Non-reference based magnetic resonance image quality assessment method, apparatus, and readable medium

Abstract

The application discloses a non-reference-based magnetic resonance image quality evaluation method and device and readable medium, and relates to an image processing module, comprising: constructing a non-reference-based magnetic resonance image quality evaluation model and training; acquiring a magnetic resonance image to be evaluated and inputting the magnetic resonance image to the trained magnetic resonance image quality evaluation model, the magnetic resonance image is first input to an image estimation module, sequentially passes through a plurality of first convolution layers, and a plurality of corresponding preliminary estimation images are obtained; the preliminary estimation image output by the first first convolution layer is input to a multi-scale feature extraction module to obtain a multi-scale feature image; the magnetic resonance image, the plurality of preliminary estimation images and the multi-scale feature image are input to a magnetic resonance curve estimation module to obtain a magnetic resonance curve estimation image; the magnetic resonance curve estimation image, the multi-scale feature image and the magnetic resonance image are input to a magnetic resonance multi-element Gaussian blind quality module to obtain a quality score. The application solves the problem of lacking reference magnetic resonance images.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a non-reference-based magnetic resonance image quality assessment method, device and readable medium. Background Art

[0002] Magnetic resonance imaging (MRI) has been widely used in clinical diagnosis and computer-aided diagnosis due to its advantages of clearly displaying soft tissue structures, lacking ionizing radiation hazards, and being able to perform cross-sectional examinations in any orientation. However, the imaging quality of MRI equipment and the acquisition process can cause distortion, affecting the accuracy of subsequent image analysis (such as segmentation and detection). Therefore, image quality assessment is an important benchmark for comparing the quality of magnetic resonance (MR) images before analysis and after image enhancement.

[0003] Currently, image quality assessment methods can be categorized as full-reference, semi-reference, and non-reference. Both full-reference and semi-reference methods require paired or unpaired high- and low-quality images for training. However, obtaining paired or unpaired MRI images is challenging, and large-scale subjective evaluations to obtain mean opinion scores (MASs) through human labeling are difficult. Therefore, due to the lack of reference images for comparison, non-reference image quality assessment becomes important. Traditional non-reference image quality assessment methods and shallow non-reference image quality assessment networks rely primarily on low-level and statistical features of images to assess image quality. These methods fail to deeply understand the semantic information and structure of images and are susceptible to factors such as noise and artifacts. This is because low-level features are insufficient to accurately represent the complex distortions encountered in real-world scenes. Deep learning-based non-reference image quality assessment methods primarily rely on deep features and the learning capabilities of neural network models to assess image quality. They can learn high-level image features such as texture, structure, and semantic information from training data, and can more accurately simulate human visual perception.

[0004] One method proposed in the literature uses the concept of signal-to-noise as an objective metric for measuring the difference between two images. This approach converts the image quality assessment into a ratio of signal (original image) to noise (distorted portion). However, global image quality assessment methods cannot accurately analyze image details and over-analyze background noise, thus affecting the image quality assessment results. Another method proposed in the literature uses one image as the undistorted image and the other as the distorted image. The structural similarity between the two images can be considered as an image quality indicator for the distorted image. Structural similarity is more consistent with the human eye's perception of image quality when measuring image quality. However, such methods require a reference image for evaluation, which is difficult to obtain, and the quality of the reference image can affect model training and evaluation results. Summary of the Invention

[0005] The purpose of this application is to propose a non-reference-based magnetic resonance image quality assessment method, device and readable medium to address the above-mentioned technical problems.

[0006] In a first aspect, the present invention provides a non-reference-based magnetic resonance image quality assessment method, comprising the following steps:

[0007] A non-reference-based magnetic resonance image quality assessment model is constructed and trained to obtain a trained magnetic resonance image quality assessment model, wherein the magnetic resonance image quality assessment model includes an image estimation module, a multi-scale feature extraction module, a magnetic resonance curve estimation module, and a magnetic resonance multivariate Gaussian blind quality module; the image estimation module includes several first convolutional layers, each of which outputs a corresponding preliminary estimated image;

[0008] A magnetic resonance image to be evaluated is obtained and input into a trained magnetic resonance image quality assessment model. The magnetic resonance image is first input into an image estimation module and sequentially passes through several first convolutional layers to obtain several corresponding preliminary estimated images. The preliminary estimated image output by the first first convolutional layer is input into a multi-scale feature extraction module to extract a multi-scale feature image. The magnetic resonance image, several preliminary estimated images and multi-scale feature images are input into a magnetic resonance curve estimation module to predict a magnetic resonance curve estimation image. The magnetic resonance curve estimation image, the multi-scale feature image and the magnetic resonance image are input into a magnetic resonance multivariate Gaussian blind quality module to obtain a corresponding quality score.

[0009] Preferably, in the magnetic resonance curve estimation module, local dynamic estimation is performed on each preliminary estimated image output by the first convolutional layer to obtain the corresponding magnetic resonance curve feature, as shown in the following formula:

[0010] ;

[0011] in, Represents the magnetic resonance curve features corresponding to the preliminary estimated image output by the first convolutional layer, represents a magnetic resonance image; Represents the trainable curve parameters of the first recursive operation, Represents the preliminary estimated image output by the first convolutional layer, Indicates the pixels;

[0012] The magnetic resonance curve features corresponding to the preliminary estimated image output by the first convolutional layer are weightedly fused with the multi-scale feature image to obtain the advanced estimated image of the first recursive operation, as shown in the following formula:

[0013] ;

[0014] in, represents the advanced estimated image of the first recursive operation, and represents the weight coefficient, Represent multi-scale feature images;

[0015] By performing a recursive operation on the advanced estimated image of the i-1th recursive operation and the preliminary estimated image output by the i-th first convolutional layer, the magnetic resonance curve feature corresponding to the preliminary estimated image output by the i-th first convolutional layer is obtained, as shown in the following formula:

[0016] ;

[0017] in, Represents the magnetic resonance curve features corresponding to the preliminary estimated image output by the first convolutional layer i, represents the preliminary estimated image output by the first convolutional layer i, Represents the trainable curve parameters of the i-th recursive operation, i=2,3,…,n, n represents the total number of recursive operations, the same as the total number of the first convolutional layer, It represents the advanced estimated image of the i-1th recursive operation, and its expression is as follows:

[0018] ;

[0019] After n recursive operations, the advanced estimated image of the nth recursive operation is obtained. The advanced estimated image of the nth recursive operation first passes through a ReLU activation function and then passes through the second convolution layer and another ReLU activation function to obtain the magnetic resonance curve estimation image E.

