Computer equipment, product and medium
By quantitatively analyzing grayscale values on ultrasound images, the characteristic values of boundary clarity and placeholding effect strength are calculated, the problem of thyroid nodules relying on doctors' experience in diagnosis is solved, the accuracy and consistency of diagnosis is improved, and the Unnecessary medical intervention.
Patent Information
- Application Number
- CN202410959335.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-15
AI Technical Summary
The diagnosis of thyroid nodules is overly dependent on doctors' experience, and the lack of direct and quantitative objective indicators makes it difficult to guarantee the accuracy and consistency of diagnosis, and is prone to non-essential punctures and surgeries.
The ultrasound image is quantitatively analyzed by computer equipment, and the characteristic values of boundary clarity and placeholding effect strength are calculated, and provided to doctors for reference to improve the accuracy and objectivity of diagnosis.
It improves the accuracy and objectivity of thyroid nodules diagnosis, and reduces unnecessary puncture and surgery caused by subjective judgment differences.
Smart Images

Figure CN118967580A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of computer technology, and more specifically, to a computer device, product, and medium. Background Art
[0002] At present, the diagnostic technologies for thyroid nodules and thyroid cancer mainly include conventional ultrasound detection technology, ultrasound angiography technology and puncture technology.
[0003] Conventional ultrasound examination is currently the most sensitive examination method for thyroid nodules and is the preferred imaging examination method for diagnosing thyroid nodules and assessing the risk of benign and malignant nodules. Experienced ultrasound physicians' prediction of the diagnosis of thyroid lesions can improve the positive rate of thyroid puncture detection. Among them, diagnostic indicators include the shape and size of the nodules, echo, internal structure, whether the edges are regular, whether the boundaries are clear, whether there is protruding capsule growth, acoustic halo, blood flow, whether there is calcification, whether the aspect ratio is greater than 1, hardness, and whether there are abnormally enlarged lymph nodes in the neck. Among them, the risk of nodules being malignant is greater when there are solid, hypoechoic, irregular shape, unclear boundaries, lack of acoustic halo, rich blood flow or central blood supply, tiny clustered calcification, and aspect ratio greater than 1. However, in conventional ultrasound examination, the morphology, boundaries, internal echoes, and blood flow of thyroid nodules are sometimes difficult to identify, especially for thyroid nodules less than 1 cm in size, which leads to certain limitations of conventional ultrasound in distinguishing between benign and malignant nodules. For this purpose, ultrasound contrast imaging technology was proposed. Ultrasound contrast imaging is based on conventional ultrasound detection. Ultrasound contrast agents are injected through peripheral veins in a bolus manner. The contrast agents are used to increase the ultrasonic interface between tissues and blood vessels, improve the sound wave scattering energy (especially second harmonic imaging), and display the microcirculation perfusion status of tissues and organs. Since malignant tumors have unique microcirculation perfusion characteristics, ultrasound doctors can further distinguish the benign and malignant nature of tumors through the microcirculation perfusion status displayed by ultrasound contrast imaging. Thyroid puncture technology, especially fine needle aspiration biopsy, is often used for further diagnosis of thyroid nodules. This technology uses fine needle puncture of thyroid lesions to extract cell tissue for pathological examination to determine the benign and malignant nature of the nodules.
[0004] Among the above-mentioned thyroid detection technologies, although ultrasound images can provide rich image information, the judgment of the nature of nodules is still highly dependent on the subjective experience and visual judgment of ultrasound physicians, especially for tiny nodules, nodules with blurred boundaries, atypical features, and inflammatory nodules. The accuracy and consistency of diagnosis face challenges, and further puncture pathological examination is often required to identify the nature of the nodules, which may lead to unnecessary puncture and surgery for low-risk nodules due to differences in subjective judgment. Summary of the invention
[0005] The purpose of the present disclosure is to provide a computer device, product and medium to solve the technical problems in the related art that the diagnosis of thyroid nodules is overly dependent on the doctor's experience, lacks direct and quantitative objective indicators, the accuracy and consistency of the diagnosis are difficult to ensure, and unnecessary punctures and surgeries are prone to occur.
