Image quality evaluation method and device, electronic equipment and storage medium
By preprocessing the image and fusion of feature parameters, combined with machine learning algorithms, the comprehensive and accurate image quality evaluation is achieved, the problem of difficult to reflect image quality information in the existing technology is solved, and targeted and practical Evaluation method.
Patent Information
- Application Number
- CN202411272791.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2024-12-03
AI Technical Summary
Existing reference-free image quality evaluation methods are difficult to fully reflect image quality information, and require comparison evaluation based on original or standard images, and it is difficult to obtain these reference images in applications.
By preprocessing the target image, feature parameters are extracted and fusion processed, comprehensive feature vectors are generated, quality evaluation models are trained to obtain image quality scores, and feedback information is generated based on the scores. The method includes steps such as denoising and enhancement processing, feature parameter extraction and fusion, principal component analysis, etc., and is trained using machine learning algorithms to adapt to images of different types and quality.
It realizes the comprehensive and accurate evaluation of image quality without relying on original or standard images, which is highly adaptable and practical, and can provide effective guidance for image quality improvement.
Smart Images

Figure CN119068317A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of image transmission technology, and in particular to an image quality assessment method and device, an electronic device, and a storage medium. Background Art
[0002] With the rapid development of information technology, graphics and images, as important carriers of information, have been widely used in various fields. However, since images may be affected by various factors during acquisition, transmission and storage, image quality may deteriorate, affecting the transmission and understanding of information. Therefore, evaluating the quality of graphics and images and providing feedback information are of great significance for improving image quality and optimizing image processing algorithms.
[0003] At present, image quality assessment methods are mainly divided into two categories: reference and no-reference. Reference methods need to rely on original images or standard images for comparative evaluation, but in practical applications, it is often difficult to obtain original or standard images. Therefore, no-reference image quality assessment methods have become a hot topic of research. However, most of the existing no-reference image quality assessment methods are based on the statistical characteristics of images, which are difficult to fully reflect the quality information of images. Summary of the invention
[0004] The purpose of the present disclosure is to overcome the deficiencies of the prior art and to provide an image quality assessment method and device, an electronic device and a storage medium. The image quality assessment method can comprehensively assess image quality and provide accurate feedback information.
[0005] According to a first aspect of an embodiment of the present disclosure, there is provided an image quality assessment method, the method comprising:
[0006] Preprocess the target image;
[0007] Extracting characteristic parameters of the preprocessed target image;
[0008] The characteristic parameters are fused to obtain a comprehensive characteristic vector; wherein the fusion process includes at least one of weighted average and principal component analysis;
[0009] Training a quality assessment model, and inputting the comprehensive feature vector into the trained quality assessment model to obtain a quality score of the target image;
[0010] According to the quality score of the target image, corresponding feedback information is generated.
[0011] In one embodiment, the preprocessing of the target image includes:
[0012] Performing denoising processing on the target image;
[0013] The target image after denoising is enhanced.
[0014] In one embodiment, extracting characteristic parameters of the preprocessed target image includes:
[0015] Converting the preprocessed target image into a grayscale image;
[0016] Calculating a pixel average value of the grayscale image, wherein the pixel average value is used to feedback an average brightness of the image;
[0017] Obtaining the grayscale values of pixels of the grayscale image, and calculating the degree of dispersion of the pixel grayscale values relative to the pixel average value, wherein the degree of dispersion is used to feedback the dispersion degree of the grayscale level of the image;
[0018] Calculating detail contrast and texture transformation degree in the target image, wherein the detail contrast and texture transformation degree are used to provide feedback on the clarity of the image;
[0019] The average amount of information of the target image is calculated, and the average amount of information is used to feedback the amount of information in the image.
[0020] In one embodiment, fusing the feature parameters to obtain a comprehensive feature vector includes:
[0021] Write feature fusion function;
[0022] The feature parameters are input into the feature fusion function to calculate and obtain the comprehensive feature vector.
[0023] In one embodiment, the feedback information includes at least one of a quality level of the target image, possible problems of the target image, and improvement suggestions for the target image.
[0024] According to a second aspect of an embodiment of the present disclosure, there is provided an image quality assessment device, the device comprising: a preprocessing module, an extraction module, a fusion module, a training module and a generation module; wherein:
[0025] A preprocessing module is used to preprocess the target image;
[0026] An extraction module, extracting characteristic parameters of the preprocessed target image;
[0027] A fusion module, which performs fusion processing on the feature parameters to obtain a comprehensive feature vector; wherein the fusion processing includes at least one of weighted average and principal component analysis;
[0028] A training module, training a quality assessment model, and inputting the comprehensive feature vector into the trained quality assessment model to obtain a quality score of the target image;
[0029] A generation module generates corresponding feedback information according to the quality score of the target image.