[0020] Preferably, the multi-scale feature extraction module includes several multi-scale feature extraction layers, the multi-scale feature extraction layer includes a third convolutional layer, an average pooling module, a maximum pooling module, a fourth convolutional layer, a fifth convolutional layer, a sixth convolutional layer, a feature fusion layer and a seventh convolutional layer, the average pooling module includes an average pooling layer and a fully connected layer, the maximum pooling module includes a maximum pooling layer and a fully connected layer, and the ReLU activation function is used in the fully connected layer; the input features of the current multi-scale feature extraction layer are input to the third convolutional layer and then pass through the average pooling layer and the maximum pooling layer respectively to obtain the average pooling feature and the maximum pooling feature, the average pooling feature and the maximum pooling feature are respectively input to the corresponding fully connected layer and then fused to obtain the first fusion feature, as shown in the following formula:

[0021] ;

[0022] in, represents the output features of the third convolutional layer, F avg and F max Represent the average pooling and maximum pooling operations respectively, α0 and α1 represent the parameters of the fully connected layer, σ represents the sigmoid activation function, and ReLU represents the ReLU activation function. represents the first fusion feature;

[0023] The first fusion feature is weightedly fused with the input feature of the current multi-scale feature extraction layer and then input into the fourth, fifth, and sixth convolutional layers of different scales respectively to obtain the first scale feature, the second scale feature, and the third scale feature. The first scale feature, the second scale feature, and the third scale feature are input into the feature fusion layer to obtain the second fusion feature, as shown in the following formula:

[0024] ;

[0025] Among them, Z represents the second fusion feature, F j represents the j-th scale feature, j=1,2,..,m, m represents the total number of scales, Concat represents the feature fusion operation, w j represents the weight of the j-th scale feature, and its expression is as follows:

[0026] ;

[0027] Among them, [] represents the splicing operation, represents the Softmax activation function;

[0028] The second fusion feature is activated by the Softmax function and then residually connected with the input feature of the current multi-scale feature extraction layer to obtain the third fusion feature. The third fusion feature is weightedly fused with the second fusion feature after passing through the seventh convolution layer to obtain the output feature of the current multi-scale feature extraction layer. The output feature of the current multi-scale feature extraction layer is then input to the next multi-scale feature extraction layer. The output feature of the last multi-scale feature extraction layer is the multi-scale feature image f b .

[0029] Preferably, the number of multi-scale feature extraction layers in the multi-scale feature extraction module is 6, the input features of the first multi-scale feature extraction layer are the features obtained by the ReLU activation function of the preliminary estimated image output by the first convolutional layer, the input features of the second to fifth multi-scale feature extraction layers are the features obtained by splicing the output features of the previous multi-scale feature extraction layer and the preliminary estimated image output by the first convolutional layer and then passing the ReLU activation function; the input features of the sixth multi-scale feature extraction layer are the features obtained by splicing the output features of the previous multi-scale feature extraction layer and the preliminary estimated image output by the first convolutional layer and then passing the Tanh activation function; the convolution kernel size of the third convolutional layer and the fourth convolutional layer is 3×3, the convolution kernel size of the second convolutional layer, the fifth convolutional layer and the seventh convolutional layer is 5×5; the convolution kernel size of the sixth convolutional layer is 7×7.

[0030] Preferably, in the magnetic resonance multivariate Gaussian blind mass module, the magnetic resonance curve estimation image and the multi-scale feature image are first weighted fused to obtain the final fusion feature, as shown in the following formula:

[0031] ;

[0032] Among them, f b represents the multi-scale feature image, E represents the magnetic resonance curve estimation image, represents the final fusion feature, t1 and t2 represent the fusion parameters;

[0033] The Gaussian distribution of the mean, standard deviation, and deviation between the magnetic resonance image and its corresponding magnetic resonance curve estimation image are calculated using the following formulas, respectively:

[0034] ;

[0035] ;

[0036] ;

[0037] Among them, p m 、p d and p srepresents the Gaussian distribution of the mean, standard deviation, and skewness difference between the MRI image and its corresponding MRI curve estimation image, σ m , σ d and σ s They represent the average of the average values ​​of the gray values ​​calculated for several local areas in the magnetic resonance image, the average of the standard deviations of the gray values ​​calculated for several areas, and the average of the skewness of the gray values ​​calculated for several areas, respectively. m 、μ d and μ s They represent the standard deviation of the mean value of the grayscale values ​​calculated for a number of regions in the magnetic resonance image, the standard deviation of the standard deviation of the grayscale values ​​calculated for a number of regions, and the standard deviation of the skewness of the grayscale values ​​calculated for a number of regions;

[0038] By using the 3-sigma rule to calculate the normal distribution, analyzing the mean, standard deviation, and skewness between the MRI image and its corresponding MRI curve to estimate the image, the quality score of the MRI image is obtained, as shown in the following formula:

[0039] ;

[0040] in, represents the quality score of the magnetic resonance image, and T represents transposition.

[0041] Preferably, the convolution kernel size of the first convolutional layer is 3×3, and the number of first convolutional layers is 7, wherein the output features of the first three first convolutional layers are the corresponding preliminary estimated images, and the output features of the third first convolutional layer are jump-connected with the output features of the fourth first convolutional layer to obtain the preliminary estimated image output by the fourth first convolutional layer; the output features of the second first convolutional layer are jump-connected with the output features of the fifth first convolutional layer to obtain the preliminary estimated image output by the fifth first convolutional layer; the output features of the first first convolutional layer are jump-connected with the output features of the sixth first convolutional layer to obtain the preliminary estimated image output by the sixth first convolutional layer, and the preliminary estimated images output by the first five first convolutional layers need to pass through the ReLU activation function before being input into the next first convolutional layer, and the preliminary estimated image output by the sixth first convolutional layer needs to pass through the Tanh activation function before being input into the seventh first convolutional layer.