[0006] In order to achieve the above objectives, the present invention adopts the following technical solutions:
[0007] A first aspect of the present disclosure provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the following steps are implemented:
[0008] Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected;
[0009] Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected;
[0010] The first characteristic value and / or the second characteristic value are displayed.
[0011] Optionally, the step of calculating the first eigenvalue according to the ultrasonic image to be detected includes:
[0012] Acquire the grayscale value of the target feature point at the boundary line of the target to be detected in the ultrasonic image to be detected, wherein the target feature point includes a first feature point located inside the boundary line of the target to be detected and close to the boundary line, and a second feature point located outside the boundary line of the target to be detected and adjacent to the position of the first feature point;
[0013] The ratio of the grayscale value of the second feature point to the grayscale value of the first feature point is used as the first feature value, and the value of the first feature value is positively correlated with the boundary clarity of the target to be detected.
[0014] Optionally, the target feature point includes a plurality of feature point pairs, each feature point pair includes one first feature point and one second feature point, and for any feature point pair, a ratio of a grayscale value of the second feature point to a grayscale value of the first feature point is recorded as a first ratio; and the step of calculating the first feature value according to the ultrasonic image to be detected further includes:
[0015] For multiple feature point pairs, an average of the first ratios of the multiple feature point pairs is calculated, and the average is used as the first feature value.
[0016] Optionally, the step of calculating the second eigenvalue according to the ultrasonic image to be detected includes:
[0017] Acquire the gray value of each point in the first region where the target to be detected is located in the ultrasonic image to be detected;
[0018] The second eigenvalue is calculated based on the grayscale value of each point in the first area and a preset formula, and the preset formula is expressed as: O=(Lmax-L) / L, wherein O represents the second eigenvalue, Lmax represents the maximum grayscale value of the first area, and L represents the average grayscale value of each point in the first area.
[0019] Optionally, when the processor executes the program, the processor further implements the following steps:
[0020] A three-dimensional grayscale distribution graph is generated according to the grayscale values of each point in the first area, wherein any point in the three-dimensional grayscale distribution graph includes a first coordinate, a second coordinate and a third coordinate, wherein the first coordinate and the second coordinate represent the position of the point in the first area, and the third coordinate represents the grayscale value of the point.
[0021] Optionally, when the processor executes the program, the processor further implements the following steps:
[0022] The target type of the target to be detected is determined according to the first characteristic value and / or the second characteristic value.
[0023] Optionally, when the processor executes the program, the processor further implements the following steps:
[0024] Calculating, according to a plurality of sample ultrasound images, a first eigenvalue range and a second eigenvalue range corresponding to when the target to be detected is an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively; the first eigenvalue range is used to characterize the boundary clarity corresponding to when the target to be detected is an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively, and the second eigenvalue range is used to characterize the strength of the space-occupying effect corresponding to when the target to be detected is an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively;
[0025] The target type of the target to be detected is determined according to the first eigenvalue and the first eigenvalue range, and / or the target type of the target to be detected is determined according to the second eigenvalue and the second eigenvalue range.
[0026] Optionally, when the processor executes the program, the processor further implements the following steps:
[0027] Determining the target feature point in response to a first user operation, where the first user operation is an operation by a user to mark the target feature point in the ultrasonic image to be detected; and / or
[0028] The first area is determined in response to a second user operation, where the second user operation is an operation of a user marking a first area where the object to be detected is located in the ultrasonic image to be detected.
[0029] A second aspect of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:
[0030] Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected;
[0031] Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected;
[0032] The first characteristic value and / or the second characteristic value are displayed.
[0033] A third aspect of the present disclosure provides a computer program product, including a computer program, which, when executed by a processor, performs the following steps:
[0034] Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected;
[0035] Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected;
[0036] The first characteristic value and / or the second characteristic value are displayed.
