CN113397600A - Hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction - Google Patents
Hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction Download PDFInfo
- Publication number
- CN113397600A CN113397600A CN202110679242.8A CN202110679242A CN113397600A CN 113397600 A CN113397600 A CN 113397600A CN 202110679242 A CN202110679242 A CN 202110679242A CN 113397600 A CN113397600 A CN 113397600A
- Authority
- CN
- China
- Prior art keywords
- ultrasonic
- ultrasonic radio
- frequency signal
- information
- frame
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/08—Detecting organic movements or changes, e.g. tumours, cysts, swellings
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/48—Diagnostic techniques
- A61B8/485—Diagnostic techniques involving measuring strain or elastic properties
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
Abstract
The invention provides a hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction, which comprises the following steps: stressing hepatic tissues with known hepatic fibrosis grade, and collecting two continuous frames of ultrasonic radio-frequency signals before and after stressing as a first frame of ultrasonic radio-frequency signal and a second frame of ultrasonic radio-frequency signal respectively; performing elastic reconstruction on the liver tissue based on the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal to acquire displacement information, strain information and elastic information of the liver tissue; reconstructing a B-mode map of the liver tissue according to the first frame of ultrasonic radio-frequency signals; marking a liver parenchymal region in the B-type map as a region of interest; acquiring data corresponding to the region of interest from the displacement information, the strain information and the elasticity information; and constructing a hepatic fibrosis grading prediction model by a machine learning method. The method can evaluate hepatic fibrosis grade more accurately and simply.
Description
Technical Field
The invention relates to the technical field of computer-aided diagnosis, in particular to a hepatic fibrosis assessment method based on ultrasonic radio-frequency signal elastic reconstruction.
Background
Chronic hepatitis B and C are worldwide diseases and also the main causes of liver fibrosis and cancer, and the degree of liver fibrosis (liver fibrosis) is closely related to the development of liver fibrosis. In detail, liver fibrosis will occur after the liver is inflamed and damaged, and the final result of liver fibrosis will show symptoms of liver fibrosis, and the liver fibrosis patient will have an opportunity to develop liver cancer, so if the degree of liver fibrosis of the patient can be accurately and timely evaluated to prevent the patient from entering the stage of liver fibrosis, the liver cancer prevention will be greatly helped.
Liver histopathological biopsy is the current gold standard for clinically grading liver fibrosis stages, but the method is a traumatic inspection, a patient is anesthetized before puncture, and then the liver part is punctured by using biopsy with the length of about 70mm, and the method is used as an invasive means, the detection cost is high, the patient can suffer from great physical and psychological pains, and the method has a plurality of complications. In addition, it is limited by sampling, which tends to result in inaccurate assessment or missed diagnosis, and has a high false negative rate, thus being limited in clinical use.
The ultrasound technology has the advantages of non-invasiveness, no radiation, convenient operation, low cost and the like, and is increasingly widely applied to hepatic fibrosis diagnosis. There are two main types of data obtained using ultrasound technology: ultrasound B-mode images and ultrasound radio frequency signals.
At present, most of research on liver fibrosis prediction is to predict the grade of liver fibrosis from the perspective of an image, calculate features such as an angular second-order matrix, entropy, contrast, correlation and the like of the image, and then use a support vector machine or a neural network and other methods as a classifier to classify the features extracted from the image. However, in the imaging process of the ultrasound image, many important information, such as phase information and partial tissue information, may be lost, and besides, the imaging result is affected by the model of the instrument, the setting of the instrument, and the like, and the consistency is poor. The ultrasonic radio frequency signal is only related to the physical sending and receiving mode of the imaging equipment, cannot be influenced by the processing measures of the imaging process, contains rich acoustic information including phase, attenuation, backscattering, sound velocity and the like, can better reflect the microstructure of the tissue, but can have another problem when the grade prediction of the hepatic fibrosis is directly carried out on the basis of the ultrasonic radio frequency signal: it is difficult to extract useful features. In addition, the single frame of ultrasound rf signal cannot reflect the hardness information directly related to liver fibrosis.
Therefore, there is a need for improvement in feature extraction based on the existing ultrasound rf signal to more accurately and easily assess the level of liver fibrosis.
Disclosure of Invention
The invention provides a hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction, which aims to construct a hepatic fibrosis grading prediction model, and improves the characteristic extraction aspect on the basis of the existing ultrasonic radio frequency signal so as to more accurately and simply assess the hepatic fibrosis grade.