[0042] In a second aspect, the present invention provides a non-reference-based magnetic resonance image quality assessment device, comprising:

[0043] a model construction module configured to construct and train a non-reference-based magnetic resonance image quality assessment model to obtain a trained magnetic resonance image quality assessment model, wherein the magnetic resonance image quality assessment model includes an image estimation module, a multi-scale feature extraction module, a magnetic resonance curve estimation module, and a magnetic resonance multivariate Gaussian blind quality module; the image estimation module includes a plurality of first convolutional layers, each of which outputs a corresponding preliminary estimated image;

[0044] The evaluation module is configured to obtain a magnetic resonance image to be evaluated and input it into a trained magnetic resonance image quality assessment model. The magnetic resonance image is first input into the image estimation module and passes through several first convolutional layers in sequence to obtain several corresponding preliminary estimated images; the preliminary estimated image output by the first first convolutional layer is input into the multi-scale feature extraction module to extract a multi-scale feature image; the magnetic resonance image, several preliminary estimated images and the multi-scale feature image are input into the magnetic resonance curve estimation module to predict a magnetic resonance curve estimation image; the magnetic resonance curve estimation image, the multi-scale feature image and the magnetic resonance image are input into the magnetic resonance multivariate Gaussian blind quality module to obtain a corresponding quality score.

[0045] In a third aspect, the present invention provides an electronic device comprising one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement the method described in any implementation manner in the first aspect.

[0046] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0047] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which implements the method described in any implementation manner in the first aspect when the computer program is executed by a processor.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] (1) The non-reference-based MRI image quality assessment method proposed in the present invention proposes an image estimation module and a MRI curve estimation module, which estimate the original brightness and detail features through nonlinear curve mapping to solve the problem of lack of reference MRI images.

[0050] (2) The non-reference-based MRI quality assessment method proposed in this invention proposes a multi-scale feature extraction module, which estimates lesion and tissue features by decomposing the MRI into multi-scale features to fully extract contextual feature information and solve the problems of variability of tissue features in MRI images and randomness of lesion areas.

[0051] (3) The non-reference-based MRI quality assessment method proposed in the present invention proposes a MRI multivariate Gaussian blind quality module, which compares and evaluates various features between the MRI image and the MRI curve estimation image to obtain the quality score of the MRI image. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 Schematic diagram of a flow chart of a non-reference-based magnetic resonance image quality assessment method according to an embodiment of the present application;

[0054] Figure 2 is a schematic diagram of a magnetic resonance image quality assessment model of a non-reference-based magnetic resonance image quality assessment method according to an embodiment of the present application;

[0055] Figure 3 Schematic diagram of a non-reference-based magnetic resonance image quality assessment device according to an embodiment of the present application;

[0056] Figure 4 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] To make the objectives, technical solutions, and advantages of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings. It is apparent that the embodiments described are only some, not all, of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort are intended to fall within the scope of protection of the present invention.

[0058] Figure 1 A non-reference-based magnetic resonance image quality assessment method provided in an embodiment of the present application is shown, comprising the following steps:

[0059] S1, construct and train a non-reference magnetic resonance image quality assessment model to obtain a trained magnetic resonance image quality assessment model, the magnetic resonance image quality assessment model includes an image estimation module, a multi-scale feature extraction module, a magnetic resonance curve estimation module and a magnetic resonance multivariate Gaussian blind quality module; the image estimation module includes several first convolutional layers, each of which outputs a corresponding preliminary estimated image.

[0060] Specifically, refer to Figure 2 First, magnetic resonance images are acquired in a magnetic resonance imaging system to form a magnetic resonance image set. The magnetic resonance image set is divided into a certain proportion to obtain a training set, a validation set, and a test set. A non-reference-based magnetic resonance image quality assessment model is constructed, which includes an image estimation module, a multi-scale feature extraction module, a magnetic resonance curve estimation module, and a magnetic resonance multivariate Gaussian blind quality module. The magnetic resonance image quality assessment model is trained using the magnetic resonance image set to obtain a trained magnetic resonance image quality assessment model. During the training process of the magnetic resonance image quality assessment model, the calculated corresponding magnetic resonance curve features are used as a loss function. By minimizing the magnetic resonance curve features, the optimal trainable parameters are obtained and saved. Subsequently, the trained magnetic resonance image quality assessment model can be used to analyze the brightness, lesion features, and edge features of the magnetic resonance image to obtain the corresponding quality score of the magnetic resonance image.

[0061] S2, obtain the magnetic resonance image to be evaluated and input it into the trained magnetic resonance image quality assessment model. The magnetic resonance image is first input into the image estimation module, and passes through several first convolutional layers in sequence to obtain several corresponding preliminary estimated images; the preliminary estimated image output by the first first convolutional layer is input into the multi-scale feature extraction module to extract the multi-scale feature image; the magnetic resonance image, several preliminary estimated images and the multi-scale feature image are input into the magnetic resonance curve estimation module to predict the magnetic resonance curve estimation image; the magnetic resonance curve estimation image, the multi-scale feature image and the magnetic resonance image are input into the magnetic resonance multivariate Gaussian blind quality module to obtain the corresponding quality score.

[0062] In a specific embodiment, the convolution kernel size of the first convolutional layer is 3×3, and the number of first convolutional layers is 7, where the output features of the first three first convolutional layers are the corresponding preliminary estimated images, and the output features of the third first convolutional layer are jump-connected with the output features of the fourth first convolutional layer to obtain the preliminary estimated image output by the fourth first convolutional layer; the output features of the second first convolutional layer are jump-connected with the output features of the fifth first convolutional layer to obtain the preliminary estimated image output by the fifth first convolutional layer; the output features of the first first convolutional layer are jump-connected with the output features of the sixth first convolutional layer to obtain the preliminary estimated image output by the sixth first convolutional layer, and the preliminary estimated images output by the first five first convolutional layers need to pass through the ReLU activation function before being input into the next first convolutional layer, and the preliminary estimated image output by the sixth first convolutional layer needs to pass through the Tanh activation function before being input into the seventh first convolutional layer.