[0037] The beneficial effects of the present disclosure are as follows:
[0038] The disclosed embodiment implements a computer device, which performs quantitative analysis on the grayscale value of the ultrasonic image to be detected, calculates a first eigenvalue characterizing the boundary clarity of the target to be detected and / or a second eigenvalue characterizing the strength of the space effect of the target to be detected, and displays the first eigenvalue and / or the second eigenvalue for reference by the doctor. By performing quantitative analysis on important indicators (boundary clarity and / or space effect) of the target to be detected in the thyroid ultrasonic image, on the one hand, the accuracy and objectivity of the diagnosis results of thyroid nodules are improved, and on the other hand, unnecessary punctures and surgeries for low-risk nodules due to differences in subjective judgment are avoided to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] The specific implementation methods of the present disclosure are further described in detail below with reference to the accompanying drawings.
[0040] Figure 1 A schematic diagram of a computer device provided for an embodiment of the present disclosure;
[0041] Figure 2 A flow chart of a detection method provided in an embodiment of the present disclosure;
[0042] Figure 3 A schematic diagram of an embodiment of an ultrasonic image to be detected provided by an embodiment of the present disclosure;
[0043] Figure 4 for Figure 3 A local magnified image of the target to be detected;
[0044] Figure 5 A schematic diagram of a three-dimensional grayscale distribution diagram when the target to be detected is a positive nodule provided in an embodiment of the present disclosure;
[0045] Figure 6 A schematic diagram of a three-dimensional grayscale distribution diagram when the target to be detected is a negative nodule provided in an embodiment of the present disclosure;
[0046] Figure 7 A schematic diagram of an embodiment of a thyroid ultrasound image provided in an embodiment of the present disclosure. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.
[0048] Unless otherwise defined, the technical terms or scientific terms used in the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, similar words such as "one", "one" or "the" do not indicate quantity restrictions, but indicate that there is at least one. Similar words such as "include" or "comprise" mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Similar words such as "connect" or "connected" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0049] In order to better understand the technical solutions of the computer equipment, products and media disclosed in the present invention, the design concept of the present invention is first briefly introduced.
[0050] The present disclosure is based on ultrasound images and improves the accuracy and objectivity of thyroid nodule diagnosis by quantitatively analyzing the grayscale value of ultrasound images. In ultrasound images, grayscale values are important parameters that reflect tissue characteristics in ultrasound images and can indirectly reflect the feedback intensity of tissues to ultrasound (i.e., echo intensity). Clinically, doctors judge the properties of various tissues, such as glands, blood vessels, soft tissues, stones, bone tissues, etc., based on the different grayscales in ultrasound images. The ultrasound image can be a black and white ultrasound image.
[0051] In the specific implementation, a specific image analysis software is designed. The input of the image analysis software is a black-and-white ultrasound image. The software can extract grayscale values and perform quantitative grayscale value analysis on the input ultrasound image. The characteristic values of certain features related to the nodule type of thyroid nodules are calculated through quantitative grayscale value analysis, and the characteristic values and reference values are used to give a diagnosis result of the nodule type of thyroid nodules, or the characteristic values are directly provided to doctors for reference, and the doctors use the quantified characteristic values to give a diagnosis result.
[0052] Please refer to Figure 1 to Figure 2 , Figure 1 A schematic diagram of a computer device provided in an embodiment of the present disclosure, Figure 2 Flow chart of the image analysis method provided by the embodiment of the present disclosure. Figure 1 to Figure 2 As shown, the computer device includes a memory 10, a processor 20, and a computer program stored in the memory 10 and executable on the processor 20. When the processor 20 executes the program, Figure 2 The steps of the image analysis method are shown.
[0053] Exemplarily, the computer device can be a computer, tablet computer or other device with display function and image processing function, which are not listed here in this disclosure. The beneficial effects of the computer device can be found in the detailed discussion in the above method embodiment, which will not be repeated here.
[0054] Specifically, Figure 2 As shown, the image analysis method comprises the following steps:
[0055] Step S101 : acquiring an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is a region with abnormal grayscale in the ultrasonic image to be detected.