[0030] In one embodiment, the preprocessing module includes a denoising submodule and an enhancement submodule.
[0031] A denoising submodule, performing denoising processing on the target image;
[0032] The enhancement submodule performs enhancement processing on the target image after the denoising processing.
[0033] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, including: a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any of the above methods when executing the computer program.
[0034] According to a fourth aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above methods are implemented.
[0035] In an image quality assessment method provided by an embodiment of the present disclosure, the quality of an input graphic image can be directly assessed without relying on an original or standard image as a reference. Specifically, by extracting a variety of feature parameters and fusing the multiple feature parameters, the image quality assessment method in the present disclosure can more comprehensively and accurately reflect the quality information of the image, that is, it can be fully assessed; by training with a machine learning algorithm, the image quality assessment method in the present disclosure can adapt to images of different types and qualities, and has a strong generalization ability, that is, it has strong adaptability; by generating corresponding feedback information based on the quality score of the target image, the image quality assessment method in the present disclosure can be targeted and practical, thereby providing effective guidance for improving image quality, that is, it has strong practicality. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.
[0037] Figure 1 A flowchart of an image quality assessment method provided by an embodiment of the present disclosure.
[0038] Figure 2 A flowchart of an image quality assessment method provided by an embodiment of the present disclosure.
[0039] Figure 3 A logical schematic diagram of an image quality assessment method provided by an embodiment of the present disclosure.
[0040] Figure 4 An architectural diagram of an image quality assessment device provided in an embodiment of the present disclosure.
[0041] Figure 5 An architectural diagram of an image quality assessment device provided in an embodiment of the present disclosure.
[0042] Figure 6 An internal structure diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION
[0043] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.
[0044] Figure 1 The following is a flow chart of an image quality assessment method provided by an embodiment of the present disclosure. Figure 1 As shown, the method includes:
[0045] Step 101, preprocessing the target image;
[0046] In one embodiment, the preprocessing of the target image includes:
[0047] Performing denoising processing on the target image;
[0048] The target image after denoising is enhanced.
[0049] In this embodiment, the input graphic image is subjected to denoising and enhancement operations to improve the accuracy of image quality assessment. Specifically, in this embodiment, methods such as median filtering and Gaussian filtering can be used to denoise the image to reduce the impact of noise on image quality assessment; according to the actual situation of the image, methods such as contrast enhancement and brightness enhancement can be used to enhance the image to improve the clarity and contrast of the image.
[0050] Step 102: extracting characteristic parameters of the preprocessed target image;
[0051] In one embodiment, Figure 2 As shown, the feature parameters of the target image after extraction preprocessing include:
[0052] Step 201, converting the preprocessed target image into a grayscale image;
[0053] Exemplarily, the pre-processed target image can be converted into a grayscale image using the rgb2gray function of MATLAB. MATLAB is a commercial mathematical software produced by MathWorks, USA, used in data analysis, wireless communication, deep learning, image processing and computer vision, signal processing, quantitative finance and risk management, robotics, control systems and other fields. gb2gray is a function for processing images in matlab. It converts RGB images or color images into grayscale images by eliminating image hue and saturation information while retaining brightness, that is, the grayscale processing function. The format of calling this function is I=rgb2gray(RGB), which means converting the true color image RGB into a grayscale intensity image I.
[0054] Step 202, calculating the pixel average value of the grayscale image, wherein the pixel average value is used to feedback the average brightness of the image;
[0055] Step 203, obtaining the pixel grayscale values of the grayscale image, and calculating the degree of discreteness of the pixel grayscale values relative to the pixel average value, wherein the degree of discreteness is used to feedback the dispersion degree of the grayscale level of the image; it should be noted that the technical means for obtaining the pixel grayscale values of the grayscale image and calculating the degree of discreteness of the pixel grayscale values relative to the pixel average value in this embodiment already have very mature technical solutions in the prior art, and therefore, the embodiments of the present disclosure will not be described in detail.
[0056] Step 204: Calculate the detail contrast and texture transformation degree in the target image, and the detail contrast and texture transformation degree are used to feedback the clarity of the image; it should be noted that the technical means for calculating the detail contrast and texture transformation degree in the target image in this embodiment have already been very mature technical solutions in the prior art, so the embodiments of this disclosure will not be repeated. Step 205: Calculate the average amount of information of the target image, and the average amount of information is used to feedback the amount of information in the image. It should be noted that the technical means for calculating the average amount of information of the target image in this embodiment have already been very mature technical solutions in the prior art, so the embodiments of this disclosure will not be repeated.
