Artificial intelligence-based ultrasound elastography liver fibrosis assessment method

Abstract

The present application relates to the technical field of elastic imaging, in particular to an ultrasonic elastic imaging liver fibrosis evaluation method based on artificial intelligence. The present application analyzes the gray scale and gradient distribution of the pixel points in the gray scale image of each image to obtain the fat deposition degree of the patient's liver in each image; further obtains the fat display degree of the patient's liver in each parameter combination to screen out high display parameter combinations; obtains the corresponding correction parameter combination according to the imaging depth of each high display parameter combination and the fat deposition degree of the corresponding shear wave elastic image; obtains the image resolution of the corresponding correction parameter combination according to the gradient characteristics of the edge pixel points in the different elastic modulus range areas; obtains the optimal parameter combination, and obtains the liver fibrosis risk degree of the patient according to the elastic modulus distribution in the shear wave elastic image under the optimal parameter combination. The present application improves the accuracy of liver fibrosis risk evaluation by obtaining the optimal parameter combination for elastic imaging.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of elastic imaging, in particular to an ultrasonic elastic imaging liver fibrosis evaluation method based on artificial intelligence. BACKGROUND

[0002] Ultrasonic elastic imaging can detect the elasticity and stiffness of the liver, and the hardness of the liver tissue increases with the increase of the degree of fibrosis, so that by analyzing the ultrasonic elastic imaging can help reflect the degree of fibrosis of the liver.

[0003] In the prior art, shear wave elastic imaging based on fixed parameters is used to perform elastic imaging on the livers of different patients, but in clinical practice, fatty liver and liver fibrosis can both cause changes in the elasticity of the liver, and if there is a large amount of fat droplet deposition on the surface of the liver, it will cause changes in the mechanical properties of the liver tissue, and the shear wave elastic imaging result may be inaccurate due to the fat distribution covering the local or early fibrosis changes, resulting in errors in the evaluation of liver fibrosis. SUMMARY

[0004] In order to solve the technical problem that the evaluation of liver fibrosis is poor when fixed parameters are used for shear wave elastic imaging due to changes in the elasticity of the liver caused by fatty liver and liver fibrosis in clinical practice, the purpose of the present application is to provide an ultrasonic elastic imaging liver fibrosis evaluation method based on artificial intelligence, and the technical solution adopted is as follows:

[0005] The present application provides an ultrasonic elastic imaging liver fibrosis evaluation method based on artificial intelligence, which comprises:

[0006] Obtaining B-mode images of normal livers and patient livers, and shear wave elastic images under different preset imaging depths and probe frequencies to form parameter combinations, wherein the shear wave elastic images contain the elastic modulus of each pixel point;

[0007] For the B-mode images or shear wave elastic images, the fat deposition degree of the patient liver in each image is obtained according to the gray scale and gradient distribution of the pixel points in the gray scale images of each image between the normal liver and the patient liver, and the fat display degree of the patient liver under each parameter combination is obtained according to the difference in fat deposition degree and the difference in gray scale distribution between the B-mode images of the patient liver and the shear wave elastic images under each parameter combination, and the high display parameter combination is screened out;

[0008] According to the imaging depth of each high display parameter combination and the fat deposition degree of the corresponding shear wave elasticity image, a corresponding modified parameter combination is obtained; a plurality of preset elastic modulus range regions of the shear wave elasticity image under each modified parameter combination are obtained, and according to the gradient characteristics of the edge pixel points in different elastic modulus range regions, the image resolution of the corresponding modified parameter combination is obtained; and according to the image resolution and the fat display degree of different modified parameter combinations, an optimal parameter combination is obtained.

[0009] According to the elastic modulus distribution in the shear wave elasticity image under the optimal parameter combination, the liver fibrosis risk degree of the patient is obtained.

[0010] Further, the method for obtaining the fat deposition degree comprises:

[0011] For the gray-scale image of each image, the number of pixel points at different gray-scale values is used to construct a gray-scale histogram, and the average slope level between different gray-scale values and the previous gray-scale value in the gray-scale histogram is obtained as the average slope level of each image.

