will be introduced later and will not be repeated here.

[0065] Methods for measuring texture feature similarity include but are not limited to: Euclidean distance, correlation coefficient, and cosine similarity.

[0066] This application is based on the color migration method and adopts the idea of bidirectional mapping. Two enhanced images are obtained through the two methods of "enhancement first, then migration" and "migration first, then enhancement". Specifically:

[0067] “Enhance first, then transfer” means that the X-ray image to be enhanced is enhanced first, and then the grayscale distribution of the X-ray enhanced image is mapped to the natural image through color transfer technology to obtain the first enhanced image.

[0068] “Migration first, then enhancement” means first mapping the grayscale distribution of the X-ray image to the natural image, and then enhancing the natural image to obtain a second enhanced image.

[0069] It should be noted that if the grayscale distribution of image A is mapped to image B through color transfer technology to obtain image C, then A is the input image and B is the target image. Regardless of whether it is migrated first and then enhanced or enhanced first and then migrated, the target image B is a natural image and C is the output image.

[0070] Specifically, the purpose of color migration is to transfer the grayscale distribution characteristics of the image to be evaluated to a natural image. The steps are as follows:

[0071] First, perform color space conversion, converting the image from RGB space to CIELab color space. The channels in Lab space are independent (L is the brightness channel, representing the brightness information of the image, a represents the color component from green to red, and b represents the color component from blue to yellow). However, there is a certain correlation between the color channels in RGB space, which may affect the color migration effect.

[0072] Next, we calculate the mean and variance of the three color channels of the target and input images in Lab space. The mean represents the average brightness of the channel, representing the overall details of the image and is used to adjust the overall color. The variance represents the distribution range of the channel color, representing the image's hue and contrast, and is used to control the range of color expansion.

[0073] Then, color migration is achieved by the following formula:

[0074]

[0075] in, For output image Channel value, , For the input image Channel value, For the target image Channel value, For the input image The standard deviation of the channel, For the target image The standard deviation of the channel, For the input image The mean of the channel, For the target image The mean of the channel.

[0076] In the above formula, the ratio of standard deviations , adjust the color contrast of the input image to make it consistent with the target image; : Align the center of the color to 0 by subtracting the mean of the input image; add the mean of the target image , complete the alignment of the color center position.

[0077] Due to the particularity of the X-ray image structure, it is extremely difficult to find a clear image with a similar structure to the input image. Therefore, this application adopts the idea of bidirectional mapping to construct two different enhanced images, compare them with the input image respectively, and then confirm the two comparison results with each other to obtain the final result. The processing of the constructed reference image can be divided into two steps: "enhancement first, then migration" and "migration first, then enhancement". The former can highlight the grayscale consistency and detail enhancement effect of the X-ray image, and the latter starts from the natural image and can emphasize the texture and structural characteristics of the X-ray image, which is closer to the enhancement effect in the subjective perception of the human eye. The accuracy of the evaluation is guaranteed by the complementarity of the two enhanced images.

[0078] Traditional X-ray image quality assessment primarily evaluates image quality based on grayscale and texture features. While this provides a quantitative basis, it often overlooks the complexity of human visual perception and the subjective experience of images. Based on Retinex theory, the human visual system perceives X-ray images based on not only grayscale and texture but also reflective and illumination components. By separating image illumination changes from the inherent characteristics of an object, the visual system can more reliably perceive the object's reflective information.

[0079] Human visual perception characteristics

[0080] Based on the Retinex theory to simulate human visual perception, the image is decomposed into the grayscale features of the reflection component and the illumination component, as follows:

[0081] 1) Decomposition of input image

[0082] Assume that the input image The reflection component and light components The product of , namely:

[0083]

[0084] 2) Multi-scale Gaussian filtering

[0085] In order to separate the illumination component and the reflection component, a multi-scale filtering technique is used. By performing multi-scale Gaussian filtering on the image, illumination information at different scales is obtained.

[0086] 3) Calculation of illumination components

[0087] Convert the input image to the logarithmic domain and then calculate the illumination component and the reflection component

[0088]

[0089] in, is an image processed by Gaussian filtering, representing large-scale lighting information.