In order to achieve the above objects and other related objects, the present invention provides a hepatic fibrosis assessment method based on elastic reconstruction of ultrasonic rf signals, comprising:
stressing hepatic tissues with known hepatic fibrosis grade, and collecting two continuous frames of ultrasonic radio-frequency signals as a first frame of ultrasonic radio-frequency signal and a second frame of ultrasonic radio-frequency signal respectively, wherein the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal are respectively an ultrasonic radio-frequency signal before stress and an ultrasonic radio-frequency signal after stress;
performing elastic reconstruction on the liver tissue based on the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal to acquire displacement information, strain information and elastic information of the liver tissue;
reconstructing a B-mode map of the liver tissue according to the first frame of ultrasonic radio-frequency signals;
marking a liver parenchymal region in the B-type map as a region of interest;
acquiring data corresponding to the region of interest from the displacement information, the strain information and the elasticity information;
and constructing a hepatic fibrosis grading prediction model by a machine learning method based on the data of the displacement information, the strain information and the elasticity information corresponding to the region of interest, wherein the hepatic fibrosis grading prediction model is used for evaluating the hepatic fibrosis grade of the hepatic tissue.
Preferably, the elastic reconstruction is performed on the liver tissue to obtain displacement information, strain information and elasticity information of the liver tissue, and specifically includes: firstly, calculating to obtain the displacement information of the liver tissue, then obtaining the strain information through the displacement information, and finally obtaining the elastic information through the closing operation and the opening operation of morphological operation.
Preferably, the calculating to obtain the displacement information of the liver tissue specifically includes:
calibrating the first frame of ultrasonic radio frequency signals to represent:
according to the time shift of the signal window, the second frame of ultrasonic radio frequency signals are expressed as:
wherein θ is the phase of the first frame of ultrasonic RF signal, w0Is the center frequency of the ultrasonic carrier wave, A (t) is the amplitude of the ultrasonic radio frequency signal, and tau is the time shift of the signal window;
calculating the first frame of ultrasonic radio frequency signals and the second frame of ultrasonic radio frequency signals by using a complex cross-correlation function, wherein the calculation formula is as follows:
wherein, T is the size of a signal window for calculating the complex cross correlation, and represents the complex conjugate of the term;
calculating the displacement information u (t) from a complex cross-correlation function:
wherein arg [ R ]ab(t)]Is the phase difference, lambda, of the first frame of ultrasonic RF signals and the second frame of ultrasonic RF signals0Is the wavelength of the ultrasonic radio frequency signal.
Preferably, the obtaining the strain information from the displacement information specifically includes:
carrying out strain estimation on each sampling point on each ultrasonic radio-frequency signal line by using a least square method to obtain the strain at the sampling point; and obtaining the strain information of the liver tissue according to the strain at each sampling point on each ultrasonic radio-frequency signal line.
Preferably, the strain estimation is performed on each sampling point on each ultrasonic radio frequency signal line by using a least square method to obtain the strain at the sampling point, and specifically includes:
fitting the ith sampling point to the (i + N) th sampling point on the jth ultrasonic radio-frequency signal line;
solving the slope of the ith sampling point on the jth ultrasonic radio-frequency signal line by using a least square method to obtain the strain of the liver tissue at the sampling point, wherein the calculation formula is as follows:
preferably, the selection criteria of the region of interest at least include: the tissue depth is the same, the blood vessel avoiding area and the marked shape are polygons.
Preferably, the hepatic fibrosis grading prediction model construction comprises:
extracting high-flux characteristics of the data of the elastic information;
carrying out feature screening on the extracted features by a sparse representation method to obtain different classification categories;
and carrying out classification and judgment based on a support vector machine so that the classification category corresponds to the liver fibrosis grade, and classifying in a machine learning manner so as to construct the liver fibrosis grade prediction model.
Preferably, the high-flux features include at least three aspects of first-order histogram, texture, and wavelet.
Preferably, the extracted features are subjected to feature screening by a sparse representation method by using the following expression:
wherein y is a classification category, and D ═ D1,D2,...,DI]Is a set of all dictionaries, a is a sparse coefficient, mu is a regularization parameter greater than 0, | · | | survivalpIs represented bypThe process of the regularization is carried out,
an evaluation value representing the sparse coefficient a.