[0063] Specifically, the image estimation module and magnetic resonance curve estimation module in the magnetic resonance image quality assessment model proposed in the embodiments of this application primarily design curves based on magnetic resonance imaging principles to estimate the original brightness and detail features of the magnetic resonance image. Based on the universal imaging formula of MRI, the purpose is to dynamically adjust the grayscale range of the input magnetic resonance image through the curve to estimate the grayscale value of the magnetic resonance image at optimal contrast and brightness, rather than mapping from image to image.

[0064] refer to Figure 2 The MRI quality assessment model learns the mapping between an MRI image and its best estimated MRI curve through an image estimation module consisting of seven first convolutional layers. Each first convolutional layer consists of a convolutional layer with a kernel size of 3×3. The first convolutional layers of two adjacent layers are connected sequentially with skip connections, without upsampling or downsampling. After the MRI image is input into the MRI quality assessment model, a corresponding preliminary estimated image is obtained after each first convolutional layer, thereby obtaining a set of preliminary estimated images. As the number of network layers increases, more features can be learned at different scales to gradually adapt to the true characteristics of MRI images.

[0065] In a specific embodiment, the multi-scale feature extraction module includes several multi-scale feature extraction layers, the multi-scale feature extraction layer includes a third convolutional layer, an average pooling module, a maximum pooling module, a fourth convolutional layer, a fifth convolutional layer, a sixth convolutional layer, a feature fusion layer and a seventh convolutional layer, the average pooling module includes an average pooling layer and a fully connected layer, the maximum pooling module includes a maximum pooling layer and a fully connected layer, and the fully connected layer uses a ReLU activation function; the input features of the current multi-scale feature extraction layer are input to the third convolutional layer and then pass through the average pooling layer and the maximum pooling layer respectively to obtain the average pooling feature and the maximum pooling feature, the average pooling feature and the maximum pooling feature are respectively input to the corresponding fully connected layer and then fused to obtain the first fusion feature, as shown in the following formula:

[0066] ;

[0067] in, represents the output features of the third convolutional layer, F avg and F max Represent the average pooling and maximum pooling operations respectively, α0 and α1 represent the parameters of the fully connected layer, σ represents the sigmoid activation function, and ReLU represents the ReLU activation function. represents the first fusion feature;

[0068] The first fusion feature is weightedly fused with the input feature of the current multi-scale feature extraction layer and then input into the fourth, fifth, and sixth convolutional layers of different scales respectively to obtain the first scale feature, the second scale feature, and the third scale feature. The first scale feature, the second scale feature, and the third scale feature are input into the feature fusion layer to obtain the second fusion feature, as shown in the following formula:

[0069] ;

[0070] Among them, Z represents the second fusion feature, F j represents the j-th scale feature, j=1,2,..,m, m represents the total number of scales, Concat represents the feature fusion operation, w j represents the weight of the j-th scale feature, and its expression is as follows:

[0071] ;

[0072] Among them, [] represents the splicing operation, represents the Softmax activation function;

[0073] The second fusion feature is activated by the Softmax function and then residually connected with the input feature of the current multi-scale feature extraction layer to obtain the third fusion feature. The third fusion feature is weightedly fused with the second fusion feature after passing through the seventh convolution layer to obtain the output feature of the current multi-scale feature extraction layer. The output feature of the current multi-scale feature extraction layer is then input to the next multi-scale feature extraction layer. The output feature of the last multi-scale feature extraction layer is the multi-scale feature image f b .

[0074] In a specific embodiment, the number of multi-scale feature extraction layers in the multi-scale feature extraction module is 6, the input features of the first multi-scale feature extraction layer are the features obtained by applying the ReLU activation function to the preliminary estimated image output by the first convolutional layer, and the input features of the second to fifth multi-scale feature extraction layers are the features obtained by splicing the output features of the previous multi-scale feature extraction layer with the preliminary estimated image output by the first convolutional layer and then applying the ReLU activation function; the input features of the sixth multi-scale feature extraction layer are the features obtained by splicing the output features of the previous multi-scale feature extraction layer with the preliminary estimated image output by the first convolutional layer and then applying the Tanh activation function; the convolution kernel size of the third and fourth convolutional layers is 3×3, the convolution kernel size of the second, fifth and seventh convolutional layers is 5×5; and the convolution kernel size of the sixth convolutional layer is 7×7.

[0075] Specifically, in clinical diagnosis and image analysis, the main focus is on local features in magnetic resonance images, such as tumors, tissue edges, and detailed features. However, the variability of tissue features and the randomness of damaged areas make it difficult for deep learning networks to learn their features. Therefore, the embodiment of the present application proposes a multi-scale feature extraction module, which inputs the preliminary estimated image y1 output by the first convolutional layer, and after passing through 6 layers of multi-scale feature extraction layers, obtains a multi-scale feature image. Specifically, after the input features of the current multi-scale feature extraction layer are input to the current multi-scale feature extraction layer, they first pass through the third convolution layer with a convolution kernel size of 3×3, the average pooling module, and the maximum pooling module to obtain the first fusion feature, which is fused with the estimated image y1 and subjected to convolution operations with convolution kernel sizes of 3×3, 5×5, and 7×7, respectively, to obtain features of different scales, and a second fusion feature is obtained through feature fusion operation. After passing through the Softmax activation function, a residual operation is performed on the input features of the current multi-scale feature extraction layer to obtain the third fusion feature, which is then passed through the seventh convolution layer with a convolution kernel size of 5×5 and a weighted fusion operation is performed on the third fusion feature to obtain the output feature of the current multi-scale feature extraction layer. The input features of the first multi-scale feature extraction layer are the first preliminary estimated image y1, and the input features of the 2nd to 5th multi-scale feature extraction layers are the features obtained by skipping the output features of the previous multi-scale feature extraction layer and the preliminary estimated image y1 output by the first convolution layer and then passing through the ReLU activation function. The input features of the sixth multi-scale feature extraction layer are the features obtained by skipping the output features of the previous multi-scale feature extraction layer and the preliminary estimated image y1 output by the first convolutional layer, and then passing through the Tanh activation function.