[0090] 4) Restoration of reflection components

[0091]

[0092] This is equivalent to performing a logarithmic transformation on the input image, subtracting the illumination component, and then performing exponential restoration.

[0093] Based on the Retinex theory, the human visual perception is simulated, and the image is decomposed into reflection components and illumination components. The visual perception feature value is obtained by calculating the reflection components and illumination components of the evaluation image and the reference image. , used to quantify the difference in image perception in human eyes.

[0094]

[0095] in, Used to measure the comprehensive matching degree between the decomposed reflection and illumination components of the X-ray image and the reference image:

[0096]

[0097] Used to measure the difference between the reflection components of the reference image and the image to be evaluated:

[0098]

[0099] Used to measure the difference between the illumination components of the reference image and the image to be evaluated:

[0100]

[0101] in, , , , are weighted coefficients, To enhance the reflective component of the image, To enhance the illumination component of the image, To enhance the pixel value of the image; is the reflection component of the original image, is the lighting component of the original image, is the pixel value of the original image; is the matrix element-wise multiplication, is the L1 norm, is the number of image pixels, Light component The gradient, is a weighting factor based on the reflection component In one illustrated embodiment, and Both are 0.1, and Both are 0.01, is 1.

[0102] Grayscale features

[0103] Grayscale features can be represented by a variety of indicators, including but not limited to: image grayscale value variance, standard deviation, information entropy and two-dimensional entropy. This embodiment measures the amount of information in grayscale distribution by calculating the information entropy of the X-ray image. Information entropy The calculation formula is as follows:

[0104]

[0105] in, Grayscale value The probability value of occurrence can be obtained from the grayscale histogram.

[0106] Texture features

[0107] Texture features can be represented by a variety of indicators, including but not limited to: gray level co-occurrence matrix features, Tamura texture features and local binary patterns. This embodiment uses gray level co-occurrence matrix features to measure the spatial correlation between pixels. Represents the gray-level co-occurrence matrix, which is a The matrix ( Grayscale, which is the number of different grayscales or colors contained in a picture), elements Describes the grayscale value and grayscale values The possibility of appearing in a certain spatial relationship in an image.

[0108] The contrast, correlation, and energy of the image are calculated using the gray-level co-occurrence matrix as follows:

[0109]

[0110]

[0111]

[0112] The contrast, correlation and energy of the image are calculated by weighted method to obtain the image texture feature value.

[0113] The mean of the texture feature values, visual perception feature values and grayscale feature values of the two enhanced images are calculated, and the texture feature mean, visual perception feature mean and grayscale feature mean are mapped to scores and then weighted to obtain the evaluation score of the sub-image.

[0114] In some embodiments, the texture feature mean, visual perception feature mean, and grayscale feature mean are mapped to scores respectively through scoring, so as to achieve comparability of values between different features.

[0115] In some embodiments, the texture feature mean, the visual perception feature mean, and the grayscale feature mean are mapped to scores through normalization processing to achieve comparability of values between different features.

[0116] Different weights are assigned to different features based on actual needs. If the complexity of grayscale distribution is important, information entropy may be given a higher weight. If the inspection target involves texture changes (such as cracks or defect edges), texture features are given a higher weight. If the inspection target is highly relevant to visual perception, the importance of these features should be increased.

[0117] The two enhanced images obtained by the above method and the evaluation image are used to calculate the corresponding features, and a specific accuracy score is obtained based on the corresponding weights.

[0118] In this application, all weights can be set manually or optimized by particle swarm optimization.

[0119] The average of the evaluation scores of all sub-images is used as the quality evaluation result of the enhancement algorithm to be evaluated for X-ray image enhancement. By comparing the quality evaluation results of different enhancement algorithms for X-ray image enhancement, the most effective enhancement algorithm is selected.

[0120] It should be understood that the above-mentioned device is used to execute the method in the above-mentioned embodiment. The implementation principle and technical effect of the corresponding program module in the device are similar to those described in the above-mentioned method. The working process of the device can refer to the corresponding process in the above-mentioned method and will not be repeated here.

[0121] Based on the method in the above embodiment, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method in the above embodiment.

[0122] Based on the method in the above embodiment, an embodiment of the present application provides a computer program product. When the computer program product runs on a processor, the processor executes the method in the above embodiment.