[0056] In the disclosed embodiment, the ultrasound image to be detected is a black-and-white ultrasound image of the thyroid gland. Since different tissues have different grayscales in black-and-white ultrasound images, the grayscale value can characterize the nature of the tissue to a certain extent. Therefore, by observing the black-and-white ultrasound image, it can be determined whether there is suspicious tissue, and the suspicious tissue exhibits a grayscale different from that of the surrounding normal tissue (i.e., the grayscale abnormality described above). In the black-and-white ultrasound image of the thyroid gland, the suspicious tissue is usually referred to as a nodule. At this time, the target to be detected can be understood as the nodule to be detected. Since the grayscale of the nodule to be detected is different from that of the surrounding normal tissue, the grayscale difference will cause a boundary line to be formed between the nodule to be detected and the surrounding normal tissue in the ultrasound image. Usually, the nodule to be detected may be a thyroid nodule or a nodule-like lesion, and the nature of the nodule needs to be diagnosed. Please refer to Figure 3 , Figure 3 This is a schematic diagram of an embodiment of an ultrasonic image to be detected provided by an embodiment of the present disclosure. The execution subject of obtaining the ultrasonic image to be detected in this step is the aforementioned image analysis software, for example, the user can open or import the ultrasonic image to be detected in the image analysis software.
[0057] It is understandable that the image analysis method of the embodiment of the present disclosure can be applied to the detection of other structures in the medical field in addition to the thyroid gland detection. When the image analysis method of the embodiment of the present disclosure assists other diagnoses, the ultrasound images to be detected are ultrasound images of other structures, such as ultrasound images of the liver, gallbladder, pancreas, spleen, kidney, bladder, prostate, thyroid gland, breast, submandibular gland, parotid gland, etc. in the abdomen.
[0058] In the disclosed embodiment, the ultrasound image to be detected is a thyroid ultrasound image, and the target to be detected is a nodule to be detected. In the related art, doctors usually give corresponding nodule diagnosis results based on ultrasound images and combined with clinical experience. Optionally, the diagnosis result usually includes information such as the nodule type of the nodule, the risk level of the thyroid nodule, etc. Among them, the nodule type generally includes inflammatory nodules, benign nodules, high-risk nodules, etc. According to the 2020 China Ultrasound Thyroid Imaging Report and Data System Guidelines, the risk level of thyroid nodules can be divided into 6 levels, and the definitions of each level are: Level 1: No nodules; Level 2: Benign, malignant risk is 0%; Level 3: Benign may be large, and the risk of malignant nodules is less than 2%; Level 4: The risk of malignant nodules is 2%-90% (further diagnosis is required in combination with specific clinical conditions, such as puncture biopsy); Level 5: Highly suggestive of malignancy, with a malignant risk greater than 90%; Level 6: Indicates that the biopsy has been confirmed to be malignant.
[0059] The diagnosis results given by doctors based on clinical experience are usually highly subjective. For example, the risk of malignancy of grade 4 nodules and 4a-4c nodules is 2% to 90%, and they are easily confused with grade 3 nodules and inflammatory nodules. The diagnosis results are closely related to the clinical experience of doctors. Therefore, for the same ultrasound image, different doctors may give different results. On the one hand, it will lead to poor accuracy and consistency of the results. On the other hand, it may also lead to unnecessary puncture and surgery for low-risk nodules due to subjective judgment differences. For example: In clinical practice, some inflammatory nodules or benign nodules also have poor boundary clarity and irregular shapes. Some patients underwent surgery, and the postoperative pathology showed benign, which seriously affected their future thyroid function and quality of life. That is, through puncture pathological examination, although the pathological diagnosis results of inflammatory nodules or benign nodules can be obtained, the puncture biopsy is unnecessary at this time, and puncture complications may occur.
[0060] Step S102, calculating a first eigenvalue and / or a second eigenvalue based on the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, and is used to characterize the boundary clarity of the target to be detected; the second eigenvalue is a grayscale difference value between a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, and is used to characterize the strength of the space effect of the target to be detected.