Preferably, the classification and determination are performed based on a support vector machine, so that the classification category corresponds to the liver fibrosis level, and the calculation model of the support vector machine is:
yi(wTxi+b)≥1-ξi,i＝1,...,n
ξi≥0,i＝1,...,n
wherein (x)i,yi) For a given training set R ═ x1,y1),(x2,y2),...,(xn,yn)]，yiE.g., +1, -; w ═ w (w)1；w2；...；wd) Determining the direction of the hyperplane for the normal vector; b is a displacement term, and determines the distance between the hyperplane and the origin; xiiIs a relaxation variable, and C is a penalty factor;
and using a radial basis function as a kernel function of the support vector machine, the radial basis function being expressed as:
wherein, | | xi-yi| | is expressed as the squared euclidean distance between two feature vectors, σ being a free parameter.
In summary, the invention provides a hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction, which comprises the steps of firstly performing elastic reconstruction, reconstructing displacement information, strain information and elastic information of a tissue by two continuous frames of ultrasonic radio frequency signals before and after a liver tissue is stressed, performing ultrasonic gray scale imaging to obtain a B-type image of a first frame of ultrasonic radio frequency signal, selecting an ROI (region of interest) on the B-type image, and acquiring data of the ROI corresponding to the displacement information, the strain information and the elastic information; the method comprises the steps of extracting high-flux features, screening the features, classifying, distinguishing and machine learning based on a support vector machine, and constructing a hepatic fibrosis grading prediction model for predicting hepatic fibrosis grades, and has the advantages of no wound, good safety, low price, low requirement on ultrasonic equipment, good performance and the like.
Drawings
Fig. 1 is a schematic diagram of a hepatic fibrosis assessment method based on elastic reconstruction of ultrasonic rf signals according to an embodiment of the present invention.
Detailed Description
The liver fibrosis assessment method based on the elastic reconstruction of the ultrasonic radio frequency signal provided by the invention is further described in detail with reference to fig. 1 and the specific implementation manner. The advantages and features of the present invention will become more apparent from the following description. It is to be noted that the drawings are in a very simplified form and are all used in a non-precise scale for the purpose of facilitating and distinctly aiding in the description of the embodiments of the present invention. To make the objects, features and advantages of the present invention comprehensible, reference is made to the accompanying drawings. It should be understood that the structures, ratios, sizes, and the like shown in the drawings and described in the specification are only used for matching with the disclosure of the specification, so as to be understood and read by those skilled in the art, and are not used to limit the implementation conditions of the present invention, so that the present invention has no technical significance, and any structural modification, ratio relationship change or size adjustment should still fall within the scope of the present invention without affecting the efficacy and the achievable purpose of the present invention.
First, technical terms related to the present invention will be described.
B type graph: the ultrasonic image reflects the difference of acoustic parameters in a medium, and information different from optics, X rays, Y rays and the like can be obtained; the ultrasonic wave has good resolution capability on soft tissues, can obtain useful signals with a dynamic range of more than 120dB, and is beneficial to identifying tiny lesions of biological tissues; when the ultrasonic image displays the living tissue, the required image can be obtained without dyeing treatment. Sparse representation: the purpose of signal sparse representation is to represent a signal by using as few atoms as possible in a given overcomplete dictionary, and a more concise representation mode of the signal can be obtained, so that information contained in the signal can be more easily obtained, and the signal can be more conveniently processed, such as compression, encoding and the like. Signal window: the main mathematical tool for digital signal processing is the fourier transform, which studies the relationship of the entire time and frequency domains; when a computer is used for processing engineering test signals, the infinite signals cannot be measured and operated, but the finite time segments of the infinite signals are taken for analysis; the method includes the steps of intercepting a time slice from a signal, carrying out periodic continuation processing on the intercepted signal time slice to obtain a virtual infinite-length signal, and carrying out mathematical processing such as Fourier transform and correlation analysis on the signal. A support vector machine: a Support Vector Machine (SVM) is a generalized linear classifier (generalized linear classifier) that binary classifies data in a supervised learning manner, and a decision boundary of the SVM is a maximum margin hyperplane for solving a learning sample. Radial basis function: the radial basis function is a real-valued function whose value depends only on the distance from the origin, i.e., Φ (x) Φ (| x |), or may also be the distance to any point c, which is referred to as the center point, i.e., Φ (x, c) | (| x-c |); any function Φ that satisfies the property Φ (x) ═ Φ (iix |) is called a radial basis function, and the criterion typically uses euclidean distances (also called euclidean radial basis functions), although other distance functions are possible; in a neural network architecture, as a primary function of the fully-connected layer and the Re LU layer.