[0012] The image edge of each gray-scale image is obtained, and the gradient average of all edge pixel points on all image edges is obtained as the average gradient feature.

[0013] According to the difference between the average slope level of each image in the patient's liver and the normal liver, the difference in the average gradient feature, and the gray-scale average of all pixel points on the corresponding gray-scale image of the patient's liver, the fat deposition degree of the patient's liver in each image is obtained, the difference in the average gradient feature is negatively correlated with the fat deposition degree, and the difference in the average slope level and the gray-scale average of all pixel points are positively correlated with the fat deposition degree.

[0014] Further, the method for obtaining the fat display degree comprises:

[0015] The corresponding gray-scale images between the B-ultrasound image of the patient's liver and the shear wave elasticity image under each parameter combination are binarized, and the number of pixel points with consistent results after binarization is counted.

[0016] According to the difference in the fat deposition degree between the ultrasound gray-scale image and the shear wave elasticity gray-scale image, the fat display degree under each parameter combination is obtained, the number of pixel points with consistent results is positively correlated with the fat display degree, and the difference is negatively correlated with the fat display degree.

[0017] Further, the method for obtaining the high display parameter combination comprises:

[0018] If the fat display degree of the patient's liver under any parameter combination is greater than or equal to a preset display threshold, the corresponding parameter combination is taken as a high display parameter combination.

[0019] Further, the method for obtaining the corresponding modified parameter combination comprises:

[0020] According to the fat deposition degree of the shear wave elastography image under each high display parameter combination, the imaging depth corresponding to the fat deposition degree is gain-adjusted to obtain an imaging correction depth;

[0021] The imaging correction depth is substituted into the imaging depth in the high display parameter combination to form a corresponding correction parameter combination.

[0022] Further, the imaging correction depth is obtained by the following method:

[0023] A positive integer 1 and a sum of the fat deposition degrees of the shear wave elastography image under each high display parameter combination are obtained as a fat deposition weight;

[0024] A product between the fat deposition weight and the fat deposition degree is obtained as the imaging correction depth.

[0025] Further, the image resolution is obtained by the following method:

[0026] For each correction parameter combination, a mean value of the gradient amplitudes of all edge pixel points in each elastic modulus range region is obtained as an overall gradient amplitude;

[0027] According to a maximum value of the overall gradient amplitudes in all elastic modulus range regions and an amplitude difference between the maximum value of the overall gradient amplitudes and the overall gradient amplitudes of different elastic modulus range regions, an image resolution of the corresponding correction parameter combination is obtained, the maximum value of the overall gradient amplitudes is positively correlated with the image resolution, and the amplitude difference is negatively correlated with the image resolution.

[0028] Further, the optimal parameter combination is obtained by the following method:

[0029] An element in a coordinate system is formed by the fat display degree and the image resolution of each correction parameter combination, the element in the upper right corner of the coordinate system is selected, and the corresponding correction parameter combination is taken as the optimal parameter combination.

[0030] Further, the liver fibrosis degree is obtained by the following method:

[0031] For the shear wave elastography image under the optimal parameter combination, a target pixel point with an elastic modulus greater than a preset modulus threshold value is selected, a ratio between the number of all target pixel points and the total number of pixel points is obtained as a liver fibrosis risk degree.

[0032] Further, the preset display threshold value is 0.88.

[0033] The present application has the following beneficial effects:

[0034] The present application obtains the fat deposition degree of the patient's liver in each image according to the gray scale and gradient distribution of the pixel points in the gray scale image between the normal liver and the patient's liver, and more comprehensively quantifies the degree of fat deposition; the fat display degree of the patient's liver under each parameter combination is obtained according to the difference in fat deposition degree and the difference in gray scale distribution between the B-ultrasound image of the patient's liver and the shear wave elastography image under each parameter combination, and the high display parameter combination is screened out, which helps to understand the display ability of the shear wave elastography image under the parameter combination to fat deposition; the corresponding modified parameter combination is obtained according to the imaging depth of each high display parameter combination and the fat deposition degree of the corresponding shear wave elastography image; the image resolution of the corresponding modified parameter combination is obtained according to the gradient characteristics of the edge pixel points in different elastic modulus range areas, which reflects the imaging quality of the imaging to the boundary between fat and fibrosis; the optimal parameter combination is obtained, the fibrosis risk degree of the patient's liver is obtained according to the elastic modulus distribution in the shear wave elastography image under the optimal parameter combination, the imaging quality is improved, and the evaluation of the fibrosis risk degree is optimized. The present application improves the accuracy of liver fibrosis risk evaluation by obtaining the optimal parameter combination for elastic imaging. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, below, a brief introduction will be given to the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0036] Figure 1 A flowchart of an ultrasonic elastography liver fibrosis evaluation method based on artificial intelligence provided by an embodiment of the present application;

[0037] Figure 2 A flowchart of a fat deposition degree acquisition method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0038] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined invention purpose, below, the specific implementation, structure, features and effects of a kind of ultrasonic elastography liver fibrosis evaluation method based on artificial intelligence according to the present application are described in detail as follows. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0040] Specifically, the application provides a method for evaluating liver fibrosis based on artificial intelligence and ultrasonic elastography.

[0041] Please refer to Figure 1 , which shows a flowchart of a method for evaluating liver fibrosis based on artificial intelligence and ultrasonic elastography according to an embodiment of the application. The specific method comprises the following steps:

[0042] Step S1: Obtain B-mode images of normal liver and patient liver, and shear wave elastography images under different preset imaging depth and probe frequency parameter combinations, wherein the shear wave elastography images contain the elastic modulus of each pixel point.

[0043] In the embodiment of the application, it is considered that ignoring clinically fatty liver and liver fibrosis will both cause changes in the elasticity of the liver, and that fixed parameters for elastography imaging will cause errors in the evaluation of elasticity. In order to reduce the influence of fat on shear wave elastography images, it is necessary to analyze the fat deposition in combination with liver B-mode images. First, conventional ultrasonic imaging ultrasonic waves are emitted into the body through a probe, and are reflected back to form B-mode images according to the acoustic impedance difference of the tissues. The images can show the morphology, structure and tissue interface of the tissues.

[0044] Radiation force of focused ultrasonic beams in the medical ultrasonic power range is used to generate shear waves in the local area of biological viscous tissue. During ultrasonic elastography imaging, the imaging depth determines the propagation distance of the acoustic waves. The deeper the imaging depth, the stronger the penetrating power of the corresponding acoustic waves, and the deeper the imaging of the organs and tissues. The probe frequency of the acoustic waves determines the probe frequency of the shear wave elastography. The higher the frequency, the stronger the probe frequency, and the greater the image resolution. Therefore, in order to improve the imaging quality of shear wave elasticity, the imaging penetration and the probe frequency need to be considered.

[0045] In the embodiment of the application, the probe frequency range is 1-15MHz, the probe frequencies are spaced 1MHz apart, the imaging depth range is 2-6cm, the imaging depths are spaced 0.5cm apart, and the parameters are combined in order from small to large to obtain shear wave elastography images under different preset imaging depth and probe frequency parameter combinations. When performing ultrasonic elastography, the elastic modulus of the pixel points on the elastic image can be displayed by ROI statistical value to reflect the hardness characteristics of the tissues. The greater the elastic modulus, the greater the hardness. The specific means is a technology known to those skilled in the art, and will not be described here.

[0046] Step S2: For the B-mode image or the shear wave elastic image, the fat deposition degree of the patient's liver in each image is obtained according to the gray distribution of the pixel points in the gray image between the normal liver and the patient's liver; the fat display degree of the patient's liver under each parameter combination is obtained according to the difference in the fat deposition degree of the gray image between the B-mode image and the shear wave elastic image of the patient's liver under each parameter combination, and the gray distribution difference, and the high display parameter combination is screened out.