[0057] In this embodiment, multiple feature parameters are extracted from the preprocessed image, including the mean, standard deviation, average gradient, entropy, etc. of the image. These feature parameters can comprehensively reflect the quality information of the image.
[0058] Step 103: fusing the characteristic parameters to obtain a comprehensive characteristic vector; wherein the fusing process includes at least one of weighted averaging and principal component analysis;
[0059] Exemplarily, the present implementation may fuse the feature parameters into a comprehensive feature vector by using a weighted average method.
[0060] In one embodiment, fusing the feature parameters to obtain a comprehensive feature vector includes:
[0061] Write feature fusion function;
[0062] The feature parameters are input into the feature fusion function to calculate and obtain the comprehensive feature vector.
[0063] Step 104: training a quality assessment model, and inputting the comprehensive feature vector into the trained quality assessment model to obtain a quality score of the target image;
[0064] In one embodiment, the training quality assessment model includes:
[0065] Create a dataset containing images of various quality levels;
[0066] Extracting and fusing features of the data set containing images of various quality levels to obtain a comprehensive feature vector of each image;
[0067] The quality assessment model is trained using a machine learning algorithm; wherein the machine learning algorithm includes at least one of a support vector machine and a neural network.
[0068] In this embodiment, a data set containing images of various quality levels is prepared, and feature extraction and fusion are performed on the data set to obtain a comprehensive feature vector for each image. Machine learning algorithms such as support vector machines (SVM) and neural networks are used to train a quality assessment model, and the trained model is saved for subsequent use.
[0069] It should be noted that Support Vector Machine (SVM) is a generalized linear classifier that performs binary classification of data in a supervised learning manner. Its decision boundary is the maximum-margin hyperplane for solving learning samples. SVM uses the hinge loss function to calculate empirical risk and adds regularization terms to the solution system to optimize structural risk. It is a classifier with sparsity and robustness. SVM can perform nonlinear classification through the kernel method and is one of the common kernel learning methods.
[0070] Step 105: Generate corresponding feedback information according to the quality score of the target image.
[0071] In one embodiment, the feedback information includes at least one of a quality level of the target image, possible problems of the target image, and improvement suggestions for the target image.
[0072] In this embodiment, after the corresponding feedback information is generated according to the quality score of the target image, it will be displayed to the user.
[0073] It should be noted that the embodiments of the present disclosure can optimize and improve the algorithm according to actual conditions, such as adjusting denoising parameters, adding other feature parameters and improving feature fusion methods, so as to improve the performance and accuracy of the algorithm.
[0074] It should be further explained that the embodiments of the present disclosure are based on programming environments such as MATLAB or Python, and implement the reference-free graphic image quality assessment and feedback algorithm through the following steps. Python: Python was designed by Guido van Rossum of the Netherlands National Research Center for Mathematics and Computer Science in the early 1990s as a substitute for a language called ABC. Python provides efficient high-level data structures and simple and effective object-oriented programming. Python syntax and dynamic typing, as well as the nature of an interpreted language, make it a programming language for writing scripts and rapidly developing applications on most platforms. With the continuous update of versions and the addition of new language functions, it is gradually being used for the development of independent, large-scale projects.
[0075] In an image quality assessment method provided by an embodiment of the present disclosure, the quality of an input graphic image can be directly assessed without relying on an original or standard image as a reference. Specifically, by extracting a variety of feature parameters and fusing the multiple feature parameters, the image quality assessment method in the present disclosure can more comprehensively and accurately reflect the quality information of the image, that is, it can be fully assessed; by training with a machine learning algorithm, the image quality assessment method in the present disclosure can adapt to images of different types and qualities, and has a strong generalization ability, that is, it has strong adaptability; by generating corresponding feedback information based on the quality score of the target image, the image quality assessment method in the present disclosure can be targeted and practical, thereby providing effective guidance for improving image quality, that is, it has strong practicality.
[0076] Figure 3 A logic diagram of an image quality assessment method provided in an embodiment of the present disclosure.