[0061] In the diagnosis technology of the thyroid gland, the boundary clarity of the thyroid nodule is an important diagnostic indicator for whether the thyroid nodule is a high-risk nodule, and the space-occupying effect can distinguish high-risk nodules from benign nodules and inflammatory nodules. In the embodiment of the present disclosure, by performing quantitative grayscale value analysis on the ultrasound image to be detected, the quantitative data (eigenvalues) corresponding to these two indicators are calculated, thereby achieving deep refinement and precise processing of high-risk diagnostic indicators, which can improve the sensitivity and accuracy of diagnosis. It can be understood that the method of the embodiment of the present disclosure is more practical for thyroid nodules with a risk level greater than or equal to level 4, for which it is impossible to determine whether the nodule type is inflammatory or benign based on the ultrasound image, and can improve the diagnostic accuracy of thyroid nodules.
[0062] Specifically, the first eigenvalue is the grayscale difference value of the target feature point inside the boundary line of the target to be detected and outside the boundary line, so the first eigenvalue can represent the boundary clarity. Specifically, taking the boundary line as the dividing line, if the difference between the grayscale value of the point inside the boundary line and the grayscale value of the point outside the boundary line is small, the boundary clarity is low, and if the difference is large, the boundary clarity is high.
[0063] The second eigenvalue is the grayscale difference between the maximum grayscale value and the average grayscale value of the first area where the target to be detected is located. The maximum grayscale value represents the maximum grayscale value among the grayscale values of each point in the first area, and the average grayscale value is the average of the grayscale values of each point in the first area. At this time, the second eigenvalue can represent the maximum grayscale value difference intensity in the area where the target to be detected is located, based on the average grayscale value, which can characterize the strength of the space-occupying effect of the thyroid nodule. Specifically, the smaller the second eigenvalue, the weaker the space-occupying effect, and the larger the second eigenvalue, the stronger the space-occupying effect.
[0064] In specific implementation, only the boundary clarity of the target to be detected can be quantitatively analyzed, or only the space occupancy effect of the target to be detected can be quantitatively analyzed. Optionally, in order to improve the accuracy of diagnosis of the target to be detected, both the boundary clarity and the space occupancy effect of the target to be detected can be quantitatively analyzed.
[0065] Step S103: display the first eigenvalue and / or the second eigenvalue.
[0066] In the disclosed embodiment, after the image analysis software calculates the first eigenvalue and / or the second eigenvalue, it will also display it on the display interface so that the user can intuitively obtain quantitative data.
[0067] Compared with the related art, the computer device of the embodiment of the present invention performs quantitative analysis on the grayscale value of the ultrasonic image to be detected, calculates the first eigenvalue characterizing the boundary clarity of the target to be detected and / or the second eigenvalue characterizing the strength of the space effect of the target to be detected, and displays the first eigenvalue and / or the second eigenvalue for the doctor's reference. By performing quantitative analysis on the important indicators (boundary clarity and / or space effect) of the target to be detected in the thyroid ultrasonic image, on the one hand, the accuracy and objectivity of the diagnosis results of thyroid nodules are improved, and on the other hand, unnecessary puncture and surgery caused by subjective judgment differences for low-risk nodules are avoided to a certain extent.
[0068] In the embodiment of the present disclosure, different tissues in the ultrasonic image to be detected have different grayscales, and the grayscale value can characterize the properties of the tissue to a certain extent. Figure 7 , Figure 7 is a schematic diagram of an embodiment of a thyroid ultrasound image, Figure 7 The middle line segment Y is an arbitrary line segment set for the convenience of understanding. Figure 7 The ordinate of the green rendering represents the grayscale value of each point on line segment Y, and the abscissa represents the abscissa of each point on line segment Y, where the grayscale value ranges from 0 to 255. Figure 7 The areas Y1 and Y3 marked by the dashed box have higher grayscale values, which clinically correspond to the grayscale of the glands. The area Y2 between Y1 and Y3 has a lower grayscale value, which clinically corresponds to the grayscale of the nodules. Figure 7 It can be seen that the critical positions of glands and nodules have large grayscale differences. Based on the above principle, the boundary clarity of the target to be detected can be determined by quantitatively analyzing the grayscale values of each point inside and outside the boundary line.
[0069] In a possible implementation, the first characteristic value is a grayscale difference value between target characteristic points inside and outside the boundary line of the target to be detected. The grayscale difference value can be expressed in the form of a ratio or difference of grayscale values of target characteristic points inside and outside the boundary line.