[0047] The greater the gray value of the pixel, the more extensive the fat distribution, and the more pixels with the greater gray value relative to the normal liver; the worse the morphological characteristics of the liver, the more blurred the contour, and the smaller the gradient, the more likely there is more fat distribution. Compared with the normal liver and the patient's liver, the fat distribution of the liver in the image is quantified by analyzing the gray and gradient of the pixel points in the gray image of each image; for the B-mode image or the shear wave elastic image, the fat deposition degree of the patient's liver in each image is obtained according to the gray distribution of the pixel points in the gray image between the normal liver and the patient's liver.

[0048] Preferably, in an embodiment of the present application, the method for obtaining the fat deposition degree is as follows: Figure 2 , which shows a flow chart of a method for obtaining the fat deposition degree, comprising:

[0049] Step S201: For the gray image of each image, the number of pixel points at different gray values is used to construct a gray histogram, and the average slope level between different gray values and the previous gray value in the gray histogram is obtained as the average slope level of each image.

[0050] It should be noted that the slope can reflect the trend of the number of pixel points, the greater the slope, the greater the difference in the number of pixel points between each gray value and the previous gray value, the more pixel points with greater gray value, indicating more fat distribution; the slope is obtained by the ratio of the difference in the number of pixel points between each gray value and the previous gray value and the corresponding gray value difference, as the slope between adjacent gray values. The specific means is well known to those skilled in the art and is not described here.

[0051] Step S202: Obtain the image edge of each gray image, and obtain the gradient average of all edge pixel points on all image edges as the average gradient feature.

[0052] The deposition of fat will cause changes in the tissue contour of the liver, making the contour clarity worse, so by analyzing the gradient feature of the image edge, the edge clarity is reflected, and the greater the gradient, the clearer the edge; in the embodiment of the present application, the image edge of each gray image can be obtained by the existing CANNY edge detection or Sobel algorithm, and the specific means is well known to those skilled in the art and is not described here.

[0053] Step S203: According to the difference of the average slope level of each image of the patient's liver relative to the normal liver, the difference of the average gradient feature and the average gray value of all pixel points on the corresponding gray image of the patient's liver, the fat deposition degree of the patient's liver in each image is obtained. The difference of the average gradient feature is negatively correlated with the fat deposition degree, and the difference of the average slope level and the average gray value of all pixel points are positively correlated with the fat deposition degree.

[0054] It should be noted that the difference of the average gradient feature reflects the gradient change deviation of the patient's liver relative to the normal liver. The greater the gradient change deviation, the greater the edge definition change. Therefore, the greater the difference of the average gradient feature, the greater the edge definition of the patient's liver, and the smaller the degree of fat deposition. The greater the average slope level, the greater the number of pixel points with larger gray values. The greater the difference of the average slope level, the greater the number of pixel points with larger gray values on the corresponding image of the patient's liver, and the greater the density of fat. The greater the average gray value of the pixel points in the corresponding gray image of the patient's liver, the greater the number of fat appearing with larger gray values, and the greater the fat deposition degree. Therefore, the difference of the average gradient feature is negatively correlated with the fat deposition degree, and the difference of the average slope level and the average gray value of all pixel points are positively correlated with the fat deposition degree.

[0055] In an embodiment of the present application, the product of the difference of the average slope level and the average gray value of all pixel points on the corresponding gray image of the patient's liver is calculated as the liver fat density of the patient's liver. The ratio of the liver fat density of the patient to the difference of the average gradient feature is obtained and normalized as the fat deposition degree of the patient's liver in each image. Therefore, based on the above basic mathematical operation, the correlation between the difference of the average gradient feature, the difference of the average slope level, the average gray value of all pixel points and the fat deposition degree is constructed, that is, the greater the difference of the average gradient feature, the smaller the difference of the average slope level, the smaller the average gray value of all pixel points, the less the fat distribution, and the smaller the liver deposition degree.

[0056] It should be noted that in the embodiment of the present application, the ratio of the liver fat density of the patient to the difference of the average gradient feature is normalized to the range of [0, 1] by using the existing linear normalization or normalization function. The specific means is well known to those skilled in the art, and will not be described here.

