Medical ultrasonic image intelligent processing method and system

Abstract

The invention discloses a medical ultrasonic image intelligent processing method and system, and relates to the technical field of medical image.According to the main scheme, normal feature coordinates of different organs and tissues are calculated by obtaining normal ultrasonic image data, to-be-processed image data and historical medical record information of a patient; extracting coordinates of a to-be-detected image and comparing the coordinates with normal coordinates for preliminary judgment; performing secondary judgment by combining ovality and circularity; setting a lesion gray threshold to judge whether a lesion exists or not; calculating the size of a lesion area when the lesion exists; according to the method and the system, the comprehensive blood flow evaluation index is calculated in combination with color Doppler blood flow imaging data, the disease evaluation value is calculated according to known medical record data and the like, the disease level is judged, the diagnosis report is automatically generated, the accuracy and the efficiency of medical ultrasonic diagnosis are improved, and powerful support is provided for clinical diagnosis.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and in particular to a medical ultrasonic image intelligent processing method and system. Background Art

[0002] Medical ultrasound diagnosis occupies an important position in modern medicine, but it currently faces many challenges in the analysis and processing of ultrasound images. On the one hand, it is difficult to accurately judge the types of different tissues and organs and the condition of the lesions, and it is necessary to consider a variety of factors. On the other hand, there is a lack of efficient and accurate methods for quantitative analysis of lesion areas, and evaluation of tissue texture characteristics and blood flow characteristics, which means that the diagnostic accuracy and efficiency of the disease need to be improved.

[0003] In the prior art, ultrasound image data is usually collected, and after a certain degree of preprocessing, the image features are observed based on the doctor's experience to determine the tissue and organ type and the condition of the lesion. For some specific indicators, such as the size of the lesion area, there is a lack of accurate measurement methods. For blood flow characteristics, the Doppler effect principle is used to measure blood flow velocity and resistance index. In terms of texture feature analysis, some basic grayscale statistical methods may be used to preliminarily determine the uniformity of the tissue.

[0004] However, the existing technology has obvious shortcomings. First, it relies on the doctor's experience and judgment, which is highly subjective, inefficient, and prone to misjudgment and omission. Second, it is difficult to accurately and comprehensively reflect the complex characteristics of the lesion area and tissue texture. The analysis of blood flow characteristics is also relatively simple, and it cannot comprehensively consider multiple factors such as blood flow velocity, resistance index and blood flow direction to accurately evaluate vascular health and lesion conditions. Moreover, the existing technology lacks a unified comprehensive evaluation system and cannot effectively integrate multiple information such as the size of the lesion area, texture structure, and blood flow characteristics to achieve a more accurate judgment of the disease level and generate a detailed diagnostic report. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the deficiencies in the prior art, the present invention provides a method and system for intelligent processing of medical ultrasound images, which reduces misjudgments and missed judgments caused by doctors' subjective judgments by comparing the edge direction coordinate range thresholds of various organ tissues with the reference edge direction coordinate data; can quantitatively analyze the scale of the lesion by calculating the size of the lesion area; calculates the texture structure difference value and the comprehensive structure texture index by using grayscale co-occurrence matrix data and two-dimensional grayscale image data, overcoming the limitations of the simple grayscale statistical method in the prior art; can more accurately evaluate the health status of blood vessels and the condition of lesions by comprehensively considering multiple factors such as blood flow velocity, resistance index and blood flow direction; by comparing the symptom assessment value with the symptom assessment threshold, combining the preliminary diagnosis results to judge the symptom level and automatically generate a diagnosis report, a unified comprehensive assessment system is established, which effectively integrates multiple information such as lesion area size, texture structure, blood flow characteristics, etc., improves the accuracy and efficiency of symptom diagnosis, and provides solid support for medical decision-making.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a medical ultrasound image intelligent processing method, comprising:

[0009] Acquire normal ultrasound image data, which includes normal image samples of multiple normal tissues or organs, and acquire normal two-dimensional grayscale image data of normal image samples; acquire two-dimensional grayscale image test data and color Doppler blood flow imaging data of image samples to be processed; and acquire historical medical record information data of patients;

[0010] Through the two-dimensional grayscale image normal data of different normal image samples, the normal initial edge coordinate points (x i ,y i ), according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu );

[0011] The initial edge coordinates (x a ,y a ), and calculate the coordinates of the reference edge to be measured (X a , Y a );Calculate the coordinates of the reference edge to be measured (X a , Y b ) and the normal characteristic coordinates (X uu , Y uu ) uu , make a preliminary judgment on the type of organ tissue;

[0012] According to the coordinates of the reference edge to be measured (X a , Y a ) Calculate the ellipticity E and circularity Eq of the organ tissue; obtain the ellipticity interval threshold and circularity interval threshold of different organ tissues, compare the ellipticity E and circularity Eq with the ellipticity interval threshold and circularity interval threshold of different organ tissues, perform secondary judgment on the type of the organ tissue according to the comparison result, and finally determine the organ tissue type of the image sample to be processed;

[0013] Get the gray value of each coordinate point in the two-dimensional gray image to be tested; set the gray threshold Ew of the lesion, compare the gray value of each coordinate point with the gray threshold Ew of the lesion, and determine whether there is an abnormal coordinate with gray value. All abnormal coordinate points constitute an abnormal area, and the boundary coordinates of the abnormal area form the initial edge coordinate point to be tested (x c ,y c ), according to the initial edge coordinate point 2 to be measured (x c ,y c ) Calculate the coordinates of the reference edge to be measured (X c , Y c ); Get the boundary coordinates of the hypoechoic nodule from the image sample to be processed (xc g , yc g ), according to the coordinates of the reference edge to be measured (X c , Y c ) and the hypoechoic nodule boundary coordinates (xc g ,yc g ) to determine whether there are lesions in organ tissues;

[0014] When a lesion is determined, the boundary coordinates of the hypoechoic nodule (xc g , yc g ) Calculate the size A of the lesion area;

[0015] A grayscale co-occurrence matrix is ​​obtained according to the two-dimensional grayscale image data to be tested, and texture structure difference data inside the tissue is calculated according to the grayscale co-occurrence matrix and the two-dimensional grayscale image data to be tested; and a comprehensive structure texture index STI is calculated according to the texture structure difference data;

[0016] According to the color Doppler blood flow imaging data, the blood flow velocity Cv and the blood flow resistance index RII are calculated, and the blood flow direction TIF is determined. According to the blood flow velocity Cv, the blood flow resistance index RII, the blood flow direction TIF and the patient's historical medical information data, the comprehensive blood flow assessment index MPI is calculated;

[0017] The disease assessment value CEV is calculated based on the known medical record data, the size A of the lesion area, the comprehensive structural texture index STI and the comprehensive blood flow assessment index MPI. The disease assessment value CEV and the disease assessment threshold set Dv are compared, combined with the preliminary diagnosis results, the disease level is judged, and a diagnosis report is automatically generated.

[0018] In the preferred embodiment of the above-mentioned intelligent processing method for medical ultrasound images: according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu ) are as follows:

[0019] According to the normal initial edge coordinate point (x i ,y i ) Calculate the normal mean coordinates (Xi mean , Yi mean ), based on the following formula:

[0020]

[0021] Among them, x i Indicates the x-axis value of the i-th edge coordinate point, y i Represents the y-axis value of the i-th edge coordinate point, the value range of i is 1, 2, 3...n, and n represents the total number of initial mark-off coordinate points;

[0022] According to the normal initial edge coordinate point (x i ,y i ) and normal mean coordinates (Xi mean , Yi mean ), calculate the coordinate standard deviation (X m , Y m ), the specific formula is:

[0023]

[0024] According to the coordinate standard deviation (X m , Y m ) and normal mean coordinates (Xi mean , Yi mean ) Calculate the normal feature coordinates (X uu , Y uu ), based on the following specific formula:

[0025] X uu =(X mean -kX m , X mean +k m )

[0026] Y uu =(Y mean -kY m , Y mean +kY m ).

[0027] In the preferred embodiment of the medical ultrasound image intelligent processing method, the reference edge coordinate data point (X a , Y a ), and the specific steps for preliminarily determining the organ tissue type are:

[0028] Extract the initial edge coordinate point (x a ,y a ), and obtain the initial edge coordinate point to be measured (x a ,y a ) in the x-axis direction man , the minimum value x in the x-axis direction min , maximum value y in the y-axis direction man and the minimum value y in the y-axis direction min ;

[0029] According to the coordinate data point of the benchmark edge to be measured (x a ,y a ), calculate the coordinate mean (X mean , Y mean ), the specific formula is:

[0030]

[0031] Among them, x a Indicates the value of the ath coordinate on the x-axis, y b It represents the value of the b-th coordinate on the y-axis. The value range of a is 1, 2, 3...x, and x represents the total number of coordinates.

[0032] According to the maximum value x in the x-axis direction man , the minimum value x in the x-axis direction min , maximum value y in the y-axis direction man and the minimum value y in the y-axis direction min The specific formula for calculating the coordinate value normalization factor d is as follows:

[0033]

[0034] According to the initial edge coordinate point to be measured (x a ,y a ), coordinate mean (X mean , Y mean ) and the coordinate value normalization factor d, calculate the reference edge coordinates (X a, Y a ), based on the following specific formula:

[0035]

[0036] According to the reference edge coordinates (X a , Y a ) Calculate the difference S uu The specific formula is as follows:

[0037]

[0038] Set the difference standard value S w ;

[0039] When the difference S uu <Difference standard value S w When the organ tissue to be tested is initially judged to be of this difference degree S uu Corresponding organ tissue type; when the difference S uu >Difference standard value S w , it means that it is not this difference S uu Corresponding organ tissue type.

[0040] In the preferred embodiment of the intelligent processing method for medical ultrasound images, the steps of calculating the ellipticity E and the circularity Eq and further determining the organ tissue type are as follows:

[0041] Based on the coordinates of the reference edge to be measured (X a , Y a ) to obtain the organ tissue area ES and the organ tissue area perimeter ESD, and according to the reference edge coordinates to be measured (X a , Y a ) performing ellipse fitting to obtain the length a of the major semi-axis and the length b of the minor semi-axis of the fitted ellipse;

[0042] The ellipticity E is calculated based on the length of the major semi-axis a and the length of the minor semi-axis b, according to the following formula:

[0043]

[0044] Among them, the value range of the minor semi-axis length b of the fitted ellipse is b>0;

[0045] The circularity Eq is calculated based on the organ tissue area ES and the organ tissue area perimeter ESD. The specific formula is as follows:

[0046]

[0047] Among them, the value range of ESD of the perimeter of the organ tissue area is ESD>0;

[0048] The ellipticity E and circularity Eq are compared with the ellipticity interval threshold and circularity interval threshold of different organs and tissues respectively;

[0049] When the circularity Eq and the ellipticity E meet the ellipticity interval threshold and circularity interval threshold of the same organ tissue, it is judged to belong to this organ tissue type;

[0050] When the results of the preliminary determination of the organ tissue type are consistent with the secondary determination of the organ tissue type, it is finally determined to belong to this organ tissue type. When the results of the preliminary determination of the organ tissue type are inconsistent with the secondary determination of the organ tissue type, the two-dimensional grayscale image data to be tested is re-acquired and the determination is repeated.

[0051] In the preferred embodiment of the above-mentioned medical ultrasonic image intelligent processing method:

[0052] The method to determine whether there are abnormal coordinates of gray value is:

[0053] When the gray value of the coordinate point in the two-dimensional gray image to be tested is greater than the gray value threshold Ew of the lesion, the coordinate is determined to be a gray value abnormal coordinate;

[0054] The methods for determining whether there is a lesion are:

[0055] For the reference edge coordinates to be measured (X c , Y c ) coordinate point (X j , Y j ) and the hypoechoic nodule boundary coordinates (xc g , yc g ) coordinate point (xc m , yc m ), when the coordinate point (X j , Y j ) and the coordinate point (xc m , yc m ) and the reference edge coordinate data 2 (X c , Y c ) intersect, and (xc m , yc m ) is located inside the polygon formed by the reference edge coordinate data 2, it is judged that there is a lesion, and the hypoechoic nodule boundary coordinate (xc g ,yc g ) are considered as the boundary coordinates of the lesion area;

[0056] The method for calculating the lesion area size A is:

[0057] According to the coordinates of the hypoechoic nodule boundary (xc g ,yc g), calculate the size of the lesion area A, the specific formula is as follows:

[0058]

[0059] Among them, xc g represents the g-th coordinate point on the x-axis of the hypoechoic nodule boundary, g represents the coordinate point number on the hypoechoic nodule boundary, and its value is [1, k], where k is a positive integer, xc g+1 represents the x-axis g+1 coordinate point on the boundary of the hypoechoic nodule, yc g represents the g-th coordinate point on the y-axis of the hypoechoic nodule boundary, yc g+1 Represents the g+1th coordinate point on the y-axis on the boundary of the hypoechoic nodule.