[0076] In a specific embodiment, in the magnetic resonance curve estimation module, local dynamic estimation is performed on each preliminary estimated image output by the first convolutional layer to obtain the corresponding magnetic resonance curve feature, as shown in the following formula:

[0077] ;

[0078] in, Represents the magnetic resonance curve features corresponding to the preliminary estimated image output by the first convolutional layer, represents a magnetic resonance image; Represents the trainable curve parameters of the first recursive operation, Represents the preliminary estimated image output by the first convolutional layer, Indicates the pixels;

[0079] The magnetic resonance curve features corresponding to the preliminary estimated image output by the first convolutional layer are weightedly fused with the multi-scale feature image to obtain the advanced estimated image of the first recursive operation, as shown in the following formula:

[0080] ;

[0081] in, represents the advanced estimated image of the first recursive operation, and represents the weight coefficient, Represent multi-scale feature images;

[0082] By performing a recursive operation on the advanced estimated image of the i-1th recursive operation and the preliminary estimated image output by the i-th first convolutional layer, the magnetic resonance curve feature corresponding to the preliminary estimated image output by the i-th first convolutional layer is obtained, as shown in the following formula:

[0083] ;

[0084] in, Represents the magnetic resonance curve features corresponding to the preliminary estimated image output by the first convolutional layer i, represents the preliminary estimated image output by the first convolutional layer i, Represents the trainable curve parameters of the i-th recursive operation, i=2,3,…,n, n represents the total number of recursive operations, the same as the total number of the first convolutional layer, It represents the advanced estimated image of the i-1th recursive operation, and its expression is as follows:

[0085] ;

[0086] After n recursive operations, the advanced estimated image of the nth recursive operation is obtained. The advanced estimated image of the nth recursive operation first passes through a ReLU activation function and then passes through the second convolution layer and another ReLU activation function to obtain the magnetic resonance curve estimation image E.

[0087] Specifically, in order to adapt to the feature information of each layer, each preliminary estimated image in the preliminary estimated image set needs to be input into the magnetic resonance curve estimation module, which can not only estimate the magnetic resonance curve features, but also iteratively use the magnetic resonance curve features for local dynamic estimation. In the magnetic resonance curve estimation module, it is necessary to recursively operate the multi-scale feature image extracted by the multi-scale feature extraction module and the preliminary estimated image set. After the input magnetic resonance image is input into the magnetic resonance image quality assessment model, it undergoes a convolution operation of the first convolution layer with n convolution kernels of size 3×3 to obtain a preliminary estimated image. The number of iterations is n, which is the same as the number of the first convolution layers. Each first convolution layer of the magnetic resonance image has a corresponding magnetic resonance curve to estimate the best fitting feature. During the iterative process, ∈[0,1], its initial value can be calculated as:

[0088] ;

[0089] Among them, I max and I min Represents the maximum and minimum values ​​of the grayscale value in the magnetic resonance image. By setting different parameters β, the low-frequency and high-frequency characteristics of magnetic resonance images of different grayscale levels can be estimated, and the magnetic resonance curve characteristics of each grayscale level can be accurately estimated. Through iterative estimation, the parameters Gradually approaching the distribution of image features, the estimated magnetic resonance curve features are closer to the real features without distortion. After n recursive operations, an advanced estimated image is obtained for each recursive operation, and the magnetic resonance curve estimated image E is predicted by the advanced estimated image of the nth recursive operation.

[0090] In a specific embodiment, in the magnetic resonance multivariate Gaussian blind mass module, the magnetic resonance curve estimation image and the multi-scale feature image are first weighted fused to obtain the final fusion feature, as shown in the following formula:

[0091] ;

[0092] Among them, f b represents the multi-scale feature image, E represents the magnetic resonance curve estimation image, represents the final fusion feature, t1 and t2 represent the fusion parameters;

[0093] The Gaussian distribution of the mean, standard deviation, and deviation between the magnetic resonance image and its corresponding magnetic resonance curve estimation image are calculated using the following formulas, respectively:

[0094] ;

[0095] ;

[0096] ;

[0097] Among them, p m 、p d and p s represents the Gaussian distribution of the mean, standard deviation, and skewness difference between the MRI image and its corresponding MRI curve estimation image, σ m , σ d and σ s They represent the average of the average values ​​of the gray values ​​calculated for several local areas in the magnetic resonance image, the average of the standard deviations of the gray values ​​calculated for several areas, and the average of the skewness of the gray values ​​calculated for several areas, respectively. m 、μ d and μ s They represent the standard deviation of the mean value of the grayscale values ​​calculated for a number of regions in the magnetic resonance image, the standard deviation of the standard deviation of the grayscale values ​​calculated for a number of regions, and the standard deviation of the skewness of the grayscale values ​​calculated for a number of regions;

[0098] By using the 3-sigma rule to calculate the normal distribution, analyzing the mean, standard deviation, and skewness between the MRI image and its corresponding MRI curve to estimate the image, the quality score of the MRI image is obtained, as shown in the following formula:

[0099] ;

[0100] in, represents the quality score of the magnetic resonance image, and T represents transposition.

[0101] Specifically, unlike natural images, MRI images primarily use different grayscale values ​​to represent different tissue types or water contents. The node saliency score (NSS) based on image intensity distribution cannot directly capture the structural details in MRI images, and the details contained in MRI images of different organs also vary. Therefore, after inputting the MRI curve estimation image and the multi-scale feature image, a fitting curve that conforms to the MRI image distribution is constructed by combining the image's grayscale histogram distribution.

[0102] Then, since the distribution of different features in the MRI follows a normal distribution, the normal distribution is calculated using the 3-sigma rule, and the mean, standard deviation, and skewness between the MRI image and its corresponding MRI curve estimation image are analyzed to obtain the quality score of the MRI image.

[0103] The following is a detailed experiment to illustrate the technical effects of the non-reference-based magnetic resonance image quality assessment method proposed in the embodiments of the present application.