Fig. 1 is a liver fibrosis assessment method based on elastic reconstruction of ultrasonic rf signals according to an embodiment of the present invention, which constructs a liver fibrosis classification prediction model in a machine learning manner through a sample of liver fibrosis classification previously assessed under a known Scheuer scoring standard, and then performs liver fibrosis classification prediction through the constructed liver fibrosis classification prediction model. The method comprises the following steps:
stressing hepatic tissues with known hepatic fibrosis grade, and collecting two continuous frames of ultrasonic radio-frequency signals as a first frame of ultrasonic radio-frequency signal and a second frame of ultrasonic radio-frequency signal respectively, wherein the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal are respectively an ultrasonic radio-frequency signal before stress and an ultrasonic radio-frequency signal after stress;
performing elastic reconstruction on the liver tissue based on the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal to acquire displacement information, strain information and elastic information of the liver tissue;
reconstructing a B-mode map of the liver tissue according to the first frame of ultrasonic radio-frequency signals;
marking a liver parenchymal region in the B-type map as a region of interest;
acquiring data corresponding to the region of interest from the displacement information, the strain information and the elasticity information;
and constructing a hepatic fibrosis grading prediction model by a machine learning method based on the data of the displacement information, the strain information and the elasticity information corresponding to the region of interest, wherein the hepatic fibrosis grading prediction model is used for evaluating the hepatic fibrosis grade of the hepatic tissue.
In specific implementation, for example, the liver ultrasonic RF signals of 81 rats are used as data to perform an experiment, and the liver fibrosis grades are divided into 5 grades of 0,1,2,3 and 4 according to the Scheuer scoring standard, wherein stage 0 represents no liver fibrosis, stage 1 represents mild liver fibrosis, stage 2 represents significant liver fibrosis, stage 3 represents severe liver fibrosis, and stage 4 represents liver cirrhosis. The data distribution is as follows: 6 cases of rats with S0 stage, 28 cases of rats with S1 stage, 28 cases of rats with S2 stage, 10 cases of rats with S3 stage and 9 cases of rats with S4 stage.
1. Firstly, carrying out liver tissue stress on 81 samples, and collecting two continuous frames of ultrasonic radio-frequency signals as a first frame of ultrasonic radio-frequency signal and a second frame of ultrasonic radio-frequency signal respectively, wherein the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal are respectively an ultrasonic radio-frequency signal before stress and an ultrasonic radio-frequency signal after stress; performing elastic reconstruction on the liver tissue based on the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal to acquire displacement information, strain information and elastic information of the liver tissue;
in this embodiment, performing elastic reconstruction on the liver tissue to obtain displacement information, strain information, and elasticity information of the liver tissue specifically includes: firstly, calculating to obtain the displacement information of the liver tissue, then obtaining the strain information through the displacement information, and finally obtaining the elastic information through the closing operation and the opening operation of morphological operation.
In this embodiment, the calculating the displacement information of the liver tissue specifically includes:
calibrating the first frame of ultrasonic radio frequency signals to represent:
according to the time shift of the signal window, the second frame of ultrasonic radio frequency signals are expressed as:
wherein θ is the phase of the first frame of ultrasonic RF signals, w0Is the center frequency of the ultrasonic carrier wave, A (t) is the amplitude of the ultrasonic radio frequency signal, and tau is the time shift of the signal window;
calculating the first frame of ultrasonic radio frequency signals and the second frame of ultrasonic radio frequency signals by using a complex cross-correlation function, wherein the calculation formula is as follows:
wherein, T is the size of the signal window, and represents the complex conjugate of the term;
calculating the displacement information u (t) from a complex cross-correlation function:
wherein arg [ R ]ab(t)]Is the phase difference, lambda, of the first frame ultrasonic radio frequency signal and the second frame ultrasonic radio frequency signal received by the ultrasonic probe before and after tissue compression0Is the wavelength of the ultrasonic radio frequency signal.
In this embodiment, the obtaining the strain information from the displacement information specifically includes:
carrying out strain estimation on each sampling point on each ultrasonic radio-frequency signal line by using a least square method to obtain the strain at the sampling point; and obtaining the strain information of the liver tissue according to the strain at each sampling point on each ultrasonic radio-frequency signal line.