[0060] In the preferred embodiment of the intelligent processing method for medical ultrasonic images, the specific steps of calculating the comprehensive structure texture index STI are as follows:

[0061] The specific steps for calculating the comprehensive structural texture index STI are:

[0062] Texture structure difference data include contrast Contrast, grayscale correlation Cor, energy value Eu, homogeneity Hu, moment of inertia IDu and entropy Entropy;

[0063] The two-dimensional grayscale image test data also includes the grayscale value er of the normal coordinate point, the grayscale value jd of the abnormal coordinate point, and the standard deviation σ of the grayscale value of the normal coordinate point er , the standard deviation of the gray value of the abnormal coordinate point σ jd 、The mean gray value μ of normal coordinate points er , mean gray value μ of abnormal coordinate points jd ;

[0064] Obtain the normal-abnormal coordinate point pair frequency P(er, jd) according to the gray level co-occurrence matrix;

[0065] According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the normal-abnormal coordinate point pair frequency P (er, jd), the contrast Contrast is calculated based on the following specific formula:

[0066]

[0067] According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the gray value mean μ of the normal coordinate point er , mean gray value μ of abnormal coordinate points jd As well as the normal-abnormal coordinate point pair frequency P(er, jd), the grayscale correlation Cor is calculated based on the following specific formula:

[0068]

[0069] According to the frequency of normal-abnormal coordinate point pairs P(er, jd), the entropy is calculated based on the following specific formula:

[0070]

[0071] According to the normal-abnormal coordinate point frequency P (er, jd), the energy value Eu is calculated based on the following specific formula:

[0072]

[0073] Among them, the gray value er is [1,200], and the gray value jd is [201,255];

[0074] The homogeneity Hu is calculated based on the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the frequency P (er, jd) of the normal-abnormal coordinate point pair. The specific formula is as follows:

[0075]

[0076] According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the normal-abnormal coordinate point pair frequency P(er, jd), the moment of inertia IDu is calculated based on the following specific formula:

[0077]

[0078] According to the texture structure difference data including contrast Contrast, grayscale correlation Cor, energy value Eu, homogeneity Hu, moment of inertia IDu and entropy Entropy, the comprehensive structure texture index STI is calculated. The specific formula is as follows:

[0079]

[0080] In the preferred embodiment of the medical ultrasound image intelligent processing method, the color Doppler blood flow imaging data includes the transmitting frequency fb, the receiving frequency f S , the angle θ between the ultrasound beam and the blood flow direction, the peak blood flow velocity V during systole ma and end-diastolic blood flow velocity V mi ;

[0081] According to the transmitting frequency fb, receiving frequency f S The angle θ between the ultrasound beam and the blood flow direction is used to calculate the blood flow velocity Cv. The specific formula is as follows:

[0082]

[0083] Wherein, df represents the influence coefficient of color Doppler blood flow imaging data, and its value is [0.1, 10];

[0084] According to the peak systolic blood flow velocity V ma and end-diastolic blood flow velocity V mi , calculate the blood flow resistance index RII, the specific formula is as follows:

[0085]

[0086] When (fb-fs)>0, it means that the blood flows toward the probe; when (fb-fs)<0, it means that the blood flows away from the probe.

[0087] In the preferred embodiment of the medical ultrasound image intelligent processing method, the specific steps of calculating the comprehensive blood flow assessment index MPI are as follows:

[0088] Patient history information data including past medical history index QS si , Patient age index BC Z and patient gender index BB Z ;

[0089] According to blood flow velocity Cv, blood flow resistance index RII, blood flow direction TIF, medical history impact index QS si , Gender Impact Index BC Z And age impact index BB Z , calculate the comprehensive blood flow assessment index MPI, the specific formula is as follows:

[0090]

[0091] Where d is the weight factor of blood flow velocity Cv, which is 0.1≤d≤0.3, g is the weight factor of blood flow resistance index RII, which is 0.2≤g≤0.4, p is the weight factor of blood flow direction TIF, which is 0.3≤p≤0.5, QS si is the influence index of the si-th medical record of the patient. The value range of si is 1, 2, 3...sn. Sn is the total number of medical records of the medical history data Q. The value is a positive integer. CX si is the weight coefficient of the si-th medical record, and its value is 0.1<CX si <0.3, and d+g+p+CX si =1.

[0092] In the preferred embodiment of the medical ultrasound image intelligent processing method, the specific method for calculating the evaluation value CEV of the medical record to be tested is:

[0093] Patient history information data also includes medical record evaluation value GCX uu;

[0094] Set the reference value μ for the lesion area size S , comprehensive structural texture index reference value μ STI And the comprehensive blood flow assessment index reference value μ CFEI ;

[0095] According to the known medical records, the size of the lesion area A, the comprehensive structural texture index STI, the comprehensive blood flow assessment index MPI, and the reference value of the lesion area size μ S , comprehensive structural texture index reference value μ STI And the comprehensive blood flow assessment index reference value μ CFEI , calculate the CEV of the medical record to be tested, the specific formula is as follows:

[0096]

[0097] Among them, β1 is the weight index of the lesion area size A, and the value of β1 is 0.1<β1≤0.3; β2 is the weight index of the comprehensive structure texture index STI, and the value is 0.3≤β2≤0.6; β3 is the weight index of the comprehensive blood flow assessment index MPI, and the value is 0.2≤β3≤0.5, and β1+β2+β3=1; GCX uu represents the evaluation value of the uuth medical record, the value range of uu is 1, 2, 3...jo, jo is the total number of medical records, and its value is a positive integer;

[0098] Set a set of symptom assessment thresholds Dv, which includes a primary threshold Dv0 and a secondary threshold Dv1, where Dv0 <Dv1;

[0099] When the symptom evaluation value CEV≤≤Dv0, the indicator is judged as a mild pathological feature and a first-level diagnosis report is generated;

[0100] When Dv0<symptom assessment value CEV≤Dv1, the indicator is judged as moderate pathological characteristics, and a secondary diagnosis report is generated;

[0101] When Dv1<symptom assessment value CEV, the indicator is judged as severe pathological characteristics and a third-level diagnosis report is generated.

[0102] The present invention also discloses a medical ultrasonic image intelligent processing system, comprising:

[0103] The data acquisition module acquires normal ultrasound image data, which includes normal image samples of multiple normal tissues or organs, and acquires normal data of two-dimensional grayscale images of normal image samples; acquires two-dimensional grayscale image test data and color Doppler blood flow imaging data of image samples to be processed; and acquires historical medical record information data of patients;

[0104] The preliminary diagnosis module is used to extract the normal initial edge coordinate points (x i ,y i ), according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu ); extract the initial edge coordinates (x) to be measured from the two-dimensional grayscale image to be measured a ,y a ), and calculate the coordinates of the reference edge to be measured (X a , Y a );Calculate the coordinates of the reference edge to be measured (X a , Y b ) and the normal characteristic coordinates (X uu , Y uu ) uu , make a preliminary judgment on the type of organ tissue;

[0105] The secondary diagnosis module, based on the coordinates of the reference edge to be tested (X a , Y a ) Calculate the ellipticity E and circularity Eq of the organ tissue; obtain the ellipticity interval threshold and circularity interval threshold of different organ tissues, compare the ellipticity E and circularity Eq with the ellipticity interval threshold and circularity interval threshold of different organ tissues, perform secondary judgment on the type of the organ tissue according to the comparison result, and finally determine the organ tissue type of the image sample to be processed;

[0106] The data processing module is used to obtain the gray value of each coordinate point in the two-dimensional gray image to be tested; set the lesion gray threshold Ew, compare the gray value of each coordinate point with the lesion gray threshold Ew, and determine whether there is an abnormal gray value coordinate. All abnormal coordinate points constitute an abnormal area, and the boundary coordinates of the abnormal area form the initial edge coordinate point to be tested (x c ,y c ), according to the initial edge coordinate point 2 to be measured (x c ,y c ) Calculate the coordinates of the reference edge to be measured (X c , Y c ); Get the boundary coordinates of the hypoechoic nodule from the image sample to be processed (xc g ,yc g ), according to the coordinates of the reference edge to be measured (X c , Y c ) and the hypoechoic nodule boundary coordinates (xc g , yc g), to determine whether there is a lesion in the organ tissue; when it is determined that there is a lesion, according to the coordinates of the reference edge to be measured (X c , Y c ) calculate the size A of the lesion area; obtain a grayscale co-occurrence matrix according to the two-dimensional grayscale image data to be tested, calculate the texture structure difference data inside the tissue according to the grayscale co-occurrence matrix and the two-dimensional grayscale image data to be tested; calculate the comprehensive structure texture index STI according to the texture structure difference data; calculate the blood flow velocity Cv and the blood flow resistance index RII according to the color Doppler blood flow imaging data, and determine the blood flow direction TIF, and calculate the comprehensive blood flow evaluation index MPI according to the blood flow velocity Cv, the blood flow resistance index RII, the blood flow direction TIF and the patient's historical medical record information data;

[0107] The summary report module is used to calculate the disease assessment value CEV based on the known medical record data, the size A of the lesion area, the comprehensive structural texture index STI and the comprehensive blood flow assessment index MPI, compare the disease assessment value CEV and the disease assessment threshold set Dv, combine the preliminary diagnosis results, judge the disease level, and automatically generate a diagnosis report.

[0108] (III) Beneficial effects

[0109] The present invention provides a medical ultrasound image intelligent processing method and system, which has the following beneficial effects:

[0110] (1) Comprehensively collect normal ultrasound image data, image data to be processed, and patient medical history information to provide a rich data basis for subsequent analysis and diagnosis. Normal image samples of multiple normal tissues or organs help to establish accurate normal feature coordinate standards. Two-dimensional grayscale image data and color Doppler blood flow imaging data provide key information in terms of morphology and blood flow, respectively. Patient medical history information can be combined with individual conditions for more accurate diagnosis.

[0111] (2) By extracting the normal initial edge coordinate points and calculating the normal feature coordinate standard, a reference basis is provided for judging the organ tissue type of the image to be processed. The coordinate points of the image to be tested are extracted and compared with the normal standard to make a preliminary judgment on the difference, which reduces the uncertainty of subjective judgment, improves the objectivity and accuracy of diagnosis, and lays the foundation for subsequent further analysis.

[0112] (3) Shape feature parameters such as ellipticity and circularity are calculated and compared with specific thresholds for secondary judgment. The organ and tissue type can be further confirmed from the perspective of shape features. This can make up for the shortcomings of judgment based solely on coordinate differences and improve the reliability of diagnosis, especially for some organs and tissues with obvious shape features, which can be identified more accurately.

[0113] (4) Automatically segment the lesion area and calculate the size, which can quantitatively analyze the degree of lesions and provide important indicators for disease assessment. It calculates the texture structure difference value and the comprehensive structure texture index through the gray-level co-occurrence matrix, deeply analyzes the internal characteristics of the tissue, reflects the uniformity and complexity of the tissue, calculates the blood flow velocity, blood flow resistance index and comprehensive blood flow assessment index, comprehensively evaluates the health status of blood vessels and lesions, and provides multi-dimensional information for comprehensive judgment of diseases.

[0114] (5) The disease assessment value is calculated based on the known medical record data and ultrasound image data, and the disease level is determined based on the preliminary diagnosis results and a diagnosis report is automatically generated, which improves the efficiency and accuracy of diagnosis, provides doctors with detailed diagnostic conclusions and treatment recommendations, and facilitates patients to understand their own condition and examination results, thus promoting doctor-patient communication and the scientific nature of medical decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] Figure 1 The present invention is a flowchart of a medical ultrasound image intelligent processing method and system. DETAILED DESCRIPTION

[0116] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0117] See also Figure 1 The present invention provides a medical ultrasound image intelligent processing method, comprising:

[0118] Acquire normal ultrasound image data, which includes normal image samples of multiple normal tissues or organs, and acquire normal two-dimensional grayscale image data of normal image samples; acquire two-dimensional grayscale image test data and color Doppler blood flow imaging data of image samples to be processed; and acquire historical medical record information data of patients.

[0119] Specifically, ultrasonic image data of normal organ tissues are obtained based on clinical medical cases and medical literature, which provides rich reference basis for the subsequent establishment of accurate normal feature coordinates. The two-dimensional grayscale image normal data of normal image samples can be obtained through image processing software lmage or Fiji. The two-dimensional grayscale image can present the grayscale level of the tissue, reflecting the structure and density of the tissue and other information. The two-dimensional grayscale image test data and color Doppler blood flow imaging data of the image samples to be processed can also be obtained through image processing software 1mage or Fiji. The two-dimensional grayscale image test data is used to analyze the morphology and grayscale characteristics of the tissue to be processed, and the color Doppler blood flow imaging data can provide the blood flow conditions inside the tissue, providing multi-dimensional information for comprehensive evaluation of the health status of the tissue. Obtaining the patient's historical medical information data is helpful for more accurate diagnosis and disease assessment based on the patient's individual situation.