[0070] The target feature points include a first feature point located inside the boundary line of the target to be detected and close to the boundary line, and a second feature point located outside the boundary line of the target to be detected and adjacent to the first feature point. Assume that the grayscale value of the first feature point is recorded as the first grayscale value G1, and the grayscale value of the second feature point is recorded as the second grayscale value G2.
[0071] Exemplarily, the first eigenvalue K1 can be expressed as the difference between the second grayscale value G2 and the first grayscale value G1, such as K1 = G2-G1. The larger the absolute value of K1, the greater the grayscale difference on both sides of the boundary line, and the clearer the boundary line. Similarly, the smaller the absolute value of K1, the smaller the grayscale difference on both sides of the boundary line, and the lower the clarity of the boundary line.
[0072] Exemplarily, the first eigenvalue K1 may also be expressed as K1=(G2-G1) / G1.
[0073] Exemplarily, the ratio of the gray value G2 of the second feature point to the gray value G1 of the first feature point is taken as the first eigenvalue K1. If the first eigenvalue K1 is expressed as K1=G2 / G1, then the first eigenvalue K1 can be understood as the boundary slope of the boundary line. The larger the absolute value of the difference between the first eigenvalue K1 and 1 (the absolute value of (K1-1)), the higher the boundary clarity is, and the smaller the absolute value of the difference between the first eigenvalue K1 and 1 is, the lower the boundary clarity is.
[0074] Optionally, the step of calculating the first eigenvalue according to the ultrasonic image to be detected includes:
[0075] (1) obtaining grayscale values of target feature points at a boundary line of the target to be detected in the ultrasonic image to be detected, wherein the target feature points include a first feature point located inside and close to the boundary line of the target to be detected, and a second feature point located outside and adjacent to the first feature point;
[0076] (2) The ratio of the grayscale value of the second feature point to the grayscale value of the first feature point is used as the first feature value.
[0077] In a possible implementation, the target feature points include multiple feature point pairs, each feature point pair includes one first feature point and one second feature point, and for any feature point pair, the ratio of the grayscale value of the second feature point to the grayscale value of the first feature point is recorded as a first ratio; the step of calculating the first eigenvalue based on the ultrasonic image to be detected also includes: for multiple feature point pairs, calculating the average of the first ratios of the multiple feature point pairs, and taking the average as the first eigenvalue.
[0078] For example, please refer to Figure 3 to Figure 4 , Figure 4 for Figure 3 A local enlarged image of the target area to be detected, such as Figure 3 to Figure 4As shown, the target feature points include 5 feature point pairs, each of which is represented by a red dot, which are respectively recorded as 1 and 2, 3 and 4, 5 and 6, 7 and 8, 9 and 10, where 2, 4, 6, 8, 10 represent the 5 first feature points located inside and close to the boundary line, and 1, 3, 5, 7, 9 represent the 5 second feature points located outside the boundary line and adjacent to the first feature points 2, 4, 6, 8, 10, respectively. At this time, the first eigenvalue is the mean of the first ratios of each feature point pair. Assuming that the grayscale values of the target feature points 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 are recorded as L1, L2, L3, L4, L5, L6, L7, L8, L9, L10, respectively, then the first eigenvalue K1 is expressed as K1 = (L1 / L2+L3 / L4+L5 / L6+L7 / L8+L9 / L10) / 5. The doctor can then use the first characteristic value K1 to make a diagnosis.
[0079] In the embodiment of the present disclosure, the target feature points are determined by the user (doctor) based on the ultrasound image to be detected. In specific implementation, the ultrasound image to be detected can be opened in the image analysis software, and the image analysis software provides a user annotation function, and the user can annotate the target feature points in the ultrasound image to be detected, such as Figure 4 When the user marks the target feature points, he first observes the ultrasonic image to be detected, and manually determines the area where the target to be detected is located and the boundary line of the target to be detected. Then, around the boundary line, at the position where the boundary line is relatively clear, multiple feature point pairs are selected, where the first feature point is located inside the boundary line and closest to the boundary line, and the second feature point is located outside the boundary line and closest to the boundary line. It should be noted that Figure 3 and Figure 4 The solid circle in the middle represents the first area, and the boundary line in Figure 3 and Figure 4 Usually, when the boundary of the target to be detected is unclear, the qualitative determination of benign / high risk requires more subjective judgment, so Figure 3 and Figure 4 The boundary line of the target to be detected may be unclear. In this case, the position of the boundary line can be clearly determined in the target to be detected, and the target feature point is selected.