[0104] The experimental parameters for the examples in this application were set as follows: the batch size was set to 16, the model was trained for 200 epochs, and the initial learning rate was 0.005. All experiments used PyTorch version 1.3.1 and were conducted on two NVIDIA RTX 3090 Ti GPUs, each with 24 GB of memory.

[0105] In order to further verify the effectiveness of each module of the non-reference-based magnetic resonance image quality model proposed in the embodiment of the present application, and taking into account that the image estimation module, the magnetic resonance curve estimation module and the magnetic resonance scene statistical blind quality model are indispensable components, a step-by-step performance evaluation was performed on the multi-scale feature extraction module. The results of the ablation study are shown in Table 1. According to the ablation results of the non-reference-based magnetic resonance image quality model proposed in the embodiment of the present application without the multi-scale feature extraction module, the average values ​​of PLCC and SROCC were 0.756 and 0.791, respectively. When the multi-scale feature extraction module was added to the magnetic resonance image quality model, the average values ​​of PLCC and SROCC increased by 12.1%, respectively. This shows that the multi-scale feature extraction module can effectively extract contextual information and detail features from the magnetic resonance image, and when fused with the estimated magnetic resonance image, it can more accurately estimate the background and detail features.

[0106] Table 1 Ablation experiment results:

[0107]

[0108] In order to evaluate the image quality assessment performance of the non-reference-based magnetic resonance image quality assessment method proposed in the embodiments of this application, this application uses four advanced RR-IQA methods for comparative experiments, including the Blind Referenceless Image Spatial Quality Evaluator (BRISQUE), the Wavelet Domain Image Quality Assessment Model (WaDIOaM), the Non-Reference Quality Metric (NROM), and the Neural Image Assessmen (NIMA) model. In addition, three advanced NR-IQA methods are also used for comparison: Discrete Entropy (DE), Gray Mean Gradient (GMG), and Integrated Local Natural Image Quality Evaluator (IL-NIQE).

[0109] According to the PLCC and SROCC evaluation indicators of each image quality assessment method listed in Table 2, the values ​​obtained by the non-reference-based magnetic resonance image quality assessment method proposed in the embodiments of the present application are all optimal. Specifically, the PLCC and SROCC values ​​are 0.848 and 0.864, respectively, indicating that the non-reference-based magnetic resonance image quality assessment method proposed in the embodiments of the present application has excellent performance compared to other advanced image quality assessment methods.

[0110] Table 2 Comparative experimental results:

[0111]

[0112] Further references Figure 3 As an implementation of the methods shown in the above figures, the present application provides an embodiment of a non-reference-based magnetic resonance image quality assessment device. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0113] The present invention provides a non-reference-based magnetic resonance image quality assessment device, comprising:

[0114] A model construction module 1 is configured to construct and train a non-reference-based magnetic resonance image quality assessment model to obtain a trained magnetic resonance image quality assessment model, wherein the magnetic resonance image quality assessment model includes an image estimation module, a multi-scale feature extraction module, a magnetic resonance curve estimation module, and a magnetic resonance multivariate Gaussian blind quality module; the image estimation module includes a plurality of first convolutional layers, each of which outputs a corresponding preliminary estimated image;

[0115] Evaluation module 2 is configured to obtain a magnetic resonance image to be evaluated and input it into a trained magnetic resonance image quality assessment model. The magnetic resonance image is first input into the image estimation module and passes through several first convolutional layers in sequence to obtain several corresponding preliminary estimated images; the preliminary estimated image output by the first first convolutional layer is input into the multi-scale feature extraction module to extract a multi-scale feature image; the magnetic resonance image, several preliminary estimated images and the multi-scale feature image are input into the magnetic resonance curve estimation module to predict a magnetic resonance curve estimation image; the magnetic resonance curve estimation image, the multi-scale feature image and the magnetic resonance image are input into the magnetic resonance multivariate Gaussian blind quality module to obtain a corresponding quality score.

[0116] Figure 4 Schematic diagram of the hardware structure of the electronic device provided by the embodiment of the present invention. Figure 4 As shown, the electronic device of this embodiment includes: a processor 401 and a memory 402; wherein the memory 402 is used to store computer-executable instructions; and the processor 401 is used to execute the computer-executable instructions stored in the memory to implement the various steps performed by the electronic device in the above embodiment. For details, please refer to the relevant description of the above method embodiment.

[0117] Optionally, the memory 402 may be independent or integrated with the processor 401 .

[0118] When the memory 402 is independently provided, the electronic device further includes a bus 403 for connecting the memory 402 and the processor 401 .

[0119] An embodiment of the present invention further provides a computer storage medium, in which computer-executable instructions are stored. When the processor 401 executes the computer-executable instructions, the above method is implemented.

[0120] An embodiment of the present invention further provides a computer program product, including a computer program. When the computer program is executed by the processor 401, the above method is implemented.

[0121] In the embodiments provided herein, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the module division is merely a logical functional division. In actual implementation, other division methods may be used. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not implemented. In addition, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or module, which may be electrical, mechanical or other forms.

[0122] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of these modules may be selected to implement the solution of this embodiment based on actual needs.

[0123] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each module may exist physically separately, or two or more modules may be integrated into a single unit. The units formed by the above modules may be implemented in the form of hardware or hardware plus software functional units.

[0124] The above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or processor 401 to perform some steps of the methods of various embodiments of the present application.

[0125] It should be understood that the processor 401 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), or application-specific integrated circuits (ASIC). A general-purpose processor may be a microprocessor, or the processor 401 may be any conventional processor 401. The steps of the method disclosed in the present invention may be directly implemented as being executed by the hardware processor 401, or may be implemented by a combination of hardware and software modules in the processor 401.

[0126] The memory 402 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk, or an optical disk.

[0127] Bus 403 can be an Industry Standard Architecture (ISA), a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Bus 403 can be classified as an address bus, a data bus, a control bus, etc. For ease of illustration, the bus 403 in the drawings of this application is not limited to a single bus 403 or a single type of bus 403.