[0120] Through the two-dimensional grayscale image normal data of different normal image samples, the normal initial edge coordinate points (x i ,y i ), according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu ).

[0121] Specifically, the normal data of two-dimensional grayscale images of different normal image samples can be used through existing tools such as image processing libraries, such as Opencv open source computer vision library, Scikit-image image processing library; or edge detection algorithms, such as Canny edge detection; or MATLAB automation tools; or U-Net convolutional neural network architecture and FCN full convolutional network, etc., to extract normal initial edge coordinate points. These coordinate points represent the edge position of normal tissues or organs and reflect their morphological characteristics. Based on these normal initial edge coordinate points, the normal feature coordinates of different organ tissues are calculated. These normal feature coordinates can be used as a reference standard for judging whether the organ tissue in the image to be processed is normal.

[0122] The initial edge coordinates (x a ,y a ), and calculate the coordinates of the reference edge to be measured (X a , Y a );Calculate the coordinates of the reference edge to be measured (X a , Y b ) and the normal characteristic coordinates (X uu , Y uu ) uu , make a preliminary judgment on the type of organ tissue.

[0123] Specifically, the two-dimensional grayscale image data to be tested can be extracted through existing tools such as image processing libraries, such as Opencv open source computer vision library, Scikit-image image processing library; or edge detection algorithms, such as Canny edge detection; or MATLAB automation tools; or U-Net convolutional neural network architecture and FCN full convolutional network, to calculate the initial edge coordinates to be tested and the reference edge coordinates to be tested. This coordinate is a more accurate description of the edge of the tissue to be tested. The difference between the reference edge coordinates to be tested and the normal characteristic coordinates of different organ tissues is calculated. By comparing the difference, the type of organ tissue can be preliminarily judged to determine which organ type the tissue to be processed may belong to.

[0124] According to the coordinates of the reference edge to be measured (X a , Y a ) calculate the ellipticity E and circularity Eq of the organ tissue; obtain the ellipticity interval threshold and circularity interval threshold of different organ tissues, compare the ellipticity E and circularity Eq with the ellipticity interval threshold and circularity interval threshold of different organ tissues, respectively, make a secondary judgment on the type of the organ tissue according to the comparison result, and finally determine the organ tissue type of the image sample to be processed.

[0125] Specifically, the ellipticity and circularity of the organ tissue are calculated according to the coordinates of the reference edge to be measured. Ellipticity and circularity are parameters that describe the shape characteristics of the tissue, which can reflect whether the morphology of the tissue is regular. The ellipticity interval threshold and circularity interval threshold of different organ tissues are obtained. These thresholds are obtained through statistical analysis of a large number of normal organ tissues. The ellipticity and circularity are compared with the ellipticity interval threshold and circularity interval threshold of different organ tissues. According to the comparison results, it can be further confirmed whether the shape characteristics of the tissue meet the characteristics of a specific organ tissue, so as to make a secondary judgment on the type of the organ tissue, and finally determine the organ tissue type of the image sample to be processed.

[0126] Get the gray value of each coordinate point in the two-dimensional gray image to be tested; set the gray threshold Ew of the lesion, compare the gray value of each coordinate point with the gray threshold Ew of the lesion, and determine whether there is an abnormal coordinate with gray value. All abnormal coordinate points constitute an abnormal area, and the boundary coordinates of the abnormal area form the initial edge coordinate point to be tested (x c ,y c ), according to the initial edge coordinate point 2 to be measured (x c ,y c ) Calculate the coordinates of the reference edge to be measured (X c , Y c ); Get the boundary coordinates of the hypoechoic nodule from the image sample to be processed (xc g, yc g ), according to the coordinates of the reference edge to be measured (X c , Y c ) and the hypoechoic nodule boundary coordinates (xc g , yc g ) to determine whether there is any lesion in the organ tissue.

[0127] Specifically, existing tools such as image processing libraries, such as Opencv open source computer vision library, Scikit-image image processing library; or edge detection algorithms, such as Canny edge detection; or MATLAB automation tools; or U-Net convolutional neural network architecture and FCN full convolutional network, etc., can be used to obtain the grayscale value of each coordinate point in the two-dimensional grayscale image to be tested, and set the lesion grayscale threshold. This threshold is determined based on the difference in grayscale values ​​between normal tissue and lesion tissue. The specific data of a specific medical case can be referred to, and the grayscale value of each coordinate point is compared with the lesion grayscale threshold to determine whether there is an abnormal grayscale value coordinate. If the grayscale value of a coordinate point exceeds the normal range, it is considered to be an abnormal grayscale value coordinate. All abnormal coordinate points constitute an abnormal area, which may be a potential lesion area. The coordinates of the boundary of the abnormal area are obtained to form coordinate point 2, which provides a basis for further analysis of the lesion area. When there are abnormal grayscale value coordinates, all abnormal coordinate points constitute an abnormal area, and the initial edge coordinate point 2 to be tested (x) is formed according to the boundary coordinates of the abnormal area. c ,y c ) Calculate the coordinates of the reference edge to be measured (X c , Y c ), and then according to the coordinates of the reference edge to be measured (X c , Y c ) and the hypoechoic nodule boundary coordinates (xc g , yc g ), the presence of lesions can be further confirmed.

[0128] When a lesion is determined, the boundary coordinates of the hypoechoic nodule (xc g , yc g ) Calculate the size A of the lesion area.

[0129] Specifically, hypoechoic nodules usually appear in ultrasound images as areas with lower echo intensity compared with surrounding normal tissues. The boundary coordinates of these nodules are crucial for determining the extent of the lesion. By carefully observing and measuring the ultrasound images, the coordinate points (xcg, ycg) on ​​the boundary of the hypoechoic nodules can be obtained. These coordinate points constitute the outline of the lesion area.

[0130] According to the two-dimensional grayscale image data to be tested, the grayscale co-occurrence matrix is ​​obtained through existing tools such as Scikit-image image processing library or MATLAB automation tools. According to the grayscale co-occurrence matrix and the two-dimensional grayscale image data to be tested, the texture structure difference data inside the tissue is calculated; according to the texture structure difference data, the comprehensive structure texture index STI is calculated.

[0131] Specifically, a grayscale co-occurrence matrix is ​​obtained based on the two-dimensional grayscale image data to be tested. The grayscale co-occurrence matrix reflects the frequency of occurrence of pixel pairs with different grayscale values ​​in the image in a specific direction and distance. Grayscale values ​​and grayscale values ​​are obtained based on the grayscale co-occurrence matrix. The grayscale value here can be a specific element value in the matrix or a statistic obtained by calculating the matrix. Based on the two-dimensional grayscale image data, the texture structure difference data inside the tissue is calculated. The texture structure difference data may include parameters such as contrast, homogeneity, and energy. These parameters reflect the texture characteristics of the tissue. Based on the texture structure difference data, a comprehensive structure texture index is calculated through a specific calculation formula. The comprehensive structure texture index can comprehensively reflect the characteristics of the texture complexity and uniformity of the tissue, and provide an important basis for judging the pathological condition of the tissue.

[0132] According to the color Doppler blood flow imaging data, the blood flow velocity Cv and the blood flow resistance index RII are calculated, and the blood flow direction TIF is determined. According to the blood flow velocity Cv, the blood flow resistance index RII, the blood flow direction TIF and the patient's historical medical information data, the comprehensive blood flow assessment index MPI is calculated.

[0133] According to the color Doppler blood flow imaging data, the blood flow velocity and blood flow resistance index are calculated and the blood flow direction is determined using the principle of the Doppler effect. The blood flow velocity reflects the flow speed of blood in the tissue, the blood flow resistance index reflects the patency and resistance of the blood vessels, and the blood flow direction can help determine the relationship between the flow direction of the blood vessels and the location of the lesion. According to the blood flow velocity, blood flow resistance index, blood flow direction and the patient's historical medical information data, the comprehensive blood flow assessment index is calculated through a specific mathematical model. The comprehensive blood flow assessment index can comprehensively reflect the blood perfusion of the tissue and the functional status of the blood vessels, and provide important information for evaluating the impact and severity of the lesion.

[0134] The disease evaluation value CEV is calculated based on the known medical record data, the size A of the lesion area, the comprehensive structural texture index STI and the comprehensive blood flow evaluation index MPI. The disease evaluation value CEV and the disease evaluation threshold set Dv are compared. Combined with the preliminary diagnosis results, the disease level is judged and a diagnosis report is automatically generated. According to the known medical record data, the size of the lesion area, the comprehensive structural texture index and the comprehensive blood flow evaluation index, the disease evaluation value is calculated using statistical analysis and machine learning methods. The disease evaluation value is a comprehensive quantitative assessment of the patient's condition. The disease evaluation value and the disease evaluation threshold set are compared, and the severity level of the disease is determined based on the comparison results. Combined with the preliminary diagnosis results, the patient's disease is comprehensively analyzed and judged. Finally, a diagnosis report is automatically generated, which includes the patient's basic information, examination results, disease analysis, diagnosis conclusions and treatment recommendations, etc., to provide decision support for doctors and facilitate patients to understand their own conditions.

[0135] In the preferred embodiment of the above-mentioned intelligent processing method for medical ultrasound images: according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu ) are as follows:

[0136] According to the normal initial edge coordinate point (x i ,y i ) Calculate the normal mean coordinates (Xi mean , Yi mean ), based on the following formula:

[0137]

[0138] Among them, x i Indicates the x-axis value of the i-th edge coordinate point, y i Represents the y-axis value of the i-th edge coordinate point. The value range of i is 1, 2, 3...n, and n represents the total number of initial far-field coordinate points.

[0139] It should be noted that for a given number of normal tissues or organs, the normal initial edge coordinate points (x i ,y i ), by summing the x-axis values ​​and y-axis values ​​of all coordinate points, and then dividing by the total number of coordinate points n, we get the normal mean coordinate (Xi mean , Yi mean ), this mean coordinate represents the average position of all normal initial edge coordinate points in the x and y directions, reflecting the central trend of the edge position of the organ tissue under normal circumstances.

[0140] According to the normal initial edge coordinate point (xi ,y i ) and normal mean coordinates (Xi mean , Yi mean ), calculate the coordinate standard deviation (X m , Y m ), the specific formula is:

[0141]

[0142] It should be noted that the deviation between each normal initial edge coordinate point and the normal mean coordinate is calculated, that is, (x i -Xi mean ) and (y i -Yi mean ), the square of the deviation represents the amplification of the degree of deviation, the square of the deviation of all coordinate points is summed and divided by the total number of coordinate points n, the average square of the deviation is obtained, and then the square root is taken to obtain the coordinate standard deviation X m and Y m The standard deviation reflects the degree of dispersion of the normal initial edge coordinate points in the x and y directions relative to the mean coordinates, that is, the fluctuation range of the edge position of organ tissue under normal circumstances.

[0143] According to the coordinate standard deviation (X m , Y m ) and normal mean coordinates (Xi mean , Yi mean ) Calculate the normal feature coordinates (X uu , Y uu ), based on the following specific formula:

[0144] X uu =(X mean -kX m , X mean +k m )

[0145] Y uu =(Y mean -kY m , Y mean +kY m )

[0146] It should be noted that, with the normal mean coordinate as the center, by adding or subtracting a multiple k multiplied by the standard deviation of the coordinates, an interval range is obtained as the normal characteristic coordinates. This interval represents the reasonable value range of the edge coordinates of the organ tissue in the x and y directions under normal circumstances. The multiple k can be adjusted according to actual conditions and is usually used to determine a confidence interval. For example, when k = 2, it covers approximately 95% of the normal data range. If the initial edge coordinates to be measured in the image to be processed fall within this range, it can be preliminarily determined that the tissue belongs to the normal type; if it exceeds this range, there may be an abnormality.

[0147] By calculating the normal mean coordinates, this formula can integrate the edge coordinate information of multiple normal tissues and determine a representative center position, providing a benchmark for subsequent judgments; the calculation of the coordinate standard deviation can quantify the degree of discreteness of the edge coordinates of normal tissues, reflecting the fluctuation range under normal conditions. Combined with the mean coordinates, it can more comprehensively describe the characteristics of the edge position of normal tissues; the determination of normal characteristic coordinates is based on the mean and standard deviation to form a reasonable value range. When judging the organ tissue type, it can accurately and preliminarily judge whether the tissue to be tested is normal, thereby improving the accuracy and objectivity of the diagnosis. At the same time, this method provides a reliable reference standard for subsequent further analysis and judgment, helps to more efficiently perform intelligent processing of medical ultrasound images, and provides strong support for clinical diagnosis and treatment.