[0080] After the user marks the target feature points, the image analysis software will perform image processing on the ultrasonic image to be detected and extract the grayscale values corresponding to the target feature points. Taking the target feature points shown above as an example, which include 5 feature point pairs, the image analysis software can generate a grayscale value list after image processing on the ultrasonic image to be detected, as shown in Table 1. Subsequent quantitative analysis is performed based on the data in Table 1.
[0081] Table 1
[0082] The first feature point Gray value The second feature point Gray value X2 L2 X1 L1 X4 L4 X3 L3 X6 L6 X5 L5 X8 L8 X7 L7 X10 L10 X9 L9
[0083] Correspondingly, the image analysis method disclosed herein further includes the following steps: determining the target feature point in response to a first user operation, wherein the first user operation is an operation of a user marking the target feature point in the ultrasound image to be detected.
[0084] In a possible implementation, the second eigenvalue is the grayscale difference between the maximum grayscale value and the average grayscale value of the first area where the target to be detected is located. The grayscale difference can be expressed by the ratio or difference between the maximum grayscale value and the average grayscale value or other forms. Assume that the second eigenvalue is denoted as O, the maximum grayscale value is expressed as Lmax, and the average grayscale value is expressed as L.
[0085] Exemplarily, the second eigenvalue O is expressed as O=Lmax-L. The larger the second eigenvalue is, the greater the maximum grayscale value difference intensity is within the first region based on the average grayscale value, that is, the greater the difference between the maximum grayscale value and the average grayscale value is.
[0086] Exemplarily, (Lmax-L) may be further normalized, such as the second eigenvalue O is expressed as O=(Lmax-L) / L.
[0087] Correspondingly, the step of calculating the second eigenvalue based on the ultrasonic image to be detected includes: (1) obtaining the grayscale value of each point in the first area where the target to be detected is located in the ultrasonic image to be detected; (2) calculating the second eigenvalue based on the grayscale value of each point in the first area and a preset formula, and the preset formula is expressed as: O=(Lmax-L) / L, wherein O represents the second eigenvalue, Lmax represents the maximum grayscale value of the first area, and L represents the average grayscale value of each point in the first area.
[0088] In the embodiment of the present disclosure, the first area is determined by the user based on the ultrasound image to be detected. In specific implementation, the ultrasound image to be detected can be opened in the image analysis software, and the image analysis software provides a user annotation function, and the user can annotate the first area in the ultrasound image to be detected, such as Figure 3 When the user marks the first area, he first observes the ultrasound image to be detected, preliminarily manually determines the area where the target to be detected is located, and then accurately delineates the area (i.e., it is relatively clear that it is a nodule area, so the first area is usually within the manually determined boundary line), and then the image analysis software can identify the first area delineated by the user. Correspondingly, the image analysis method disclosed in the present invention also includes the following steps: determining the first area (such as Figure 3 The first area is represented by reference numeral 11 in the figure), and the second user operation is an operation of the user marking the first area where the target to be detected is located in the ultrasonic image to be detected, that is, an operation of the user circling the first area.
[0089] In a possible implementation method, the image analysis method provided in the embodiment of the present disclosure further includes the following steps:
[0090] Step S104, generating a three-dimensional grayscale distribution graph according to the grayscale values of each point in the first area, wherein any point in the three-dimensional grayscale distribution graph includes a first coordinate, a second coordinate and a third coordinate, wherein the first coordinate and the second coordinate represent the position of the point in the first area, and the third coordinate represents the grayscale value of the point.