In this embodiment, the strain estimation using the least square method specifically includes: the length of the signal window is N, the displacement field function image between the ith sampling point and the (i + N) th sampling point on the jth ultrasonic radio-frequency signal line is fitted, and the displacement field function image in the signal window is expressed as:
ui,j＝ai,jzi,j+bi,j,i＝1,...,N
wherein z isi,jRepresenting the tissue depth of the ith sampling point on the jth ultrasonic radio-frequency signal line, ai,jAnd bi,jIs a strain value to be estimated;
solving the slope of the fitted displacement function image, namely the strain of the tissue at the sampling point by using a least square method, wherein the calculation formula is as follows:
the displacement field function image u in the window can also be represented in the form of a matrixi,jThe following were used:
where A is an Nx 2 matrix with the first column being the tissue depth zi,jThe second column is a column vector with a length N of all 1 values. The above equation is solved using the least squares method, and is expressed as:
wherein the content of the first and second substances,
is the displacement data that contains noise and is,
is tissue strain.
And dividing the displacement information subjected to average pooling by the absolute value of the strain information, and performing closed operation and open operation of morphological operation to obtain elastic information.
2. Reconstructing a B-mode map of the liver tissue according to the first frame of ultrasonic radio-frequency signals; marking a liver parenchymal region in the B-type map as a region of interest; acquiring data corresponding to the region of interest from the displacement information, the strain information and the elasticity information;
in this embodiment, the selection criteria of the region of interest at least include: the tissue depth is the same, the blood vessel avoiding area and the marked shape are polygons.
3. And constructing a hepatic fibrosis grading prediction model by a machine learning method based on the data of the displacement information, the strain information and the elasticity information, wherein the hepatic fibrosis grading prediction model is used for evaluating the hepatic fibrosis grade of the hepatic tissue.
In this embodiment, the constructing of the hepatic fibrosis classification prediction model includes: extracting high-flux characteristics from the data of the displacement information, the strain information and the elasticity information; carrying out feature screening on the extracted features by a sparse representation method to obtain different classification categories; and carrying out classification and judgment based on a support vector machine so that the classification category corresponds to the liver fibrosis grade, and constructing the liver fibrosis grade prediction model in a machine learning manner.
In this embodiment, the high-throughput features at least include three aspects of a first-order histogram, a texture, and a wavelet, and the method for performing sparse representation on the extracted features performs feature screening, where the screening process is represented as:
wherein the content of the first and second substances,
representing the strain information, y is a classification category, D ═ D1,D2,...,DI]For the set of all dictionaries, a is the sparse coefficientMu is a regularization parameter greater than 0 | · | | non-calculationpIs represented bypAnd (4) regularizing.
In this embodiment, the classification and determination based on the support vector machine is performed to make the classification category correspond to the liver fibrosis level, and the calculation model of the support vector machine is:
yi(wTxi+b)≥1-ξi,i＝1,...,n
ξi≥0,i＝1,...,n
wherein (x)i,yi) For a given training set R ═ x1,y1),(x2,y2),...,(xn,yn)]，yiE.g., +1, -; w ═ w (w)1；w2；...；wd) Determining the direction of the hyperplane for the normal vector; b is a displacement term, and determines the distance between the hyperplane and the origin; xiiIs a relaxation variable, and C is a penalty factor;
and using a radial basis function as a kernel function of the support vector machine, the radial basis function being expressed as:
wherein, | | xi-yi| | is expressed as the squared euclidean distance between two feature vectors, σ being a free parameter.
Extracting high-throughput features of the data, extracting high-throughput features of three aspects of a first-order histogram, texture and wavelet from each information graph, wherein the total number of the high-throughput features is 350, the specific features and the number of the high-throughput features are shown in table 1, and the features of the three information graphs are spliced together to form 1050 features;
TABLE 1
The classification performance results of 81 samples were obtained, and are shown in table 2 below: mild hepatic fibrosis F is more than or equal to F1Significant hepatic fibrosis F is more than or equal to F2Severe hepatic fibrosis F is more than or equal to F3And hepatic fibrosis F ═ F4。
TABLE 2
Wherein the Accuracy (ACC), sensitivity (SENS), Specificity (SPEC), Positive Predictive Value (PPV), Negative Predictive Value (NPV) are calculated as follows:
wherein, TP, TN, FP and FN are the number of true positive, true negative, false positive and false negative respectively.