[0057] The presence of fat causes the image color to deepen, the greater the fat deposition degree, the greater the difference in image color distribution reflecting the fat distribution in the liver, the greater the difference in fat deposition degree, the smaller the reliability of the image texture similarity, the more the parameter combination cannot show the complete fat state of the patient, the smaller the fat display degree, therefore, according to the difference in fat deposition degree between the B-ultrasound image of the liver of the patient and the gray-scale image under each parameter combination, and the difference in gray-scale distribution, the fat display degree of the liver of the patient under each parameter combination is obtained, and a high-display parameter combination is screened out.

[0058] Preferably, in an embodiment of the present application, the method for obtaining the fat display degree comprises:

[0059] It should be noted that the B-ultrasound image reflects the echo intensity of the tissue, and the shear wave elasticity image reflects the hardness of the tissue, but in the case of liver cirrhosis, the change of the echo intensity is positively correlated with the change of the hardness of the tissue, and the display ratio is consistent relative to the overall range;

[0060] The corresponding gray-scale images between the B-ultrasound image of the liver of the patient and the shear wave elasticity image under each parameter combination are binarized, and the number of pixel points consistent with the binarization result is counted.

[0061] According to the number of pixel points consistent with the result and the difference in fat deposition degree between the ultrasound gray-scale image and the shear wave elasticity gray-scale image, the fat display degree under each parameter combination is obtained, the number of pixel points consistent with the result is positively correlated with the fat display degree, and the difference is negatively correlated with the fat display degree.

[0062] In an embodiment of the present application, the difference between the fat deposition degrees of the ultrasound gray-scale image and the shear wave elasticity gray-scale image or the difference between the ratio of the fat deposition degrees of the ultrasound gray-scale image and the shear wave elasticity gray-scale image and a positive integer 1 is calculated, and is normalized as a first display coefficient; the greater or smaller the ratio is, the greater the difference in fat deposition degree between the images is, the greater the difference between the ratio and the positive integer 1 is, the smaller the fat distribution similarity is, the more incomplete the shear wave elasticity image is in displaying fat, and the smaller the fat display degree is; the product between the superposition similarity and the first display coefficient is obtained as the fat display degree of each parameter combination; therefore, the correlation between the number of pixel points consistent with the result, the difference and the fat display degree is constructed based on the above basic mathematical operation, that is, the smaller the number of pixel points consistent with the result is, the greater the difference is, the more inconsistent the fat distribution is, and the smaller the fat display degree is.

[0063] Preferably, in an embodiment of the present application, the method for obtaining the high-display parameter combination comprises:

[0064] If the fat display degree of the liver of the patient under any parameter combination is greater than or equal to a preset display threshold, the corresponding parameter combination is taken as a high-display parameter combination.

[0065] It should be noted that in one embodiment of the present application, the preset display threshold is 0.88, and in other embodiments of the present application, the size of the preset display threshold can be set according to specific circumstances, which is not limited or described here.

[0066] Step S3: obtaining a corresponding modified parameter combination according to the imaging depth of each high display parameter combination and the fat deposition degree of the corresponding shear wave elasticity image; obtaining a plurality of preset elastic modulus range regions of the shear wave elasticity image under each modified parameter combination, and obtaining the image resolution of the corresponding modified parameter combination according to the gradient characteristics of the edge pixel points in different elastic modulus range regions; and obtaining the optimal parameter combination according to the image resolution and the fat display degree of different modified parameter combinations.

[0067] The deeper the imaging depth, the more deep liver tissue can be imaged, the fat deposition degree reflects the severity of the interference of fat on the ultrasonic signal, the greater the fat deposition degree, the more serious the fat interference, and the more deep imaging depth is needed to penetrate the fat layer; therefore, the corresponding modified parameter combination is obtained according to the imaging depth of each high display parameter combination and the fat deposition degree of the corresponding shear wave elasticity image.