[0148] In the preferred embodiment of the medical ultrasound image intelligent processing method, the reference edge coordinate data point (X a , Y a ), and the specific steps for preliminarily determining the organ tissue type are:

[0149] Extract the initial edge coordinate point (x a ,y a ), and obtain the initial edge coordinate point to be measured (x a ,y a ) in the x-axis direction man , the minimum value x in the x-axis direction min , maximum value y in the y-axis direction man and the minimum value y in the y-axis direction min .

[0150] According to the coordinate data point of the benchmark edge to be measured (x a ,y a ), calculate the coordinate mean (Xm ean , Ym ean ), the specific formula is:

[0151]

[0152] Among them, xa Indicates the value of the ath coordinate on the x-axis, y b It represents the value of the b-th coordinate on the y-axis. The value range of a is 1, 2, 3...x, and x represents the total number of coordinates.

[0153] It should be noted that first, the initial edge coordinate point to be measured (x a ,y a ), and determine the maximum value x in the x-axis direction man , the minimum value x in the x-axis direction min , maximum value y in the y-axis direction man and the minimum value y in the y-axis direction min These extreme values ​​can define the range of the area to be measured in the direction and then calculate the coordinate mean (X mean , Y mean ), sum up the x and y values ​​of all the initial edge coordinate points to be measured and divide them by the total number of coordinates to get the average position in the x and y directions, which reflects the central trend of the edge coordinates of the area to be measured.

[0154] According to the maximum value x in the x-axis direction man , the minimum value x in the x-axis direction min , maximum value y in the y-axis direction man and the minimum value y in the y-axis direction min The specific formula for calculating the coordinate value normalization factor d is as follows:

[0155]

[0156] It should be noted that this formula calculates a normalization factor based on the span of the area to be measured in the x and y directions. It is calculated by calculating the span in the x direction (x man -x min ) and the span in the y direction (y man -y min ). This factor can be used to normalize coordinate values ​​to a specific range for subsequent calculations and comparisons.

[0157] According to the initial edge coordinate point to be measured (x a ,y a ), coordinate mean (X mean , Y mean ) and the coordinate value normalization factor d, calculate the reference edge coordinates (X a , Y a ), based on the following specific formula:

[0158]

[0159] It should be noted that these two formulas are to subtract the coordinate mean from the x value and y value of each initial edge coordinate point to be measured, and then divide by the normalization factor d. The purpose of this is to normalize the coordinate values ​​so that the coordinates of the measured areas of different sizes and positions can be compared and analyzed on a relatively uniform scale. After such processing, the obtained reference edge coordinates (X a , Y a ) is more comparable and stable, which is conducive to subsequent operations such as difference calculation with normal feature coordinates, so as to more accurately judge the type and status of organ tissue.

[0160] According to the reference edge coordinates (X a , Y a ) Calculate the difference S uu The specific formula is as follows:

[0161]

[0162] Set the difference standard value S w .

[0163] When the difference S uu <Difference standard value S w When the organ tissue to be tested is initially judged to be of this difference degree S uu Corresponding organ tissue type: When the difference S uu >Difference standard value S w , it means that it is not this difference S uu Corresponding organ tissue type.

[0164] It should be noted that X a and Y a is the reference edge coordinate obtained by processing the image data to be measured, X uu and Y uu It is the normal characteristic coordinates of different organs and tissues calculated by normal image data. This formula calculates the Euclidean distance between the coordinates to be measured and the normal characteristic coordinates on a two-dimensional plane. The principle is to measure the degree of deviation between the edge position of the tissue to be measured and the edge position of the normal tissue by comparing the difference between the two coordinate points in the x and y directions.

[0165] Furthermore, the standard value of the difference is S w By analyzing a large amount of existing normal and abnormal ultrasound image data, the differences between the normal characteristic coordinates of different organ tissues and various abnormal conditions are calculated. By performing statistical analysis on these difference data and calculating the mean, a reasonable difference standard value is determined and a difference standard value S is set. w , when the calculated difference S uu When S is less than the standard value of differencew , which means that the edge position of the tissue to be tested is close to the edge position of a normal organ tissue, and the organ tissue to be tested is preliminarily judged to be the organ tissue type corresponding to this difference. uu Greater than the difference standard value S w When , it means that the edge position of the tissue to be tested deviates greatly from the edge position of the normal organ tissue, which means that it is not the organ tissue type corresponding to this difference. In this way, by setting a threshold for judgment, the organ tissue type can be quickly screened preliminarily, providing a basis for subsequent more accurate judgment.

[0166] By extracting the maximum value of the coordinates and calculating the mean value of the coordinates, this formula can quickly determine the approximate range and center position of the area to be tested, providing a basic reference for subsequent analysis. At the same time, the coordinate mean reflects the central trend of the edge coordinates to be tested, which helps to preliminarily understand the overall position characteristics of the tissue to be tested. The coordinate value normalization factor can be calculated to normalize the coordinates of the areas to be tested of different sizes and positions, so that different images have a unified scale when analyzed, which improves the versatility and comparability of the method. The normalized formula is used to calculate the reference edge coordinates, so that different image data to be tested can be compared under the same standard, enhancing the accuracy and objectivity of the judgment of organ tissue types. The use of this series of formulas helps to more efficiently perform intelligent processing of medical ultrasound images, provide a more reliable basis for clinical diagnosis, and improve diagnostic efficiency and accuracy.

[0167] In the preferred embodiment of the intelligent processing method for medical ultrasound images, the steps of calculating the ellipticity E and the circularity Eq and further determining the organ tissue type are as follows:

[0168] Based on the coordinates of the reference edge to be measured (X a , Y a ) to obtain the organ tissue area ES and the organ tissue area perimeter ESD, and according to the reference edge coordinates to be measured (X a , Y a ) performs ellipse fitting and obtains the major semi-axis length a and the minor semi-axis length b of the fitted ellipse.

[0169] The ellipticity E is calculated based on the length of the major semi-axis a and the length of the minor semi-axis b, according to the following formula:

[0170]

[0171] The value range of the length b of the minor semi-axis of the fitted ellipse is b>0.

[0172] It should be noted that, firstly based on (X a , Y a) The coordinates of the reference edge to be measured are used to obtain the organ tissue area ES and the organ tissue area perimeter ESD, and an ellipse is fitted to obtain the length of the major semi-axis a and the length of the minor semi-axis b of the fitted ellipse. The ellipticity is calculated by measuring the ratio of the major semi-axis to the minor semi-axis to measure the degree of closeness between the tissue shape and the ellipse. If the ratio is close to 1, it means that the tissue shape is closer to a circle; if the ratio is larger, it means that the tissue shape is narrower and longer, similar to an ellipse. For different organ tissues, their normal morphology may have different ellipticity characteristic ranges.

[0173] The circularity Eq is calculated based on the organ tissue area ES and the organ tissue area perimeter ESD. The specific formula is as follows:

[0174]

[0175] Among them, the value range of ESD of the organ tissue area perimeter is ESD>0.

[0176] It should be noted that operations such as pixel counting and edge detection can be performed through Python's Scikit-learn and Numpy libraries or MATLAB's image processing tools to calculate the area and perimeter of organ tissues. For ellipse fitting, the linear regression or optimization algorithm module in Scikit-learn can be used, combined with the principle of least squares method to adjust parameters. For example, the optimization function in the scipy.optimize module can be used to minimize the sum of the distances from the coordinate points to the ellipse to determine the parameters of the ellipse equation.

[0177] The principle of this formula is to compare the relationship between the area and circumference of the organ tissue region with that of a standard circle, and deform this relationship to obtain a circularity of 1. For actual organ tissues, the closer the circularity is to 1, the closer the shape is to a circle; the smaller the circularity is, the greater the difference between the shape and the circle is.

[0178] The ellipticity interval thresholds and circularity interval thresholds of different organ tissues are obtained according to medical literature, and the ellipticity E and circularity Eq are compared with the ellipticity interval thresholds and circularity interval thresholds of different organ tissues, respectively.

[0179] When the circularity Eq and the ellipticity E meet the ellipticity interval threshold and circularity interval threshold of the same organ tissue, it is judged to belong to this organ tissue type.

[0180] When the results of the preliminary determination of the organ tissue type are consistent with the secondary determination of the organ tissue type, it is finally determined to belong to this organ tissue type. When the results of the preliminary determination of the organ tissue type are inconsistent with the secondary determination of the organ tissue type, the two-dimensional grayscale image data to be tested is re-acquired and the determination is repeated.

[0181] It should be noted that the calculated ellipticity and circularity are compared with the ellipticity interval threshold and circularity interval threshold of different organ tissues respectively. If the circularity and ellipticity are both within the ellipticity interval threshold and circularity interval threshold range of the same organ tissue, it is judged to belong to this organ tissue type. This is because different organ tissues usually have specific shape feature ranges. By comparing the shape features of the tissue to be tested with these ranges, the type of organ tissue can be further confirmed.

[0182] When the results of the preliminary judgment of the organ tissue type are consistent with the results of the secondary judgment of the organ tissue type, it is determined to belong to this organ tissue type, which improves the accuracy and reliability of the judgment. If the preliminary judgment and the secondary judgment results are inconsistent, the two-dimensional grayscale image data to be tested is re-acquired for repeated judgment to ensure the accuracy of the diagnosis and avoid erroneous judgments due to data errors or algorithm instability.

[0183] Calculating ellipticity and circularity can quantitatively analyze organ tissues from the perspective of shape characteristics. Ellipticity can reflect the similarity between tissue shape and an ellipse, while circularity can be compared with a standard circle. The combination of the two provides a new dimension for determining organ tissue types. Different organ tissues usually have a specific shape range in their normal morphology. These formulas can be used to more accurately identify tissue types. Comparing ellipticity and circularity with specific thresholds for secondary judgment can make up for the shortcomings of initial judgments based solely on edge coordinate differences. When the initial judgment is consistent with the secondary judgment results, the accuracy and reliability of the diagnosis can be improved. If they are inconsistent, re-acquiring data for repeated judgment can help reduce misjudgments and ensure the accuracy of diagnostic results. This will provide stronger support for clinical diagnosis and improve the quality and effectiveness of intelligent processing of medical ultrasound images.

[0184] In the preferred embodiment of the above-mentioned medical ultrasonic image intelligent processing method:

[0185] When the grayscale value of the coordinate point in the two-dimensional grayscale image to be tested is greater than the lesion grayscale threshold Ew, the coordinate is determined to be a grayscale value abnormal coordinate.

[0186] It should be noted that when the grayscale value > lesion grayscale threshold Ew and the grayscale value boundary coordinates conform to the boundary coordinate area of ​​the low-echo nodule in the ultrasound image, it is judged that the organ tissue has a lesion; when the grayscale value < lesion grayscale threshold Ew and the grayscale value boundary coordinates do not conform to the boundary coordinate area of ​​the low-echo nodule in the ultrasound image, the two-dimensional grayscale image data to be tested is re-acquired and the low-echo nodule inspection in the ultrasound image is performed, and repeated judgment is performed; in this way, the presence of lesions is comprehensively judged by the two characteristics of grayscale value and low-echo nodule, thereby improving the accuracy of judgment.

[0187] The methods for determining whether there is a lesion are:

[0188] For the reference edge coordinates to be measured (X c , Y c ) coordinate point (X j , Y j ) and the hypoechoic nodule boundary coordinates (xc g , yc g ) coordinate point (xc m , yc m ), when the coordinate point (X j , Y j ) and the coordinate point (xc m , yc m ) and the reference edge coordinate data 2 (X c , Y c ) intersect, and (xc m , yc m ) is located inside the polygon formed by the reference edge coordinate data 2, it is judged that there is a lesion, and the coordinate (xc g , yc g ) are considered as the boundary coordinates of the lesion area.

[0189] The method for calculating the lesion area size A is:

[0190] According to the coordinates of the hypoechoic nodule boundary (xc g , yc g ), calculate the size of the lesion area A, the specific formula is as follows:

[0191]

[0192] Among them, xc g represents the g-th coordinate point on the x-axis of the hypoechoic nodule boundary, g represents the coordinate point number on the hypoechoic nodule boundary, and its value is [1, k], where k is a positive integer, xc g+1 represents the x-axis g+1 coordinate point on the boundary of the hypoechoic nodule, yc g represents the g-th coordinate point on the y-axis of the hypoechoic nodule boundary, yc g+1 Indicates the g+1th coordinate point on the y-axis on the boundary of the hypoechoic nodule.