According to the technical scheme, elastic reconstruction is firstly carried out, displacement information, strain information and elastic information of the tissues are reconstructed by two continuous frames of ultrasonic radio-frequency signals before and after the liver tissues are stressed, ultrasonic gray-scale imaging is carried out, a B-type image of the first frame of ultrasonic radio-frequency signals is obtained, an ROI (region of interest) is selected on the B-type image, and data of the ROI corresponding to the displacement information, the strain information and the elastic information are obtained; the method comprises the steps of extracting high-flux features, screening the features, classifying, distinguishing and machine learning based on a support vector machine, and constructing a hepatic fibrosis grading prediction model for predicting hepatic fibrosis grades, and has the advantages of no wound, good safety, low price, low requirement on ultrasonic equipment, good performance and the like.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.
Claims (10)
1. A hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction is characterized by comprising the following steps:
stressing hepatic tissues with known hepatic fibrosis grade, and collecting two continuous frames of ultrasonic radio-frequency signals as a first frame of ultrasonic radio-frequency signal and a second frame of ultrasonic radio-frequency signal respectively, wherein the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal are respectively an ultrasonic radio-frequency signal before stress and an ultrasonic radio-frequency signal after stress;
performing elastic reconstruction on the liver tissue based on the first frame of ultrasonic radio-frequency signal and the second frame of ultrasonic radio-frequency signal to acquire displacement information, strain information and elastic information of the liver tissue;
reconstructing a B-mode map of the liver tissue according to the first frame of ultrasonic radio-frequency signals;
marking a liver parenchymal region in the B-type map as a region of interest;
acquiring data corresponding to the region of interest from the displacement information, the strain information and the elasticity information;
and constructing a hepatic fibrosis grading prediction model by a machine learning method based on the data of the displacement information, the strain information and the elasticity information corresponding to the region of interest, wherein the hepatic fibrosis grading prediction model is used for evaluating the hepatic fibrosis grade of the hepatic tissue.
2. The method for assessing liver fibrosis based on elastic reconstruction of ultrasonic radio frequency signals according to claim 1, wherein the elastic reconstruction of the liver tissue is performed to obtain displacement information, strain information and elasticity information of the liver tissue, and specifically comprises: firstly, calculating to obtain the displacement information of the liver tissue, then obtaining the strain information through the displacement information, and finally obtaining the elastic information through the closing operation and the opening operation of morphological operation.
3. The method for assessing liver fibrosis based on elastic reconstruction of ultrasonic rf signals according to claim 2, wherein the calculating to obtain the displacement information of the liver tissue specifically comprises:
calibrating the first frame of ultrasonic radio frequency signals to represent:
according to the time shift of the signal window, the second frame of ultrasonic radio frequency signals are expressed as:
wherein θ is the phase of the first frame of ultrasonic RF signal, w0Is the center frequency of the ultrasonic carrier wave, A (t) is the amplitude of the ultrasonic radio frequency signal, and tau is the time shift of the signal window;
calculating the first frame of ultrasonic radio frequency signals and the second frame of ultrasonic radio frequency signals by using a complex cross-correlation function, wherein the calculation formula is as follows:
wherein, T is the size of a signal window for calculating the complex cross correlation, and represents the complex conjugate of the term;
calculating the displacement information u (t) from a complex cross-correlation function:
wherein arg [ R ]ab(t)]Is the phase difference, lambda, of the first frame of ultrasonic RF signals and the second frame of ultrasonic RF signals0Is the wavelength of the ultrasonic radio frequency signal.
4. The method for assessing liver fibrosis based on elastic reconstruction of ultrasonic RF signals according to claim 3, wherein the determining the strain information from the displacement information specifically comprises:
carrying out strain estimation on each sampling point on each ultrasonic radio-frequency signal line by using a least square method to obtain the strain at the sampling point; and obtaining the strain information of the liver tissue according to the strain at each sampling point on each ultrasonic radio-frequency signal line.
5. The method for assessing liver fibrosis based on ultrasonic radio frequency signal elastic reconstruction according to claim 4, wherein the strain estimation is performed on each sampling point on each ultrasonic radio frequency signal line by using a least square method to obtain the strain at the sampling point, specifically comprising:
fitting the ith sampling point to the (i + N) th sampling point on the jth ultrasonic radio-frequency signal line;
solving the slope of the ith sampling point on the jth ultrasonic radio-frequency signal line by using a least square method to obtain the strain of the liver tissue at the sampling point, wherein the calculation formula is as follows:
6. the method for assessing liver fibrosis based on elastic reconstruction of ultrasonic RF signals according to claim 1, wherein the selection criteria of the region of interest at least includes: the tissue depth is the same, the blood vessel avoiding area and the marked shape are polygons.