[0068] Preferably, in one embodiment of the present application, the method for obtaining the corresponding modified parameter combination comprises:

[0069] According to the fat deposition degree of the shear wave elasticity image under each high display parameter combination, the corresponding imaging depth is gain-adjusted to obtain an imaging modified depth;

[0070] In one embodiment of the present application, the sum of the positive integer 1 and the fat deposition degree of the shear wave elasticity image under each high display parameter combination is obtained as the fat deposition weight; and the product between the fat deposition weight and the fat deposition degree is obtained as the imaging modified depth; therefore, the correlation between the fat deposition degree, the imaging depth and the imaging modified depth is constructed based on the above basic mathematical operation, that is, the greater the fat deposition degree and the imaging depth, the more serious the influence of fat deposition, the more need for gain adjustment of the imaging depth, and the greater the imaging modified depth, the greater the penetration.

[0071] Based on this, the imaging modified depth is replaced by the imaging depth in the high display parameter combination to constitute the corresponding modified parameter combination, which can penetrate the fat layer and optimize the imaging quality.

[0072] In order to be able to more specifically analyze different imaging regions, a plurality of preset elastic modulus range regions of the shear wave elastography image under each correction parameter combination are obtained; it should be noted that, according to the elastic modulus performance corresponding to the imaging color in clinic, the elastic modulus range reflecting the same color constitutes a region, for example, the elastic modulus range of deep blue is 0-5, and the corresponding region in the range of 0-5 is taken as a preset elastic modulus range region; and then a plurality of preset elastic modulus range regions are obtained according to the existing clinical professional knowledge.

[0073] The edges of different color regions in the image with higher resolution are displayed clearly, and the gradient change is more intense; and the edges of different color regions in the image with low resolution are more blurred, and the gradient change is slow, so that the probe frequency condition of the image is quantified by analyzing the edge pixel gradient characteristics of different elastic modulus range regions; the image resolution corresponding to the correction parameter combination is obtained according to the gradient characteristics of the edge pixel points in different elastic modulus range regions.

[0074] Preferably, in an embodiment of the present application, the image resolution acquisition method comprises:

[0075] For each correction parameter combination, the average value of the gradient amplitudes of all edge pixel points in each elastic modulus range region is obtained as the overall gradient amplitude;

[0076] According to the maximum value of the overall gradient amplitude in all elastic modulus range regions and the amplitude difference between the maximum value of the overall gradient amplitude and the overall gradient amplitude of different elastic modulus range regions, the image resolution corresponding to the correction parameter combination is obtained, the maximum value of the overall gradient amplitude is positively correlated with the image resolution, and the amplitude difference is negatively correlated with the image resolution.

[0077] It should be noted that the greater the maximum value of the overall gradient amplitude, the more intense the edge gradient change of the region, and the greater the edge display definition; the greater the amplitude difference between the maximum value of the overall gradient amplitude and the overall gradient amplitude of the region of different elastic modulus ranges, the smaller the overall edge gradient amplitude of the region relative to the maximum value, the smaller the image resolution, and the smaller the overall edge gradient change of the region, and thus the smaller the resolution. Therefore, the maximum value of the overall gradient amplitude is positively correlated with the image resolution, and the amplitude difference is negatively correlated with the image resolution. In an embodiment of the present application, for each modified parameter combination, the average value of the amplitude difference between the maximum value of the overall gradient amplitude and the overall gradient amplitude of the region of different elastic modulus ranges is obtained, which reflects the change of the overall edge gradient amplitude. The greater the average value of the amplitude difference, the smaller the gradient amplitude change, and the smaller the resolution. The ratio of the maximum value of the overall gradient amplitude to the average value of the amplitude difference is obtained as the image resolution. Therefore, the correlation between the maximum value of the overall gradient amplitude, the amplitude difference, and the image resolution is established based on the above basic mathematical operations, that is, the greater the maximum value of the overall gradient amplitude, the smaller the amplitude difference, the greater the overall gradient amplitude, and the greater the image resolution.

[0078] The greater the fat display degree, the greater the penetration ability, and the greater the imaging of deeper liver tissue. The greater the image resolution, the clearer the detail display, and the higher the imaging quality. The optimal parameter combination is obtained according to the image resolution and the fat display degree of different modified parameter combinations.