[0123] It is understood that the processor in the embodiments of the present application may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0124] The method steps in the embodiments of the present application can be implemented by hardware or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, mobile hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and storage medium can be located in an ASIC.

[0125] The above embodiments can be implemented in whole or in part using software, hardware, firmware, or any combination thereof. When implemented using software, they can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions. When loaded and executed on a computer, the computer program instructions fully or partially produce the processes or functions described in the embodiments of this application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted via the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be magnetic media (e.g., floppy disk, hard disk, tape), optical media (e.g., DVD), or semiconductor media (e.g., solid-state drive (SSD)).

[0126] It will be understood that the various numerical numbers involved in the embodiments of the present application are merely distinctions for the convenience of description and are not intended to limit the scope of the embodiments of the present application.

[0127] It is easy for those skilled in the art to understand that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A color migration driven X-ray enhanced image quality assessment method, characterized in that: include: S1. Divide the X-ray image into sub-images of equal size and perform the following processing on each sub-image: Obtain a natural image with similar texture features to the sub-image, and perform grayscale mapping on the sub-image and the natural image respectively; Using the enhancement algorithm to be evaluated to enhance the grayscale mapped sub-image, and migrating the color of the enhanced sub-image to the grayscale mapped natural image to obtain a first enhanced image; Migrating the color of the grayscale mapped sub-image to the grayscale mapped natural image, and then enhancing the color-migrated natural image using the enhancement algorithm to be evaluated to obtain a second enhanced image; Calculate the texture eigenvalues, visual perception eigenvalues and grayscale eigenvalues of the two enhanced images respectively, and calculate the mean; The mean value of texture features, the mean value of visual perception features and the mean value of grayscale features are mapped into scores and then weighted to obtain the evaluation score of the sub-image; S2. The average of the evaluation scores of all sub-images is used as the quality evaluation result of the enhancement algorithm to be evaluated for X-ray image enhancement.

2. The image quality assessment method according to claim 1, wherein: The grayscale mapping is nonlinear, and the grayscale distribution range of each sub-image is adjusted to 0-255.

3. The image quality assessment method according to claim 1, wherein: The color migration specifically includes: Convert input and output images to CIELab color space; Calculate the mean and variance of the three color channels of the output image and the input image in Lab space respectively; Color migration is achieved through the following formula: in, For output image Channel value, , For the input image Channel value, For the target image Channel value, For the input image The standard deviation of the channel, For the target image The standard deviation of the channel, For the input image The mean of the channel, For the target image The mean of the channel.

4. The image quality assessment method according to claim 1, wherein: The calculation formula of the texture feature value is as follows: Calculate the contrast of an image , correlation and energy : in, Grayscale value and grayscale values The possibility of appearing in a certain spatial relationship in the image, , , is the grayscale level of the image; The contrast, correlation and energy of the image are calculated by weighted method to obtain the image texture feature value.

5. The image quality assessment method according to claim 1, wherein: The visual perception feature value The calculation formula is as follows: in, It is used to measure the comprehensive matching degree between the decomposed reflection and illumination components of the enhanced image and the original image. Used to measure the difference between the enhanced image and the original image reflection component, Used to measure the difference between the illumination components of the enhanced image and the original image. , , , are weighted coefficients, To enhance the reflective component of the image, To enhance the illumination component of the image, To enhance the pixel value of the image; is the reflection component of the original image, is the lighting component of the original image, is the pixel value of the original image; is the matrix element-wise multiplication, is the L1 norm, is the number of image pixels, To enhance image pixels Light components The gradient, is a weighting factor based on the enhanced image pixels Reflective component gradient.

6. The image quality assessment method according to claim 5, wherein: Based on the Retinex theory, the human visual perception is simulated and the image is decomposed into reflection component and illumination component.

7. The image quality assessment method according to claim 1, wherein: The grayscale feature value is the information entropy of the image.

8. A color migration driven X-ray enhanced image quality assessment system, characterized in that: comprising at least one processor and at least one memory; The at least one memory is configured to store computer instructions; The at least one processor is configured to execute at least part of the computer instructions to implement the image quality assessment method according to any one of claims 1 to 7.

Images