7. The hepatic fibrosis assessment method based on elastic reconstruction of ultrasonic radio frequency signals of claim 1, wherein the hepatic fibrosis grading prediction model construction comprises:
extracting high-flux characteristics of the data of the elastic information;
carrying out feature screening on the extracted features by a sparse representation method to obtain different classification categories;
and carrying out classification and judgment based on a support vector machine so that the classification category corresponds to the liver fibrosis grade, and classifying in a machine learning manner so as to construct the liver fibrosis grade prediction model.
8. The method for assessing liver fibrosis based on ultrasound radiofrequency signal elastic reconstruction of claim 7, wherein the high-throughput features include at least three aspects of first-order histogram, texture and wavelet.
9. The hepatic fibrosis assessment method based on elastic reconstruction of ultrasonic radio frequency signals according to claim 8, wherein the extracted features are subjected to feature screening by a sparse representation method by adopting the following expression:
wherein y is a classification category, and D ═ D1,D2,...,DI]Is a set of all dictionaries, a is a sparse coefficient, mu is a regularization parameter greater than 0, | · | | survivalpIs represented bypThe process of the regularization is carried out,
an evaluation value representing the sparse coefficient a.
10. The method for assessing liver fibrosis based on elastic reconstruction of ultrasound rf signals according to claim 9, wherein the classification is performed based on a support vector machine, so that the classification category corresponds to the liver fibrosis grade, and the calculation model of the support vector machine is:
yi(wTxi+b)≥1-ξi,i＝1,...,n
ξi≥0,i＝1,...,n
wherein (x)i,yi) For a given training set R ═ x1,y1),(x2,y2),...,(xn,yn)]，yiE.g., +1, -; w ═ w (w)1；w2；...；wd) Determining the direction of the hyperplane for the normal vector; b is a displacement term, and determines the distance between the hyperplane and the origin; xiiIs a relaxation variable, and C is a penalty factor;
and using a radial basis function as a kernel function of the support vector machine, the radial basis function being expressed as:
wherein, | | xi-yi| | is expressed as the squared euclidean distance between two feature vectors, σ being a free parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110679242.8A CN113397600A (en) | 2021-06-18 | 2021-06-18 | Hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110679242.8A CN113397600A (en) | 2021-06-18 | 2021-06-18 | Hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113397600A true CN113397600A (en) | 2021-09-17 |
Family
ID=77681396
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110679242.8A Pending CN113397600A (en) | 2021-06-18 | 2021-06-18 | Hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113397600A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080064956A1 (en) * | 2006-09-13 | 2008-03-13 | Medison Co., Ltd. | Method of displaying an elastic image |
| US20100036250A1 (en) * | 2008-08-05 | 2010-02-11 | Medison Co., Ltd. | Formation Of A Color Map For An Elastic Image |
| CN102283679A (en) * | 2011-08-04 | 2011-12-21 | 中国科学院深圳先进技术研究院 | Ultrasonic imaging system for elasticity measurement and method for measuring elasticity of biological tissue |
| CN102860842A (en) * | 2012-09-26 | 2013-01-09 | 浙江大学 | Real-time quasi-static ultrasonic elastography method |
| CN103637821A (en) * | 2013-11-27 | 2014-03-19 | 华南理工大学 | Hepatic fibrosis degree Fisher identification method based on ultrasonic radio frequency (RF) time sequence |
| US20140094702A1 (en) * | 2011-05-23 | 2014-04-03 | The Regents Of The University Of Michigan | Methods and apparatuses for measuring tissue stiffness changes using ultrasound elasticity imaging |
| US20150148658A1 (en) * | 2012-07-11 | 2015-05-28 | University Of Mississippi Medical Center | Method for the detection and staging of liver fibrosis from image acquired data |
| CN112862797A (en) * | 2021-02-23 | 2021-05-28 | 复旦大学附属华山医院 | Liver fibrosis nondestructive prediction method and system |
-
2021
- 2021-06-18 CN CN202110679242.8A patent/CN113397600A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080064956A1 (en) * | 2006-09-13 | 2008-03-13 | Medison Co., Ltd. | Method of displaying an elastic image |