[0128] The storage medium may be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0129] An exemplary storage medium is coupled to the processor 401, so that the processor 401 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 401. The processor 401 and the storage medium can be located in an application-specific integrated circuit (ASIC). Of course, the processor 401 and the storage medium can also exist as discrete components in an electronic device or a host control device.

[0130] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A non-reference-based magnetic resonance image quality assessment method, characterized in that: The following steps are involved: A non-reference-based magnetic resonance image quality assessment model is constructed and trained to obtain a trained magnetic resonance image quality assessment model, wherein the magnetic resonance image quality assessment model includes an image estimation module, a multi-scale feature extraction module, a magnetic resonance curve estimation module, and a magnetic resonance multivariate Gaussian blind quality module; the image estimation module includes a plurality of first convolutional layers, each of which outputs a corresponding preliminary estimated image; A magnetic resonance image to be evaluated is obtained and input into the trained magnetic resonance image quality assessment model. The magnetic resonance image is first input into the image estimation module and sequentially passes through several first convolutional layers to obtain several corresponding preliminary estimation images. The preliminary estimation image output by the first first convolutional layer is input into the multi-scale feature extraction module to extract a multi-scale feature image. The magnetic resonance image, several preliminary estimation images and the multi-scale feature image are input into the magnetic resonance curve estimation module to predict a magnetic resonance curve estimation image. The magnetic resonance curve estimation image, the multi-scale feature image and the magnetic resonance image are input into the magnetic resonance multivariate Gaussian blind quality module. In the magnetic resonance multivariate Gaussian blind quality module, the magnetic resonance curve estimation image and the multi-scale feature image are first weighted fused to obtain a final fusion feature, as shown in the following formula: AND a =t1·f b +t2·E; Among them, f b represents the multi-scale feature image, E represents the magnetic resonance curve estimation image, and E a represents the final fusion feature, t1 and t2 represent the fusion parameters; The Gaussian distribution of the mean, standard deviation and skewness between the magnetic resonance image and its corresponding magnetic resonance curve estimation image is calculated using the following formulas, as shown in the following formulas: Among them, p m 、p d and p s represents the Gaussian distribution of the mean, standard deviation and skewness between the magnetic resonance image and its corresponding magnetic resonance curve estimation image, σ m , σ d and σ s They represent the average of the average values ​​of the gray values ​​calculated for several local areas in the magnetic resonance image, the average of the standard deviations of the gray values ​​calculated for several areas, and the average of the skewness of the gray values ​​calculated for several areas, respectively. m 、μ d and μ s They represent the standard deviation of the mean value of the grayscale values ​​calculated for a number of regions in the magnetic resonance image, the standard deviation of the standard deviation of the grayscale values ​​calculated for a number of regions, and the standard deviation of the skewness of the grayscale values ​​calculated for a number of regions; By calculating the normal distribution using the 3-sigma rule and analyzing the mean, standard deviation, and skewness between the magnetic resonance image and its corresponding magnetic resonance curve estimation image, the quality score of the magnetic resonance image is obtained, as shown in the following formula: Where Q represents the quality score of the magnetic resonance image and T represents transposition.

2. The non-reference-based magnetic resonance image quality assessment method according to claim 1, wherein: In the magnetic resonance curve estimation module, local dynamic estimation is performed on each preliminary estimated image output by the first convolutional layer to obtain the corresponding magnetic resonance curve feature, as shown in the following formula: Wherein, MRC1(x) represents the magnetic resonance curve feature corresponding to the preliminary estimated image output by the first convolutional layer, I(x) represents the magnetic resonance image; β1 represents the trainable curve parameter of the first recursive operation, y1(x) represents the preliminary estimated image output by the first convolutional layer, and x represents the x-th pixel point; The magnetic resonance curve features corresponding to the preliminary estimated image output by the first convolutional layer are weightedly fused with the multi-scale feature image to obtain the advanced estimated image of the first recursive operation, as shown in the following formula: E1=W1·f b +W2·MRC1(x); Among them, E1 represents the advanced estimated image of the first recursive operation, W1 and W2 represent weight coefficients, and f b Represent multi-scale feature images; By performing a recursive operation on the advanced estimated image of the i-1th recursive operation and the preliminary estimated image output by the i-th first convolutional layer, the magnetic resonance curve feature corresponding to the preliminary estimated image output by the i-th first convolutional layer is obtained, as shown in the following formula: Among them, MRC i (x) represents the magnetic resonance curve feature corresponding to the preliminary estimated image output by the first convolutional layer i, y i (x) represents the initial estimated image output by the first convolutional layer i, β i represents the trainable curve parameter of the i-th recursive operation, i = 2, 3, ..., n, n represents the total number of recursive operations, which is the same as the total number of the first convolutional layers, E i-1 (x) represents the advanced estimated image of the i-1th recursive operation, and its expression is as follows: E i-1 (x)=W1·E i-2 (x)+W2·MRC i-1 (x); After n recursive operations, the advanced estimated image of the nth recursive operation is obtained. The advanced estimated image of the nth recursive operation first passes through a ReLU activation function and then passes through the second convolution layer and another ReLU activation function to obtain the magnetic resonance curve estimation image E.