[0079] Preferably, the clearer the edge of the image, the greater the image resolution, the higher the probe frequency of the image, the greater the fat display degree, and the stronger the penetration of the parameter combination to fat. In an embodiment of the present application, the method for obtaining the optimal parameter combination comprises:

[0080] An element in the coordinate system is formed by the fat display degree and the image resolution of each modified parameter combination. The element in the upper right corner of the coordinate system is selected, and the corresponding modified parameter combination is taken as the optimal parameter combination.

[0081] It should be noted that the fat display degree of the modified parameter combination is obtained by analyzing the shear wave elasticity image of the modified parameter combination according to the method for obtaining the fat display degree of each parameter combination in step S2.

[0082] Step S4: obtaining the liver fibrosis risk degree of the patient according to the elastic modulus distribution in the shear wave elasticity image under the optimal parameter combination.

[0083] Since liver fibrosis can cause increased liver cell hardness and decreased elasticity, the corresponding color region is darker, and thus the liver fibrosis risk degree of the patient is obtained according to the color distribution characteristics in the shear wave elasticity image of the optimal parameter combination.

[0084] Preferably, in one embodiment of the present application, the method for obtaining the degree of liver fibrosis comprises:

[0085] For the shear wave elastography image under the optimal parameter combination, target pixel points with an elastic modulus greater than a preset modulus threshold value are selected, and a ratio between the number of all target pixel points and the total number of pixel points is obtained as the liver fibrosis risk degree.

[0086] It should be noted that, in the embodiments of the present application, according to existing clinical knowledge, when the elastic modulus is greater than 30, the region will be considered as hard tissue, and the color will be darker, therefore, the preset modulus threshold value is 30.

[0087] Based on this, after obtaining the liver fibrosis risk degree, the liver fibrosis can be evaluated, the greater the liver fibrosis risk degree, the greater the hardness of liver cells, and the more need to prompt cirrhosis or severe liver damage; in the embodiments of the present application, if the liver fibrosis risk degree is in the range , the liver fibrosis is a low risk, and if the liver fibrosis risk degree is greater than or equal to 0.4, the liver fibrosis is a high risk, which is helpful to improve the accuracy of liver fibrosis evaluation of patients with fatty liver.

[0088] In summary, the present application analyzes the gray scale and gradient distribution of the pixel points in the gray scale image of each image to obtain the fat deposition degree of the patient's liver in each image; according to the difference in fat deposition degree and the difference in gray scale distribution between the B-mode ultrasound image of the patient's liver and the shear wave elastography image under each parameter combination, the fat display degree of the patient's liver under each parameter combination is obtained, and the high display parameter combination is screened out; according to the imaging depth of each high display parameter combination and the fat deposition degree of the corresponding shear wave elastography image, the corresponding modified parameter combination is obtained; according to the gradient characteristics of the edge pixel points in different elastic modulus range regions, the image resolution of the corresponding modified parameter combination is obtained; the optimal parameter combination is obtained, and according to the elastic modulus distribution in the shear wave elastography image under the optimal parameter combination, the liver fibrosis risk degree of the patient is obtained. The present application improves the accuracy of liver fibrosis risk evaluation by obtaining the optimal parameter combination for elastic imaging.

[0089] It should be noted that the above-mentioned embodiments of the present application are in the order of description only, and do not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0090] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other, and each embodiment mainly describes the difference from other embodiments.