[0193] It should be noted that the principle of this formula is similar to the method of calculating the area of ​​a polygon. The area of ​​the lesion is obtained by cross-producting the x and y coordinates of adjacent coordinate points and then taking half of the absolute value. This method is suitable for calculating lesion areas of irregular shapes and can accurately quantify the extent of the lesion.

[0194] This step judges the lesion by combining the grayscale value and the boundary coordinates of the low-echo nodule, thereby improving the accuracy of lesion judgment. This multi-feature judgment method can reduce misjudgment and more reliably identify diseased tissue. When determining the coordinates of the lesion area, the automated processing of this method reduces the errors and subjectivity of manual measurement, improves the efficiency and reliability of intelligent processing of medical ultrasound images, and provides strong support for clinical diagnosis and treatment.

[0195] In the preferred embodiment of the intelligent processing method for medical ultrasonic images, the specific steps of calculating the comprehensive structure texture index STI are as follows:

[0196] The texture structure difference data include contrast Contrast, grayscale correlation Cor, energy value Eu, homogeneity Hu, moment of inertia IDu and entropy Entropy.

[0197] When the grayscale value of the coordinate point in the two-dimensional grayscale image to be tested data is greater than the lesion grayscale threshold Ew, it is the abnormal coordinate point grayscale value jd; when the grayscale value of the coordinate point in the two-dimensional grayscale image to be tested data is less than the lesion grayscale threshold Ew, it is the normal coordinate point grayscale value er. Because of the degree of pathology and the color depth of the image, the normal coordinate point grayscale value er and the abnormal coordinate point grayscale value jd are set with ranges. For example, the normal coordinate point grayscale value er can range from [1,200], and the abnormal coordinate point grayscale value jd can range from [201,255]. This makes it easy to distinguish the lesion area and lays a data foundation for subsequent calculations.

[0198] The two-dimensional grayscale image test data also includes the grayscale value er of the normal coordinate point, the grayscale value jd of the abnormal coordinate point, and the standard deviation σ of the grayscale value of the normal coordinate point er , the standard deviation of the gray value of the abnormal coordinate point σ jd 、The mean gray value μ of normal coordinate points er , mean gray value μ of abnormal coordinate points jd .

[0199] The two-dimensional grayscale image data also includes the standard deviation of the grayscale values ​​of the normal coordinate points σ er , the standard deviation of the gray value of the abnormal coordinate point σ jd 、The mean gray value μ of normal coordinate points er , mean gray value μ of abnormal coordinate points jd .

[0200] First, calculate the mean gray value of the normal coordinate point. For each coordinate point in the normal image area, calculate the difference between its gray value and the mean gray value. Square these differences, sum all the squared differences, divide them by the total number of coordinate points, and finally take the square root to get the standard deviation of the gray value of the normal coordinate point σ er; First calculate the mean gray value of the abnormal coordinate point, for each coordinate point in the abnormal image area, calculate the difference between its gray value and the mean gray value, square these differences, sum all the squared differences, divide by the total number of coordinate points, and finally take the square root to get the gray value standard deviation σ of the abnormal coordinate point jd ; For all coordinate points in each normal image area, extract their grayscale values. Sum these grayscale values ​​and then divide them by the total number of coordinate points in the normal image area to obtain the grayscale mean μ of the normal coordinate points er ; For all coordinate points in each abnormal image area, extract their grayscale values; sum these grayscale values ​​and then divide them by the total number of coordinate points in the abnormal image area to obtain the grayscale value mean μ of the normal coordinate points jd .

[0201] Using programming languages ​​and related image processing libraries, such as Numpy and Scipy libraries in Python, obtain the normal-abnormal coordinate point pair frequency P(er, jd) from the gray-level co-occurrence matrix; er is a set of coordinate points in the area that is considered to be in a normal state, and the grayscale values ​​of these points usually have a certain characteristic range; jd is a set of coordinate points in the area that is judged to be in an abnormal state, and the grayscale values ​​of these points are significantly different from those of normal coordinate points, indicating that they are beyond the range of normal grayscale values.

[0202] According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the normal-abnormal coordinate point pair frequency P (er, jd), the contrast Contrast is calculated based on the following specific formula:

[0203]

[0204] Among them, (er, jd) represents each pair of grayscale values ​​in the gray-level co-occurrence matrix, the grayscale value er is [1,200], and the grayscale value jd is [201,255].

[0205] It should be noted that the variation of the grayscale value in the image is reflected by calculating the sum of the product of the square of the difference between different grayscale value pairs and the probability of occurrence. A larger contrast means that the grayscale value difference in the image is large and the texture may be rough; a smaller contrast means that the grayscale value difference is small and the texture is relatively smooth.

[0206] According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the gray value mean μ of the normal coordinate point er , mean gray value μ of abnormal coordinate points jd As well as the normal-abnormal coordinate point pair frequency P(er, jd), the grayscale correlation Cor is calculated based on the following specific formula:

[0207]

[0208] It should be noted that the linear correlation of gray values ​​in an image is measured by calculating the expectation of the product of the deviations of gray values ​​from their mean and then dividing it by the product of the standard deviations. If the correlation is high, it means that the change of gray values ​​in the image has a certain direction and regularity; if the correlation is low, it means that the change of gray values ​​is relatively random.

[0209] According to the frequency of normal-abnormal coordinate point pairs P(er, jd), the entropy is calculated based on the following specific formula:

[0210]

[0211] It should be noted that the complexity of the image is measured by calculating the negative value of the sum of the probability of occurrence of each gray value pair and the product of the logarithm of the probability. A larger entropy value indicates that the gray value distribution in the image is relatively uniform, with strong randomness and complex texture; a smaller entropy value indicates that the gray value distribution is relatively concentrated and the texture is relatively simple.

[0212] According to the normal-abnormal coordinate point frequency P (er, jd), the energy value Eu is calculated based on the following specific formula:

[0213]

[0214] Among them, the grayscale value er is [1,200], and the grayscale value jd is [201,255].

[0215] It should be noted that it reflects the uniformity and regularity of the image texture. A larger energy value indicates that the grayscale value distribution in the image is more concentrated and the texture is more uniform; a smaller energy value indicates that the grayscale value distribution is more dispersed and the texture is irregular.

[0216] The homogeneity Hu is calculated based on the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the frequency P (er, jd) of the normal-abnormal coordinate point pair. The specific formula is as follows:

[0217]

[0218] It should be noted that the similarity of adjacent grayscale values ​​in the image is reflected by calculating the sum of the probability of occurrence of each grayscale value pair and the inverse of their absolute difference. Higher homogeneity means that adjacent grayscale values ​​in the image are relatively similar and the texture is relatively uniform; lower homogeneity means that adjacent grayscale values ​​are relatively different and the texture is relatively rough.

[0219] According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the normal-abnormal coordinate point pair frequency P(er, jd), the moment of inertia IDu is calculated based on the following specific formula:

[0220]

[0221] It should be noted that the degree of change in the grayscale value in the image is measured by calculating the sum of the probability of each grayscale value pair and the inverse of the square of their difference. A larger moment of inertia indicates a larger difference in grayscale values ​​in the image and a high contrast; a smaller moment of inertia indicates a smaller difference in grayscale values ​​and a low contrast.

[0222] According to the texture structure difference data including contrast Contrast, grayscale correlation Cor, energy value Eu, homogeneity Hu, moment of inertia IDu and entropy Entropy, the comprehensive structure texture index STI is calculated. The specific formula is as follows:

[0223]

[0224] It should be noted that the principle is to comprehensively describe the texture characteristics of the image by comprehensively considering these six parameters. Each parameter reflects the texture characteristics of the image from a different perspective, such as contrast reflects the variation of grayscale values, correlation reflects the linear relationship of grayscale values, and entropy reflects the distribution uncertainty of grayscale values. Combining these parameters can obtain a more comprehensive and accurate description of texture characteristics, thus providing a stronger basis for judging the pathological condition of tissues.

[0225] This formula can quantify the texture features of an image from different angles by calculating various parameters of texture structure difference values, such as contrast, grayscale correlation, entropy, energy value, homogeneity and moment of inertia. Contrast reflects the amplitude of grayscale change and helps to identify the roughness of the texture; grayscale correlation reflects the linear relationship of grayscale values ​​and can determine the directionality of the texture; entropy measures the complexity of the image and can reflect the randomness of the texture; energy value indicates the uniformity of the texture; homogeneity and moment of inertia are described from the perspectives of local uniformity and contrast, respectively. The comprehensive structural texture index STI combines these parameters to provide a comprehensive description of texture features, which enables more accurate judgment of tissue lesions when analyzing medical ultrasound images, improves the accuracy and reliability of diagnosis, and provides doctors with quantitative texture information, which helps to better understand the structure and state of tissues and provide strong support for clinical diagnosis and treatment.

[0226] In the preferred embodiment of the intelligent processing method for medical ultrasound images, the color Doppler blood flow imaging data includes the transmitting frequency fb, the receiving frequency fs, the angle θ between the ultrasound beam and the blood flow direction, the peak blood flow velocity V ma and end-diastolic blood flow velocity V mi .

[0227] The blood flow velocity Cv is calculated based on the transmission frequency fb, the receiving frequency fs, and the angle θ between the ultrasound beam and the blood flow direction. The specific formula is as follows:

[0228]

[0229] Wherein, df represents the influence coefficient of color Doppler blood flow imaging data, and its value is [0.1, 10].

[0230] It should be noted that the influence coefficient df of color Doppler blood flow imaging data is obtained by collecting a large number of color Doppler blood flow imaging data samples, each of which contains the transmission frequency fb, the receiving frequency fs, the angle θ between the ultrasound beam and the blood flow direction, the peak blood flow velocity V during systole, and the peak blood flow velocity V during systole. ma and end-diastolic blood flow velocity V mi , use the regression algorithm in machine learning to build the model, such as linear regression, random forest regression, neural network and other models. ma and end-diastolic blood flow velocity V mi As input features, the accurate blood flow velocity value is used as the target output. The model is trained to learn the relationship between these features and blood flow velocity. In this process, the model will automatically adjust various parameters including the implicit df value to make the predicted blood flow velocity as close to the actual value as possible. After the ultrasound device transmits the ultrasonic wave, when the ultrasonic wave encounters scatterers such as red blood cells in the moving blood flow, part of the ultrasonic wave is reflected back. The device receives these reflected waves and measures their frequencies to determine the transmitting frequency fb and the receiving frequency fs respectively. During the ultrasound examination, the operator can observe the direction of the blood vessels and the direction of the ultrasound beam through the imaging interface of the ultrasound device. With experience and understanding of the anatomical structure, the operator can roughly estimate the angle θ between the ultrasound beam and the blood flow direction. When performing color Doppler blood flow imaging, the ultrasound device can measure the blood flow velocity at different times by analyzing and processing the blood flow signal. In one cardiac cycle of the heart, the device will automatically identify the maximum blood flow velocity during contraction, that is, the peak systolic blood flow velocity V ma ; and the blood flow velocity at the end of diastole, that is, the blood flow velocity at the end of diastole V mi These velocity values ​​are usually presented in graphical and numerical form on the display interface of the ultrasound device.

[0231] The principle of the formula is to use the Doppler effect to calculate the blood flow velocity Cv. When ultrasound waves irradiate moving blood cells, the receiving frequency will change due to the Doppler effect. The difference between the receiving frequency and the transmitting frequency (fb-fs) is related to the blood flow velocity. The blood flow velocity Cv can be calculated through this difference, the angle between the ultrasound beam and the blood flow direction, and related equipment parameters. The introduction of the angle θ is because the component of the blood flow velocity in the direction perpendicular to the ultrasound beam will cause the change in the receiving frequency. cosθ is used to convert the measured frequency change into the projection of the actual blood flow velocity in the direction of the ultrasound beam.

[0232] According to the peak systolic blood flow velocity V ma and end-diastolic blood flow velocity V mi , calculate the blood flow resistance index RII, the specific formula is as follows:

[0233]

[0234] When (fb-fs)>0, it means that the blood flows toward the probe; when (fb-fs)<0, it means that the blood flows away from the probe.

[0235] It should be noted that the principle is to measure blood flow resistance by comparing the difference in blood flow velocity between systole and diastole. The peak systolic blood flow velocity represents the fastest blood flow speed during heart contraction, and the end-diastolic blood flow velocity represents the slowest blood flow speed during heart diastole. The ratio of the difference between the two and the peak systolic blood flow velocity reflects the degree of blood vessel resistance to blood flow. The larger the ratio, the greater the vascular resistance; the smaller the ratio, the smaller the vascular resistance. When (fb-fs)>0, it means that the blood flow is toward the probe; when (fb-fs)<0, it means that the blood flow is away from the probe, which provides a basis for determining the direction of blood flow.