| US20100036250A1 (en) * | 2008-08-05 | 2010-02-11 | Medison Co., Ltd. | Formation Of A Color Map For An Elastic Image |
| US20140094702A1 (en) * | 2011-05-23 | 2014-04-03 | The Regents Of The University Of Michigan | Methods and apparatuses for measuring tissue stiffness changes using ultrasound elasticity imaging |
| CN102283679A (en) * | 2011-08-04 | 2011-12-21 | 中国科学院深圳先进技术研究院 | Ultrasonic imaging system for elasticity measurement and method for measuring elasticity of biological tissue |
| WO2013017105A1 (en) * | 2011-08-04 | 2013-02-07 | Shenzhen Institutes Of Advanced Technology Chinese Academy Of Sciences | Ultrasonic imaging system and method for measuring elasticity of biological tissues |
| US20150148658A1 (en) * | 2012-07-11 | 2015-05-28 | University Of Mississippi Medical Center | Method for the detection and staging of liver fibrosis from image acquired data |
| CN102860842A (en) * | 2012-09-26 | 2013-01-09 | 浙江大学 | Real-time quasi-static ultrasonic elastography method |
| CN103637821A (en) * | 2013-11-27 | 2014-03-19 | 华南理工大学 | Hepatic fibrosis degree Fisher identification method based on ultrasonic radio frequency (RF) time sequence |
| CN112862797A (en) * | 2021-02-23 | 2021-05-28 | 复旦大学附属华山医院 | Liver fibrosis nondestructive prediction method and system |
Non-Patent Citations (1)
| Title |
|---|
| 付甜甜等: ""计算机辅助超声影像组学预测肝纤维化进程的实验研究"", 《中国超声医学杂志》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Progress in non-invasive detection of liver fibrosis | |
| US10877136B2 (en) | Method for improved ultrasonic detection | |
| Roy-Cardinal et al. | Assessment of carotid artery plaque components with machine learning classification using homodyned-K parametric maps and elastograms | |
| Hsu et al. | Breast tumor classification using different features of quantitative ultrasound parametric images | |
| Bhatele et al. | Glioma Segmentation and Classification System Based on Proposed Texture Features Extraction Method and Hybrid Ensemble Learning. | |
| Hyun et al. | Ultrasound lesion detectability as a distance between probability measures | |
| Bharath et al. | Automated quantification of ultrasonic fatty liver texture based on curvelet transform and SVD | |
| Ergün et al. | Classification of carotid artery stenosis of patients with diabetes by neural network and logistic regression | |
| Chuang et al. | Detection of pediatric hepatic steatosis through ultrasound backscattering analysis | |
| Xing et al. | Automatic detection of A‐line in lung ultrasound images using deep learning and image processing | |
| Muhtadi | Breast tumor classification using intratumoral quantitative ultrasound descriptors | |
| Baek et al. | Improving breast cancer diagnosis by incorporating raw ultrasound parameters into machine learning | |
| CN113397600A (en) | Hepatic fibrosis assessment method based on ultrasonic radio frequency signal elastic reconstruction | |
| CN111528918A (en) | Tumor volume change trend graph generation device after ablation, equipment and storage medium | |
| Mohamed et al. | Computer-aided diagnosis for prostate cancer using support vector machine | |
| Fakhr et al. | Automatic diagnosis of liver diseases from ultrasound images | |
| Liu et al. | Diagnosis of significant liver fibrosis by using a DCNN model with fusion of features from US B-mode image and Nakagami parametric map: An animal study | |
| CN112862797B (en) | Liver fibrosis nondestructive prediction method and system | |
| Seabra et al. | A Rayleigh mixture model for IVUS imaging | |
| Roy-Cardinal et al. | Homodyned K-distribution parametric maps combined with elastograms for carotid artery plaque assessment | |
| CN111528914A (en) | Image processing method and device | |
| Goryachev et al. | Comparison of Quantitative Ultrasound Methods to Classify Dystrophic and Obese Models of Skeletal Muscle | |
| US20220361839A1 (en) | Non-invasive ultrasound detection device for liver fibrosis and method thereof | |
| Arunachalam et al. | Detection of Singularity in the Cell Nucleus of Synovial Sarcoma Using Wavelet Leaders | |
| Jarosik et al. | Estimating the ultrasound attenuation coefficient using convolutional neural networks--a feasibility study |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210917 |
|
| RJ01 | Rejection of invention patent application after publication |