Claims

1. An artificial intelligence-based ultrasound elastography method for assessing liver fibrosis, characterized in that, The method includes: Ultrasound images of normal liver and patient liver are acquired, as well as shear wave elastic images under different preset imaging depths and probe frequencies, wherein the shear wave elastic images contain the elastic modulus of each pixel. For ultrasound images or shear wave elastography images, the degree of fat deposition in the patient's liver in each image is obtained based on the grayscale and gradient distribution of pixels in the grayscale images of each image between the normal liver and the patient's liver. Based on the difference in the degree of fat deposition in the grayscale images of the patient's liver between the ultrasound images and the shear wave elastography images under each parameter combination, as well as the difference in grayscale distribution, the degree of fat display in the patient's liver under each parameter combination is obtained, and the parameter combination with high display is selected. Based on the imaging depth of each high-display parameter combination and the fat deposition degree of the corresponding shear wave elastic image, the corresponding corrected parameter combination is obtained; multiple preset elastic modulus range regions of the shear wave elastic image under each corrected parameter combination are obtained; based on the gradient characteristics of edge pixels in different elastic modulus range regions, the image resolution of the corresponding corrected parameter combination is obtained; based on the image resolution and fat display degree of different corrected parameter combinations, the optimal parameter combination is obtained. The risk of liver fibrosis in patients is obtained by analyzing the elastic modulus distribution in the shear wave elasticity image under the optimal parameter combination.

2. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the degree of fat deposition includes: For each grayscale image, a grayscale histogram is constructed based on the number of pixels at different grayscale values. The average slope between different grayscale values ​​and the previous grayscale value in the grayscale histogram is obtained and used as the average slope level of each image. Obtain the image edges of each grayscale image, and obtain the average gradient of all edge pixels on all image edges as the average gradient feature; The degree of fat deposition in the patient's liver in each image is obtained by considering the difference in the average slope level of the patient's liver relative to that of a normal liver, the difference in the average gradient features, and the mean gray value of all pixels in the corresponding grayscale image of the patient's liver. The difference in the average gradient features is negatively correlated with the degree of fat deposition, while the difference in the average slope level and the mean gray value of all pixels are positively correlated with the degree of fat deposition.

3. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the fat display degree includes: The corresponding grayscale images of the patient's liver ultrasound image and the shear wave elastic image under each parameter combination were binarized, and the number of pixels with consistent results after binarization was counted. The fat visibility was obtained for each parameter combination based on the difference in fat deposition between the ultrasound grayscale image and the shear wave elastic grayscale image. The number of consistent pixels was positively correlated with the fat visibility, while the difference was negatively correlated with the fat visibility.

4. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the high display parameter combination includes: If the fat display level of the patient's liver in any parameter combination is greater than or equal to the preset display threshold, the corresponding parameter combination will be used as the high display parameter combination.

5. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the corresponding combination of correction parameters includes: Based on the fat deposition degree of the shear wave elastic image under each high display parameter combination, the gain of the corresponding imaging depth is adjusted to obtain the imaging correction depth. Replace the imaging depth in the high-display parameter combination with the imaging correction depth to form the corresponding correction parameter combination.

6. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 5, characterized in that, The method for obtaining the imaging correction depth includes: The sum of the fat deposition degree of the shear wave elastic image under each high display parameter combination and the positive integer 1 is obtained as the fat deposition weight; The product between fat deposition weight and fat deposition degree is obtained as the imaging correction depth.

7. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the image resolution includes: For each combination of correction parameters, the average gradient magnitude of all edge pixels in each elastic modulus range region is obtained as the overall gradient magnitude. Based on the maximum value of the overall gradient magnitude across all elastic modulus ranges, and the difference between the maximum value of the overall gradient magnitude and the overall gradient magnitude across different elastic modulus ranges, the image resolution of the corresponding correction parameter combination is obtained. The maximum value of the overall gradient magnitude is positively correlated with the image resolution, while the difference in magnitude is negatively correlated with the image resolution.

8. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the optimal parameter combination includes: The fat display and image resolution of each correction parameter combination constitute an element in the coordinate system. The element in the upper right corner of the coordinate system is selected, and the corresponding correction parameter combination is taken as the optimal parameter combination.

9. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 1, characterized in that, The methods for obtaining the degree of liver fibrosis include: For the shear wave elastic image under the optimal parameter combination, target pixels with an elastic modulus greater than a preset modulus threshold are selected, and the ratio between the number of all target pixels and the total number of pixels is obtained as the risk of liver fibrosis.

10. The method for assessing liver fibrosis using ultrasound elastography based on artificial intelligence according to claim 4, characterized in that, The preset display threshold is 0.88.

Images