[0236] This formula is for calculating blood flow velocity. It uses the Doppler effect and multiple related parameters to accurately calculate the blood flow velocity. By considering factors such as the transmitting frequency, receiving frequency, and the angle between the ultrasound beam and the blood flow direction, it can more accurately reflect the actual blood flow situation. This helps doctors to fully understand the blood perfusion status of tissues and promptly discover possible vascular abnormalities or lesions. At the same time, this formula is a non-invasive test, which reduces harm to patients. The blood flow resistance index calculation formula measures vascular resistance by the difference between the peak systolic blood flow velocity and the end-diastolic blood flow velocity, which provides an important indicator for evaluating vascular function. It can help doctors judge the patency of blood vessels and the risk of lesions. In addition, combined with the judgment of blood flow direction, it can more comprehensively analyze blood flow characteristics, provide a strong basis for accurate diagnosis and formulation of treatment plans, and improve the clinical value and practicality of intelligent processing of medical ultrasound images.

[0237] In the preferred embodiment of the medical ultrasound image intelligent processing method, the specific steps of calculating the comprehensive blood flow assessment index MPI are as follows:

[0238] Patient history information data including past medical history index QS si , Patient age index BC Z and patient gender index BB Z .

[0239] According to blood flow velocity Cv, blood flow resistance index RII, blood flow direction TIF, medical history impact index QS si , Gender Impact Index BC Z And age impact index BB Z , calculate the comprehensive blood flow assessment index MPI, the specific formula is as follows:

[0240]

[0241] Among them, when the blood flow direction is toward the probe, TIF = 1, and when it is away from the probe, TIF = -1; d is the weight factor of blood flow velocity Cv, with a value of 0.1≤d≤0.3, g is the weight factor of blood flow resistance index RII, with a value of 0.2≤g≤0.4, p is the weight factor of blood flow direction TIF, with a value of 0.3≤p≤0.5, QS si is the influence index of the patient's si-th medical record. The value range of si is 1, 2, 3...sn. Sn is the total number of medical records in the medical history data Q. It is a positive integer. CX si is the weight coefficient of the si-th medical record, and its value is 0.1 <CX si <0.3, and d+g+p+CX si =1.

[0242] It should be noted that by consulting the patient's medical records, diagnosis reports and other medical documents, extracting information about the patient's past medical conditions, and quantifying them to obtain the medical history index QS si ; Directly obtain age data from the patient's personal information, divide and assign values ​​according to the degree of impact of the disease on different age groups, and then convert the age into the patient age index BC Z According to the patient's gender information, different gender-specific values ​​are assigned based on the differences in the incidence of certain diseases in different genders as the gender impact index BC Z, suppose a doctor is evaluating a patient's condition. By reviewing the patient's medical records and diagnostic reports, it is found that the patient had suffered from mild pneumonia in the past and recovered after treatment. According to the pre-set quantitative standards, this history of mild illness is assigned a certain value, such as a medical history index of 3, assuming a range of values ​​from 0 to 10, with higher values ​​indicating a greater potential impact of the medical history on the current condition. A 50-year-old patient comes to see a doctor. According to the degree of impact of different age groups on the disease, the age is divided into several intervals and assigned values. For example, the age group of 30-60 years old may be assigned a medium value, assuming that the age index corresponding to this age group is 5. This assignment is based on research showing that the risk of certain diseases is relatively high in this age group, or that changes in body function may have a certain impact on the development of a specific disease. If the patient is a male, for some diseases, such as prostate disease, the incidence rate of men is relatively high. Based on this difference in incidence, a specific value is assigned to male patients as the gender impact index. Assume that for the disease situation currently under consideration, the gender impact index of men is 4. If the patient is a female, for some specific female diseases, such as breast diseases, the incidence rate of women is higher, and different gender impact indexes are assigned accordingly. For example, female patients may be assigned a gender impact index of 6 for certain diseases related to female physiological characteristics.

[0243] The formula is a weighted sum of different parts such as blood flow velocity, blood flow resistance index, blood flow direction and patient history information, and takes into account the patient's gender impact index BC Z And age impact index BB Z , so that the role of each factor in the comprehensive evaluation can be balanced, so that the final MPI can fully and comprehensively reflect the patient's blood flow status and potential health risks. At the same time, it is required to ensure that the sum of each weight factor is 1, so that the comprehensive evaluation maintains a balance between different factors and avoids excessive dominance of a certain factor in the evaluation results.

[0244] This formula takes into account multiple factors such as blood flow velocity, blood flow resistance index, blood flow direction, and patient historical medical information, and can comprehensively evaluate the patient's blood flow condition. The setting of weight factors makes different factors have a reasonable degree of importance in the evaluation, avoiding the dominance of a single factor. The consideration of blood flow velocity reflects the dynamic situation of blood flow; the blood flow resistance index helps to judge the patency of blood vessels; the blood flow direction provides key information about the direction of blood flow. At the same time, combined with the patient's historical medical information, such as past medical conditions, age and gender, it can better adapt to individual differences and improve the accuracy and pertinence of the evaluation. Through this comprehensive evaluation, doctors can have a more comprehensive understanding of the patient's condition, provide a more reliable basis for diagnosis and formulation of treatment plans, help to detect potential health problems early, improve medical quality, reduce the risk of misdiagnosis, and provide strong support for patients' health management.

[0245] In the preferred embodiment of the medical ultrasound image intelligent processing method, the specific method for calculating the evaluation value CEV of the medical record to be tested is:

[0246] Obtain medical record evaluation value GCX through patient historical medical record information data uu , score each patient's historical medical record information and sum it up to get a comprehensive medical record evaluation value

[0247] Collect ultrasound image data of people of different ages, genders, and physical conditions to ensure that the reference value has wide applicability. For example, for liver ultrasound images, collect thousands of images of normal livers and patients with various liver diseases, measure the size of the lesion area, determine a reasonable reference range, and set the reference value μ for the lesion area size. S .

[0248] Calculate the comprehensive structural texture index of different medical records as the reference value μ for setting the comprehensive structural texture index STI 's reference.

[0249] Considering the characteristics of blood flow changes under different disease states and the differences in blood supply to different organs and tissues, the reference value μ of the comprehensive blood flow assessment index is set in combination with medical cases and medical journal materials. CFEI For example, in tumor tissue, blood flow is usually abundant, and its comprehensive blood flow evaluation index reference value may be different from that of normal tissue, so a corresponding reference value is set.

[0250] According to the known medical records, the size of the lesion area A, the comprehensive structural texture index STI, the comprehensive blood flow assessment index MPI, and the reference value of the lesion area size μ S , comprehensive structural texture index reference value μ STI And the comprehensive blood flow assessment index reference value μ CFEI, calculate the evaluation value CEV of the medical record to be measured, and the specific formula is as follows:

[0251]

[0252] Among them, β1 is the weight index of the lesion area size A, and the value of β1 is 0.1 < β1 ≤ 0.3; β2 is the weight index of the comprehensive structure texture index STI, and the value is 0.3 ≤ β2 ≤ 0.6; β3 is the weight index of the comprehensive blood flow evaluation index MPI, and the value is 0.2 ≤ β3 ≤ 0.5, and β1 + β2 + β3 = 1; GCX uu represents the evaluation value of the uu-th medical record, and the value range of uu is 1, 2, 3......jo, where jo is the total number of medical records and takes a positive integer.

[0253] It should be noted that this formula combines multiple aspects such as the lesion area size, the comprehensive structure texture index, the comprehensive blood flow evaluation index, and the comprehensive influence of multiple medical records. By standardizing each part and taking a specific exponential power, and then adding the comprehensive influence of multiple medical records, the evaluation value of the medical record to be measured is obtained. Such a design can comprehensively consider information from different aspects, comprehensively evaluate the condition of the medical record to be measured, and provide a more accurate diagnosis and treatment reference basis for doctors. At the same time, the requirement in the formula that β1 + β2 + β3 = 1 ensures that the sum of the three weight indexes is 1, so that the total weight of the lesion area size, the comprehensive structure texture index, and the comprehensive blood flow evaluation index in the evaluation is 1, that is, the sum of the relative importance is 100%. This can reasonably allocate weights between different parameters to adapt to different disease situations and evaluation needs. This formula comprehensively considers multiple key factors such as the lesion area size, the comprehensive structure texture index, and the comprehensive blood flow evaluation index, and can comprehensively evaluate the condition of the medical record to be measured. By adjusting the index, the weights of each factor can be flexibly adjusted according to different disease characteristics and clinical needs, improving the pertinence and accuracy of the evaluation. The consideration of the lesion area size helps to intuitively understand the degree of the lesion; the comprehensive structure texture index can provide more detailed information from the aspect of image texture features; the comprehensive blood flow evaluation index reflects the impact of the blood flow state on the condition. Coupled with the comprehensive influence term of multiple medical records, the experience of historical medical records can be used for reference to enhance the reliability of the evaluation. This multi-dimensional evaluation method provides a more comprehensive diagnostic basis for doctors, helps to detect diseases early, accurately judge the development of the condition, formulate more effective treatment plans, and improve the clinical value and practicality of medical ultrasonic image intelligent processing.

[0254] Set the disease evaluation threshold set Dv. The disease evaluation threshold set Dv includes a primary threshold Dv0 and a secondary threshold Dv1, where Dv0 < Dv1;

[0255] When the disease evaluation value CEV < Dv0, the judgment index is mild pathological characteristics, and a primary diagnosis report is generated;

[0256] When Dv0<symptom assessment value CEV≤Dv1, the indicator is judged as moderate pathological characteristics, and a secondary diagnosis report is generated;

[0257] When Dv1<symptom assessment value CEV, the indicator is judged as severe pathological characteristics and a third-level diagnosis report is generated.

[0258] It should be noted that medical experts determine the typical ranges of different diseases under specific indicators based on long-term clinical practice and observation and analysis of a large number of medical records, and thus set thresholds at various levels.

[0259] When the disease evaluation value CEV≤≤Dv0, there may be slight abnormalities, but these abnormalities are not sufficient to meet the diagnostic criteria of the disease.

[0260] The diagnosis report includes: Basic patient information: Clarify the patient's identity, including name, age, gender, medical record number, etc., to facilitate subsequent inquiries and comparisons. Ultrasound image description: Detailed description of normal tissue characteristics observed in ultrasound images, such as the shape, size, and regularity of the structure of organs. Analysis of symptom intermittent indicators: List specific values ​​of symptom intermittent indicators and compare them with the first-level threshold. Analyze factors that may cause slight fluctuations in indicators, such as recent lifestyle and changes in physiological status. Diagnostic conclusion: Clearly point out that the diagnostic criteria for the disease have not been met at present, but there are potential risks. Recommendations and further examinations: Propose time intervals for regular review to monitor changes in indicators. Depending on the specific situation, it may be recommended to adjust the lifestyle or take some preventive measures.

[0261] Suppose we are studying cardiovascular disease. By analyzing ultrasound image data and other relevant clinical data of a large number of patients with cardiovascular disease, we have determined the typical ranges of different cardiovascular diseases under specific indicators. For example, for myocardial lesions, specific indicators such as lesion area size, comprehensive structural texture index, and comprehensive blood flow assessment index have their corresponding typical ranges. After analysis and statistics, it is determined that the reference value of the lesion area size is within a certain range, assuming it is 1.5 square centimeters to 2.5 square centimeters; the reference value of the comprehensive structural texture index is set to 0.4 to 0.6; the reference value of the comprehensive blood flow assessment index is set to 0.3 to 0.5. According to these typical ranges, thresholds at various levels are further set. Suppose a patient undergoes a cardiovascular ultrasound examination, and the disease assessment value CEV is calculated to be 1.2. At the same time, the threshold Dv0 is set to 1.5. Since CEV1.2 is less than Dv01.5, this indicates that the patient may have a slight abnormality. For example, some subtle changes in the myocardial area may be observed in the ultrasound image, or the comprehensive structural texture index and the comprehensive blood flow assessment index may deviate slightly from the normal range, but these abnormalities are not sufficient to meet the clear diagnostic criteria for myocardial lesions. At this point, the doctor may recommend that the patient undergo further examinations, such as blood tests, electrocardiograms, etc., or ask the patient to undergo regular ultrasound examinations to closely monitor the development of the disease and prevent potential cardiovascular problems from further deteriorating.

[0262] When Dv0<symptom assessment value CEV≤Dv1, it usually indicates a lower risk or milder degree of the disease. When the symptom assessment value is lower than the first-level threshold, it may mean that the disease is relatively mild or in the early stages and requires further observation and monitoring.