[0077] Figure 4 This is a diagram of the architecture of an image quality assessment device provided by an embodiment of the present disclosure. Figure 4As shown, the device includes a preprocessing module 401, an extraction module 402, a fusion module 403, a training module 404 and a generation module 405; wherein the preprocessing module 401 is used to preprocess the target image; the extraction module 402 is used to extract the characteristic parameters of the preprocessed target image; the fusion module 403 is used to fuse the characteristic parameters to obtain a comprehensive characteristic vector; wherein the fusion processing includes at least one of weighted averaging and principal component analysis; the training module 404 is used to train a quality assessment model, and input the comprehensive characteristic vector into the trained quality assessment model to obtain a quality score of the target image; the generation module 405 is used to generate corresponding feedback information according to the quality score of the target image.
[0078] Figure 5 This is a diagram of the architecture of an image quality assessment device provided by an embodiment of the present disclosure. Figure 5 As shown, the device includes a preprocessing module 501, an extraction module 502, a fusion module 503, a training module 504 and a generation module 505; wherein the preprocessing module 501 includes a denoising submodule 5011 and an enhancement submodule 5012, wherein the denoising submodule 5011 is used to perform denoising processing on the target image; and the enhancement submodule 5012 is used to perform enhancement processing on the target image after denoising processing.
[0079] The present disclosure also provides an electronic device, the internal structure diagram of the electronic device can be as follows Figure 6 As shown. The electronic device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the electronic device is used to store data. The network interface of the electronic device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it is implemented as in the above-mentioned stress test method. It includes: a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, it implements any step in the above-mentioned image quality assessment method.
[0080] The present disclosure also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, any step in the above-mentioned image quality assessment can be implemented.
[0081] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the disclosure disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.
[0082] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims
1. A method for evaluating image quality, characterized in that: The method comprises: Preprocess the target image; Extracting characteristic parameters of the preprocessed target image; The characteristic parameters are fused to obtain a comprehensive characteristic vector; wherein the fusion process includes at least one of weighted average and principal component analysis; Training a quality assessment model, and inputting the comprehensive feature vector into the trained quality assessment model to obtain a quality score of the target image; According to the quality score of the target image, corresponding feedback information is generated.
2. The method according to claim 1, characterized in that The preprocessing of the target image comprises: Performing denoising processing on the target image; The target image after denoising is enhanced.
3. The method according to claim 1 or 2, characterized in that: The extracting of the characteristic parameters of the target image after preprocessing comprises: Converting the preprocessed target image into a grayscale image; Calculating a pixel average value of the grayscale image, wherein the pixel average value is used to feedback an average brightness of the image; Obtaining the grayscale values of pixels of the grayscale image, and calculating the degree of dispersion of the pixel grayscale values relative to the pixel average value, wherein the degree of dispersion is used to feedback the dispersion degree of the grayscale level of the image; Calculating detail contrast and texture transformation degree in the target image, wherein the detail contrast and texture transformation degree are used to provide feedback on the clarity of the image; The average amount of information of the target image is calculated, and the average amount of information is used to feedback the amount of information in the image.
4. The method according to claim 1 or 2, characterized in that: The fusing of the feature parameters to obtain a comprehensive feature vector comprises: Write feature fusion function; The feature parameters are input into the feature fusion function to calculate and obtain the comprehensive feature vector.
5. The method according to claim 1 or 2, characterized in that: The feedback information includes at least one of a quality level of the target image, possible problems of the target image, and improvement suggestions for the target image.
6. The method according to claim 1 or 2, characterized in that: The training quality assessment model includes: Create a dataset containing images of various quality levels; Extracting and fusing features of the data set containing images of various quality levels to obtain a comprehensive feature vector of each image; The quality assessment model is trained using a machine learning algorithm; wherein the machine learning algorithm includes at least one of a support vector machine and a neural network.
7. An image quality assessment device, characterized in that: The device comprises: a preprocessing module, an extraction module, a fusion module, a training module and a generation module; wherein, A preprocessing module is used to preprocess the target image; An extraction module, extracting characteristic parameters of the preprocessed target image; A fusion module, which performs fusion processing on the feature parameters to obtain a comprehensive feature vector; wherein the fusion processing includes at least one of weighted average and principal component analysis; A training module, training a quality assessment model, and inputting the comprehensive feature vector into the trained quality assessment model to obtain a quality score of the target image; A generation module generates corresponding feedback information according to the quality score of the target image.
8. The device according to claim 7, characterized in that The preprocessing module includes a denoising submodule and an enhancement submodule. A denoising submodule, performing denoising processing on the target image; The enhancement submodule performs enhancement processing on the target image after the denoising processing.
9. An electronic device, comprising: A memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the method according to any one of claims 1 to 6 when executing the computer program.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