3. The non-reference-based magnetic resonance image quality assessment method according to claim 2, characterized in that: The multi-scale feature extraction module includes several multi-scale feature extraction layers, which include a third convolutional layer, an average pooling module, a maximum pooling module, a fourth convolutional layer, a fifth convolutional layer, a sixth convolutional layer, a feature fusion layer and a seventh convolutional layer. The average pooling module includes an average pooling layer and a fully connected layer. The maximum pooling module includes a maximum pooling layer and a fully connected layer. The ReLU activation function is used in the fully connected layers. The input features of the current multi-scale feature extraction layer are input to the third convolutional layer and then pass through the average pooling layer and the maximum pooling layer respectively to obtain average pooling features and maximum pooling features. The average pooling features and the maximum pooling features are respectively input to the corresponding fully connected layers and then fused to obtain the first fusion feature, as shown in the following formula: F=σ(α1·ReLU(α0·F avg (F0))+α1·ReLU(α0·F max (F0))): Among them, F0 represents the output feature of the third convolutional layer, F avg and F max Represent the average pooling and maximum pooling operations respectively, α0 and α1 represent the parameters of the fully connected layer, σ represents the sigmoid activation function, ReLU represents the ReLU activation function, and F represents the first fusion feature; The first fusion feature is weightedly fused with the input feature of the current multi-scale feature extraction layer and then input into the fourth convolutional layer, the fifth convolutional layer, and the sixth convolutional layer of different scales respectively to obtain the first scale feature, the second scale feature, and the third scale feature. The first scale feature, the second scale feature, and the third scale feature are input into the feature fusion layer to obtain the second fusion feature, as shown in the following formula: Z=Concat(w1·F1,w2·F2,...,w m ·F m ); Among them, Z represents the second fusion feature, F j represents the j-th scale feature, j = 1, 2, .., m, m represents the total number of scales, Concat represents the feature fusion operation, w j represents the weight of the j-th scale feature, and its expression is as follows: Among them, [] represents the concatenation operation, and Softmax represents the Softmax activation function; The second fused feature is subjected to the Softmax activation function and then residually connected with the input feature of the current multi-scale feature extraction layer to obtain the third fused feature. The third fused feature is subjected to the seventh convolutional layer and weighted fusion with the second fused feature to obtain the output feature of the current multi-scale feature extraction layer. The output feature of the current multi-scale feature extraction layer is then input to the next multi-scale feature extraction layer. The output feature of the last multi-scale feature extraction layer is the multi-scale feature image f b .

4. The non-reference-based magnetic resonance image quality assessment method according to claim 3, wherein: The number of multi-scale feature extraction layers in the multi-scale feature extraction module is 6. The input features of the first multi-scale feature extraction layer are the features obtained by applying the ReLU activation function to the preliminary estimated image output by the first convolutional layer. The input features of the second to fifth multi-scale feature extraction layers are the features obtained by splicing the output features of the previous multi-scale feature extraction layer and the preliminary estimated image output by the first convolutional layer and then applying the ReLU activation function; the input features of the sixth multi-scale feature extraction layer are the features obtained by splicing the output features of the previous multi-scale feature extraction layer and the preliminary estimated image output by the first convolutional layer and then applying the Tanh activation function; the convolution kernel size of the third and fourth convolutional layers is 3×3, the convolution kernel size of the second, fifth and seventh convolutional layers is 5×5; and the convolution kernel size of the sixth convolutional layer is 7×7.

5. The non-reference-based magnetic resonance image quality assessment method according to claim 1, wherein: The convolution kernel size of the first convolution layer is 3×3, and the number of the first convolution layers is 7. The output features of the first three first convolution layers are the corresponding preliminary estimated images. The output features of the third first convolution layer are jump-connected with the output features of the fourth first convolution layer to obtain the preliminary estimated image output by the fourth first convolution layer; the output features of the second first convolution layer are jump-connected with the output features of the fifth first convolution layer to obtain the preliminary estimated image output by the fifth first convolution layer; the output features of the first first convolution layer are jump-connected with the output features of the sixth first convolution layer to obtain the preliminary estimated image output by the sixth first convolution layer. The preliminary estimated images output by the first five first convolution layers need to pass through the ReLU activation function before being input into the next first convolution layer, and the preliminary estimated image output by the sixth first convolution layer needs to pass through the Tanh activation function before being input into the seventh first convolution layer.

6. A non-reference-based magnetic resonance image quality assessment device, characterized in that: include: a model construction module configured to construct and train a non-reference-based magnetic resonance image quality assessment model to obtain a trained magnetic resonance image quality assessment model, wherein the magnetic resonance image quality assessment model includes an image estimation module, a multi-scale feature extraction module, a magnetic resonance curve estimation module, and a magnetic resonance multivariate Gaussian blind quality module; the image estimation module includes a plurality of first convolutional layers, each of which outputs a corresponding preliminary estimated image; The evaluation module is configured to obtain a magnetic resonance image to be evaluated and input it into the trained magnetic resonance image quality assessment model, wherein the magnetic resonance image is first input into the image estimation module and sequentially passes through several first convolutional layers to obtain several corresponding preliminary estimation images; the preliminary estimation image output by the first first convolutional layer is input into the multi-scale feature extraction module to extract a multi-scale feature image; the magnetic resonance image, several preliminary estimation images and the multi-scale feature image are input into the magnetic resonance curve estimation module to predict a magnetic resonance curve estimation image; the magnetic resonance curve estimation image, the multi-scale feature image and the magnetic resonance image are input into the magnetic resonance multivariate Gaussian blind quality module, and in the magnetic resonance multivariate Gaussian blind quality module, the magnetic resonance curve estimation image and the multi-scale feature image are first weighted fused to obtain a final fusion feature, as shown in the following formula: AND a =t1·f b +t2·E; Among them, f b represents the multi-scale feature image, E represents the magnetic resonance curve estimation image, and E a represents the final fusion feature, t1 and t2 represent the fusion parameters; The Gaussian distribution of the mean, standard deviation and skewness between the magnetic resonance image and its corresponding magnetic resonance curve estimation image is calculated using the following formulas, as shown in the following formulas: Among them, p m 、p d and p s represents the Gaussian distribution of the mean, standard deviation and skewness between the magnetic resonance image and its corresponding magnetic resonance curve estimation image, σ m , σ d and σ s They represent the average of the average values ​​of the gray values ​​calculated for several local areas in the magnetic resonance image, the average of the standard deviations of the gray values ​​calculated for several areas, and the average of the skewness of the gray values ​​calculated for several areas, respectively. m 、μ d and μ s They represent the standard deviation of the mean value of the grayscale values ​​calculated for a number of regions in the magnetic resonance image, the standard deviation of the standard deviation of the grayscale values ​​calculated for a number of regions, and the standard deviation of the skewness of the grayscale values ​​calculated for a number of regions; By calculating the normal distribution using the 3-sigma rule and analyzing the mean, standard deviation, and skewness between the magnetic resonance image and its corresponding magnetic resonance curve estimation image, the quality score of the magnetic resonance image is obtained, as shown in the following formula: Where Q represents the quality score of the magnetic resonance image and T represents transposition.

7. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

9. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images