[0263] The diagnosis report includes: basic information of the patient. Ultrasound image description: a detailed description of the abnormal features appearing in the ultrasound image. Analysis of the symptoms and time interval indicators: clearly list the values ​​of the symptoms and time interval indicators, and explain that they are between the primary and secondary thresholds. The possible reasons and potential risks of the analysis indicators being within this range. Diagnostic conclusion: a possible diagnosis of the disease is given, but the uncertainty is emphasized. Recommendations for further examination: other relevant imaging examinations or laboratory tests are recommended to clarify the diagnosis. Emphasize close observation of changes in the condition, and short-term follow-up can be performed if necessary.

[0264] Suppose we are studying thyroid disease. Through a large number of medical records, we have determined the relevant indicators and thresholds for thyroid disease assessment. Suppose Dv0 is 40 and Dv1 is 70. After a patient undergoes a thyroid ultrasound examination, the disease assessment value CEV is calculated to be 55 based on various indicators.

[0265] In the diagnostic report:

[0266] Patient basic information: The patient is a 35-year-old female with no special medical history.

[0267] Ultrasound image description: Ultrasound images show that some areas of the thyroid gland have uneven echoes and some tiny low-echo nodules.

[0268] Analysis of disease diagnosis indicators: The disease diagnosis indicator value is clearly listed as 55, and it is explained that it is between the first-level threshold of 40 and the second-level threshold of 70. The possible reasons for the analysis indicator being in this range may be recent stress, unbalanced iodine intake in the diet, etc. Potential risks include further enlargement of the nodules or calcification.

[0269] Diagnostic conclusion: The possible diagnosis is that the nature of the thyroid nodule is to be determined. Early changes of benign nodules are not ruled out, but the uncertainty is emphasized.

[0270] Further examinations are recommended: thyroid function tests and ultrasound reexamination after a short period of time (e.g., 3 months) are recommended to confirm the diagnosis. Close observation of changes in the condition and short-term follow-up if necessary are emphasized.

[0271] When Dv1<symptom assessment value CEV, it often represents a higher risk or more severe limit of the symptom. When the symptom assessment value is higher than the secondary threshold, it usually indicates that the condition is more serious and active treatment measures may need to be taken immediately.

[0272] The diagnostic report includes: basic information of the patient. Ultrasound image description: a comprehensive description of the serious abnormal manifestations in the ultrasound image. Analysis of the possible consequences and urgency of these abnormal manifestations. Analysis of symptom intermittent indicators: list the specific values ​​of the symptom intermittent indicators, and emphasize that they are far above the secondary threshold. Analyze the reasons why the indicators are in this highly abnormal range and the possible progression of the disease. Diagnostic conclusion: clearly give a more definite diagnosis of the disease and explain the severity of the disease. Recommendations and further examinations: It is recommended to conduct a comprehensive evaluation and formulate a treatment plan as soon as possible, which may include referral to a specialist, surgery or other treatment measures. Urgent laboratory tests or other auxiliary examinations are recommended to assess the overall condition of the body.

[0273] Still taking thyroid disease as an example, suppose another patient's CEV is 85.

[0274] In the diagnostic report:

[0275] Patient basic information: The patient is a 50-year-old male with a family history of thyroid disease.

[0276] Ultrasound image description: Ultrasound images show a large solid mass in the thyroid gland with unclear boundaries and abundant blood flow signals.

[0277] Analysis of disease diagnosis indicators: The disease diagnosis indicator value is clearly listed as 85, and it is explained that it is higher than the secondary threshold of 70. Analysis shows that due to the patient's older age and family history, the condition is more serious and there may be a risk of malignant tumors.

[0278] Diagnostic conclusion: The possible diagnosis given is that the thyroid malignancy is more likely, but further pathological examination is needed to confirm the diagnosis.

[0279] Further examination is recommended: It is recommended to immediately undergo pathological examination such as puncture biopsy to determine the nature of the mass. At the same time, active treatment measures such as surgery may need to be considered.

[0280] This formula can comprehensively evaluate the condition of the medical history to be tested and avoid the limitations of a single indicator by comprehensively considering multiple factors such as the size of the lesion area, texture index and blood flow assessment index. The setting of thresholds at each level provides clear standards for disease judgment, which helps doctors to quickly and accurately judge the severity of the disease. For cases below the first-level threshold, potential risks can be discovered in time and early intervention can be carried out; when it is between the first and second-level thresholds, it can remind doctors to conduct further examinations to clarify the diagnosis and avoid delaying the condition; when it is above the second-level threshold, emergency treatment measures can be quickly initiated. This hierarchical judgment improves the accuracy and efficiency of diagnosis and provides patients with more timely and personalized medical services. At the same time, it also provides a quantitative analysis method for medical research, which helps to deeply understand the development of the disease and promote the continuous advancement of medical ultrasound image intelligent processing technology and the expansion of clinical applications.

[0281] On the other hand, the present invention also discloses a medical ultrasonic image intelligent processing system, comprising:

[0282] The data acquisition module obtains normal ultrasound image data, which includes normal image samples of multiple normal tissues or organs, obtains normal data of two-dimensional grayscale images of normal image samples; obtains two-dimensional grayscale image test data and color Doppler blood flow imaging data of image samples to be processed; and obtains historical medical record information data of patients.

[0283] The preliminary diagnosis module is used to extract the normal initial edge coordinate points (x i ,y i ), according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu ); extract the initial edge coordinates (x) to be measured from the two-dimensional grayscale image to be measured a ,y a ), and calculate the coordinates of the reference edge to be measured (X a , Ya );Calculate the coordinates of the reference edge to be measured (X a , Y b ) and the normal characteristic coordinates (X uu , Y uu ) uu , make a preliminary judgment on the type of organ tissue.

[0284] The secondary diagnosis module, based on the coordinates of the reference edge to be tested (X a , Y a ) calculate the ellipticity E and circularity Eq of the organ tissue; obtain the ellipticity interval threshold and circularity interval threshold of different organ tissues, compare the ellipticity E and circularity Eq with the ellipticity interval threshold and circularity interval threshold of different organ tissues, respectively, make a secondary judgment on the type of the organ tissue according to the comparison result, and finally determine the organ tissue type of the image sample to be processed.

[0285] The data processing module is used to obtain the gray value of each coordinate point in the two-dimensional gray image to be tested; set the lesion gray threshold Ew, compare the gray value of each coordinate point with the lesion gray threshold Ew, and determine whether there is an abnormal gray value coordinate. All abnormal coordinate points constitute an abnormal area, and the boundary coordinates of the abnormal area form the initial edge coordinate point to be tested (x c ,y c ), according to the initial edge coordinate point 2 to be measured (x c ,y c ) Calculate the coordinates of the reference edge to be measured (X c , Y c ); Get the boundary coordinates of the hypoechoic nodule from the image sample to be processed (xc g ,yc g ), according to the coordinates of the reference edge to be measured (X c , Y c ) and the hypoechoic nodule boundary coordinates (xc g , yc g ), to determine whether there is a lesion in the organ tissue; when it is determined that there is a lesion, according to the coordinates of the reference edge to be measured (X c , Y c ) calculate the size A of the lesion area; obtain the grayscale co-occurrence matrix according to the two-dimensional grayscale image data to be tested, and calculate the texture structure difference data inside the tissue according to the grayscale co-occurrence matrix and the two-dimensional grayscale image data to be tested; calculate the comprehensive structure texture index STI according to the texture structure difference data; calculate the blood flow velocity Cv and the blood flow resistance index RII according to the color Doppler blood flow imaging data, and determine the blood flow direction TIF; calculate the comprehensive blood flow evaluation index MPI according to the blood flow velocity Cv, the blood flow resistance index RII, the blood flow direction TIF and the patient's historical medical record information data.

[0286] The summary report module is used to calculate the disease assessment value CEV based on the known medical record data, the size A of the lesion area, the comprehensive structural texture index STI and the comprehensive blood flow assessment index MPI, compare the disease assessment value CEV and the disease assessment threshold set Dv, combine the preliminary diagnosis results, judge the disease level, and automatically generate a diagnosis report.

[0287] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware or in combination with computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0288] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0289] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be covered by the protection scope of the present application.

Claims

1. A medical ultrasonic image intelligent processing method, characterized in that: include: Acquire normal ultrasound image data, which includes normal image samples of multiple normal tissues or organs, and acquire normal two-dimensional grayscale image data of normal image samples; acquire two-dimensional grayscale image test data and color Doppler blood flow imaging data of image samples to be processed; and acquire historical medical record information data of patients; Through the two-dimensional grayscale image normal data of different normal image samples, the normal initial edge coordinate points (x i ,y i ), according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , T uu ); The initial edge coordinates (x a ,y a ), and calculate the coordinates of the reference edge to be measured (X a , Y a );Calculate the coordinates of the reference edge to be measured (X a , Y b ) and the normal characteristic coordinates (X uu , Y uu ) uu , make a preliminary judgment on the type of organ tissue; According to the coordinates of the reference edge to be measured (X a , Y a ) Calculate the ellipticity E and circularity Eq of the organ tissue; obtain the ellipticity interval threshold and circularity interval threshold of different organ tissues, compare the ellipticity E and circularity Eq with the ellipticity interval threshold and circularity interval threshold of different organ tissues, perform secondary judgment on the type of the organ tissue according to the comparison result, and finally determine the organ tissue type of the image sample to be processed; Obtain the grayscale value of each coordinate point in the two-dimensional grayscale image to be tested; Set the grayscale threshold Ew of the lesion, compare the grayscale value of each coordinate point with the grayscale threshold Ew of the lesion, and determine whether there is an abnormal coordinate with grayscale value. All abnormal coordinate points constitute an abnormal area, and the boundary coordinates of the abnormal area form the initial edge coordinate point to be tested (x c ,y c ), according to the initial edge coordinate point 2 to be measured (x c ,y c ) Calculate the coordinates of the reference edge to be measured (X c , T c ); Get the boundary coordinates of the hypoechoic nodule from the image sample to be processed (xc g , yc g ), according to the coordinates of the reference edge to be measured (X c , Y c ) and the hypoechoic nodule boundary coordinates (xc g , yc g ) to determine whether there are lesions in organ tissues; When a lesion is determined, the boundary coordinates of the hypoechoic nodule (xc g , yc g ) Calculate the size A of the lesion area; A grayscale co-occurrence matrix is ​​obtained according to the two-dimensional grayscale image data to be tested, and texture structure difference data inside the tissue is calculated according to the grayscale co-occurrence matrix and the two-dimensional grayscale image data to be tested; and a comprehensive structure texture index STI is calculated according to the texture structure difference data; According to the color Doppler blood flow imaging data, the blood flow velocity Cv and the blood flow resistance index RII are calculated, and the blood flow direction TIF is determined. According to the blood flow velocity Cv, the blood flow resistance index RII, the blood flow direction TIF and the patient's historical medical information data, the comprehensive blood flow assessment index MPI is calculated; The disease assessment value CEV is calculated based on the known medical record data, the size A of the lesion area, the comprehensive structural texture index STI and the comprehensive blood flow assessment index MPI. The disease assessment value CEV and the disease assessment threshold set Dv are compared, combined with the preliminary diagnosis results, the disease level is judged, and a diagnosis report is automatically generated.

2. The medical ultrasonic image intelligent processing method according to claim 1, characterized in that: According to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu ) are as follows: According to the normal initial edge coordinate point (x i ,y i ) Calculate the normal mean coordinates (Xi mean , Yi mean ), based on the following formula: Among them, x i Indicates the x-axis value of the i-th edge coordinate point, y i Represents the y-axis value of the i-th edge coordinate point, the value range of i is 1, 2, 3...n, and n represents the total number of initial mark-off coordinate points; According to the normal initial edge coordinate point (x i ,y i ) and normal mean coordinates (Xi mean , Yi mean ), calculate the coordinate standard deviation (X m , Y m ), the specific formula is: According to the coordinate standard deviation (X m , Y m ) and normal mean coordinates (Xi mean , Yi mean ) Calculate the normal feature coordinates (X uu , Y uu ), based on the following specific formula: X uu =(X mean -kX m ,X mean +kX m ) Y uu =(Y mean -kY m ,Y mean +kY m )。 3. The medical ultrasonic image intelligent processing method according to claim 2, characterized in that: Calculate the coordinate data point of the reference edge (X a , Y a ), and the specific steps for preliminarily determining the organ tissue type are: Extract the initial edge coordinate point (x a ,y a ), and obtain the initial edge coordinate point to be measured (x a ,y a ) in the x-axis direction man , the minimum value x in the x-axis direction min , maximum value y in the y-axis direction man and the minimum value y in the y-axis direction min ; According to the coordinate data point of the benchmark edge to be measured (x a ,y a ), calculate the coordinate mean (X mean , Y mean ), the specific formula is: Among them, x a Indicates the value of the ath coordinate on the x-axis, y b It represents the value of the b-th coordinate on the y-axis. The value range of a is 1, 2, 3...x, and x represents the total number of coordinates. According to the maximum value x in the x-axis direction man , the minimum value x in the x-axis direction min , maximum value y in the y-axis direction man and the minimum value y in the y-axis direction min The specific formula for calculating the coordinate value normalization factor d is as follows: According to the initial edge coordinate point to be measured (x a ,y a ), coordinate mean (X mean , Y mean ) and the coordinate value normalization factor d, calculate the reference edge coordinates (X a , Y a ), based on the following specific formula: According to the reference edge coordinates (X a , Y a ) Calculate the difference S uu The specific formula is as follows: Set the difference standard value S w ; When the difference S uu <Difference standard value S w When the organ tissue to be tested is initially judged to be of this difference degree S uu Corresponding organ tissue type; when the difference S uu >Difference standard value S w When , it means that it is not this difference S uu Corresponding organ tissue type.

4. The medical ultrasonic image intelligent processing method according to claim 3, characterized in that: The steps for calculating the ellipticity E and the circularity Eq and further determining the organ tissue type are as follows: Based on the coordinates of the reference edge to be measured (X a , Y a ) to obtain the organ tissue area ES and the organ tissue area perimeter ESD, and according to the reference edge coordinates to be measured (X a , Y a ) performing ellipse fitting to obtain the length a of the major semi-axis and the length b of the minor semi-axis of the fitted ellipse; The ellipticity E is calculated based on the length of the major semi-axis a and the length of the minor semi-axis b, according to the following formula: Among them, the value range of the length b of the minor semi-axis of the fitted ellipse is b>0; The circularity Eq is calculated based on the organ tissue area ES and the organ tissue area perimeter ESD. The specific formula is as follows: Among them, the value range of the organ tissue area perimeter ESD is ESD>0; The ellipticity E and circularity Eq are compared with the ellipticity interval threshold and circularity interval threshold of different organs and tissues respectively; When the circularity Eq and the ellipticity E meet the ellipticity interval threshold and circularity interval threshold of the same organ tissue, it is judged to belong to this organ tissue type; When the results of the preliminary determination of the organ tissue type are consistent with the secondary determination of the organ tissue type, it is finally determined to belong to this organ tissue type. When the results of the preliminary determination of the organ tissue type are inconsistent with the secondary determination of the organ tissue type, the two-dimensional grayscale image data to be tested is re-acquired and the determination is repeated.

5. The medical ultrasonic image intelligent processing method according to claim 4, characterized in that: The method to determine whether there are abnormal coordinates of gray value is: When the gray value of the coordinate point in the two-dimensional gray image to be tested is greater than the gray value threshold Ew of the lesion, the coordinate is determined to be a gray value abnormal coordinate; The methods for determining whether there is a lesion are: For the reference edge coordinates to be measured (X c , Y c ) coordinate point (X j , Y j ) and the hypoechoic nodule boundary coordinates (xc g , yc g ) coordinate point (xc m , yc m ), when the coordinate point (X j , Y j ) and the coordinate point (xc m , yc m ) and the reference edge coordinate data 2 (X c , Y c ) intersect, and (xc m ,yC m ) is located inside the polygon formed by the reference edge coordinate data 2, it is judged that there is a lesion, and the hypoechoic nodule boundary coordinate (xc g , yc g ) are considered as the boundary coordinates of the lesion area; The method for calculating the lesion area size A is: According to the coordinates of the hypoechoic nodule boundary (xc g , yc g ), calculate the size of the lesion area A, the specific formula is as follows: Among them, xc g represents the g-th coordinate point on the x-axis of the hypoechoic nodule boundary, g represents the coordinate point number on the hypoechoic nodule boundary, and its value is [1, k], where k is a positive integer, xc g+1 represents the x-axis coordinate point g+1 on the boundary of the hypoechoic nodule, yc g represents the g-th coordinate point on the y-axis of the hypoechoic nodule boundary, yc g+1 Represents the g+1th coordinate point on the y-axis on the boundary of the hypoechoic nodule.

6. The medical ultrasonic image intelligent processing method according to claim 5, characterized in that: The specific steps for calculating the comprehensive structural texture index STI are: Texture structure difference data include contrast Contrast, grayscale correlation Cor, energy value Eu, homogeneity Hu, moment of inertia IDu and entropy Entropy; The two-dimensional grayscale image test data also includes the grayscale value er of the normal coordinate point, the grayscale value jd of the abnormal coordinate point, and the standard deviation σ of the grayscale value of the normal coordinate point er , the standard deviation of the gray value of the abnormal coordinate point σ jd 、The mean gray value μ of normal coordinate points er , mean gray value μ of abnormal coordinate points jd ; Obtain the normal-abnormal coordinate point pair frequency P(er, jd) according to the gray level co-occurrence matrix; According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the normal-abnormal coordinate point pair frequency P (er, jd), the contrast Contrast is calculated based on the following specific formula: According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the gray value mean μ of the normal coordinate point er , mean gray value μ of abnormal coordinate points jd As well as the normal-abnormal coordinate point pair frequency P(er, jd), the grayscale correlation Cor is calculated based on the following specific formula: According to the frequency of normal-abnormal coordinate point pairs P(er, jd), the entropy is calculated based on the following specific formula: According to the normal-abnormal coordinate point frequency P (er, jd), the energy value Eu is calculated based on the following specific formula: Among them, the gray value er is [1, 200], and the gray value jd is [201, 255]; The homogeneity Hu is calculated based on the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the frequency P (er, jd) of the normal-abnormal coordinate point pair. The specific formula is as follows: According to the gray value er of the normal coordinate point, the gray value jd of the abnormal coordinate point, and the normal-abnormal coordinate point pair frequency P(er, jd), the moment of inertia IDu is calculated based on the following specific formula: According to the texture structure difference data including contrast Contrast, grayscale correlation Cor, energy value Eu, homogeneity Hu, moment of inertia IDu and entropy Entropy, the comprehensive structure texture index STI is calculated. The specific formula is as follows:

7. The medical ultrasonic image intelligent processing method according to claim 6, characterized in that: Color Doppler blood flow imaging data include the transmission frequency fb, the receiving frequency fs, the angle θ between the ultrasound beam and the blood flow direction, the peak blood flow velocity V during systole, and the ma and end-diastolic blood flow velocity V mi ; The blood flow velocity Cv is calculated based on the transmission frequency fb, the receiving frequency fs, and the angle θ between the ultrasound beam and the blood flow direction. The specific formula is as follows: Wherein, df represents the influence coefficient of color Doppler blood flow imaging data, and its value is [0.1, 10]; According to the peak systolic blood flow velocity V ma and end-diastolic blood flow velocity V mi , calculate the blood flow resistance index RII, the specific formula is as follows:

8. The medical ultrasonic image intelligent processing method according to claim 7, characterized in that: The specific steps for calculating the comprehensive blood flow assessment index MPI are: Patient history information data including past medical history index QS si , Patient age index RC Z , patient gender index BB Z ; According to blood flow velocity Cv, blood flow resistance index RII, blood flow direction TIF, medical history impact index QS si , Gender Impact Index BC Z And age impact index BB Z The specific formula for calculating the comprehensive blood flow assessment index MPI is as follows: Among them, d is the weight factor of blood flow velocity Cv, which is 0.1≤d≤0.3, g is the weight factor of blood flow resistance index RII, which is 0.2≤g≤0.4, p is the weight factor of blood flow direction TIF, which is 0.3≤p≤0.5, QS si is the influence index of the si-th medical record of the patient. The value range of si is 1, 2, 3...sn. Sn is the total number of medical records of the medical history data Q. The value is a positive integer. CX si is the weight coefficient of the si-th medical record, and its value is 0.1<CX si <0.3, and d+g+p+CX si =1.

9. The medical ultrasonic image intelligent processing method according to claim 8, characterized in that: The specific calculation of the CEV of the medical record to be tested is as follows: Patient history information data also includes medical record evaluation value GCX uu ; Set the reference value μ for the lesion area size S , comprehensive structural texture index reference value μ STI And the comprehensive blood flow assessment index reference value μ CFEI ; According to the known medical records, the size of the lesion area A, the comprehensive structural texture index STI, the comprehensive blood flow assessment index MPI, and the reference value of the lesion area size μ S , comprehensive structural texture index reference value μ STI And the comprehensive blood flow assessment index reference value μ CFEI , calculate the CEV of the medical record to be tested, the specific formula is as follows: Among them, β1 is the weight index of the lesion area size A, and the value of β1 is 0.1<β1≤0.3; β2 is the weight index of the comprehensive structure texture index STI, and the value is 0.3≤β2≤0.6; β3 is the weight index of the comprehensive blood flow assessment index MPI, and the value is 0.2≤β3≤0.5, and β1+β2+β3=1; GCX uu represents the evaluation value of the uuth medical record, the value range of uu is 1, 2, 3...jo, jo is the total number of medical records, and its value is a positive integer; Setting a symptom assessment threshold set Dv, the symptom assessment threshold set Dv includes a primary threshold Dv0 and a secondary threshold Dv1, wherein Dv0<Dv1; When the symptom evaluation value CEV≤Dv0, the indicator is judged as a mild pathological feature and a first-level diagnosis report is generated; When Dv0<symptom assessment value CEV≤Dv1, the indicator is judged as moderate pathological characteristics, and a secondary diagnosis report is generated; When Dv1 is less than the symptom assessment value CEV, the indicator is judged to be a severe pathological feature, and a third-level diagnosis report is generated.

10. A medical ultrasonic image intelligent processing system, characterized in that: include: The data acquisition module acquires normal ultrasound image data, which includes normal image samples of multiple normal tissues or organs, and acquires normal data of two-dimensional grayscale images of normal image samples; acquires two-dimensional grayscale image test data and color Doppler blood flow imaging data of image samples to be processed; and acquires historical medical record information data of patients; The preliminary diagnosis module is used to extract the normal initial edge coordinate points (x i ,y i ), according to the normal initial edge coordinate point (x i ,y i ) Calculate the normal characteristic coordinates (X uu , Y uu ); extract the initial edge coordinates (x) to be measured from the two-dimensional grayscale image to be measured a ,y a ), and calculate the coordinates of the reference edge to be measured (X a , Y a );Calculate the coordinates of the reference edge to be measured (X a , Y b ) and the normal characteristic coordinates (X uu , Y uu ) uu , make a preliminary judgment on the type of organ tissue; The secondary diagnosis module, based on the coordinates of the reference edge to be tested (X a , Y a ) Calculate the ellipticity E and circularity Eq of the organ tissue; obtain the ellipticity interval threshold and circularity interval threshold of different organ tissues, compare the ellipticity E and circularity Eq with the ellipticity interval threshold and circularity interval threshold of different organ tissues, perform secondary judgment on the type of the organ tissue according to the comparison result, and finally determine the organ tissue type of the image sample to be processed; The data processing module is used to obtain the gray value of each coordinate point in the two-dimensional gray image to be tested; set the gray value threshold Ew of the lesion, compare the gray value of each coordinate point with the gray value threshold Ew of the lesion, and determine whether there is an abnormal coordinate of the gray value. All abnormal coordinate points constitute an abnormal area, and the boundary coordinates of the abnormal area form the initial edge coordinate point to be tested (x c ,y c ), according to the initial edge coordinate point 2 to be measured (x c ,y c ) Calculate the coordinates of the reference edge to be measured (X c , Y c ); Get the boundary coordinates of the hypoechoic nodule from the image sample to be processed (xc g , yc g ), according to the coordinates of the reference edge to be measured (X c , Y c ) and the hypoechoic nodule boundary coordinates (xc g , yc g ), to determine whether there is a lesion in the organ tissue; when it is determined that there is a lesion, according to the coordinates of the reference edge to be measured (X c , Y c ) calculate the size A of the lesion area; obtain a grayscale co-occurrence matrix according to the two-dimensional grayscale image to be tested data, calculate the texture structure difference data inside the tissue according to the grayscale co-occurrence matrix and the two-dimensional grayscale image to be tested data; calculate the comprehensive structure texture index STI according to the texture structure difference data; According to the color Doppler blood flow imaging data, the blood flow velocity Cv and the blood flow resistance index RII are calculated, and the blood flow direction TIF is determined. According to the blood flow velocity Cv, the blood flow resistance index RII, the blood flow direction TIF and the patient's historical medical information data, the comprehensive blood flow assessment index MPI is calculated; The summary report module is used to calculate the disease assessment value CEV based on the known medical record data, the size A of the lesion area, the comprehensive structural texture index STI and the comprehensive blood flow assessment index MPI, compare the disease assessment value CEV and the disease assessment threshold set Dv, combine the preliminary diagnosis results, judge the disease level, and automatically generate a diagnosis report.

Images