[0091] For example, please refer to Figures 5 and 6 , Figure 5 A schematic diagram of a three-dimensional grayscale distribution diagram when the target to be detected is a positive nodule (i.e., a malignant nodule determined by biopsy) provided in an embodiment of the present disclosure, Figure 6 A schematic diagram of a three-dimensional grayscale distribution diagram when the target to be detected is a negative nodule (benign nodule) provided in an embodiment of the present disclosure, wherein the spatial coordinate system of the three-dimensional grayscale distribution diagram includes an x-axis, a y-axis and a z-axis, wherein the x-axis and the y-axis define a two-dimensional plane, and the positive projection of the three-dimensional grayscale distribution diagram on the two-dimensional plane is a first area, and the z-axis represents the grayscale value. For any point in the three-dimensional grayscale distribution diagram, the first coordinate, the second coordinate and the third coordinate included therein may be an x-coordinate, a y-coordinate and a z-coordinate, respectively, and the z-coordinate is the grayscale value of the point in the ultrasonic image to be detected. Figures 5 and 6 The area marked by the dashed box in the middle represents the first area. Figures 5 and 6 It can be seen that positive nodules and negative nodules have significantly different three-dimensional images. The grayscale distribution three-dimensional image can more intuitively, quantitatively and objectively display the space-occupying effect of the target to be detected.
[0092] In the disclosed embodiment, when performing quantitative analysis on the space-occupying effect of the target to be detected, in addition to calculating the second eigenvalue, a three-dimensional image is generated based on the grayscale values of each point in the first area. The grayscale values of each point in the first area where the target to be detected is located are visualized through the three-dimensional image. The three-dimensional grayscale distribution graph can serve as a supplement to the second eigenvalue, indirectly describing the strength of the space-occupying effect, so that doctors can more intuitively observe the space-occupying effect of the target to be detected.
[0093] In a possible implementation, after the step of calculating the first eigenvalue and / or the second eigenvalue according to the ultrasonic image to be detected, the step further includes: determining a diagnosis result of the target to be detected according to the first eigenvalue and / or the second eigenvalue.
[0094] Exemplarily, the diagnosis result of the target to be detected includes the nodule type of the target to be detected, and the nodule type includes inflammatory nodules, benign nodules and high-risk nodules, wherein benign nodules or high-risk nodules refer to benign thyroid nodules or high-risk thyroid nodules, respectively.
[0095] Correspondingly, determining the diagnostic result of the target to be detected according to the first characteristic value and / or the second characteristic value is specifically: determining the target type of the target to be detected according to the first characteristic value and / or the second characteristic value. Exemplarily, when the target to be detected is a nodule to be detected, determining the target type of the target to be detected according to the first characteristic value and / or the second characteristic value is specifically: determining the nodule type of the target to be detected according to the first characteristic value and / or the second characteristic value.
[0096] In a specific implementation, the boundary clarity and space occupancy effect of multiple sample ultrasound images can be quantitatively analyzed in advance to obtain the first eigenvalue range and the second eigenvalue range corresponding to different nodule types of the target to be detected, and then the first eigenvalue and the second eigenvalue of the ultrasound image to be detected are compared with the first eigenvalue range and the second eigenvalue range respectively, and then the nodule type of the target to be detected is determined according to the comparison result. The diagnosis result has strong objectivity and accuracy.
[0097] Optionally, a doctor may manually compare the first eigenvalue and the second eigenvalue of the ultrasonic image to be detected with the first eigenvalue range and the second eigenvalue range respectively, and determine the target type of the target to be detected according to the comparison result.
[0098] Optionally, the steps of comparing the first eigenvalue and the second eigenvalue of the ultrasonic image to be detected with the first eigenvalue range and the second eigenvalue range, respectively, and determining the target type of the target to be detected according to the comparison result can also be performed by a computer device. For example, a comparison algorithm can be added to the image analysis software, and the comparison algorithm has the first eigenvalue range and the second eigenvalue range built in. After the image analysis software performs quantitative analysis and calculation on the current ultrasonic image to be detected to obtain the first eigenvalue and the second eigenvalue, it automatically compares and determines the target type of the target to be detected according to the comparison result. During output, the image analysis software can display the first eigenvalue, the second eigenvalue, the first eigenvalue range, the second eigenvalue range, and the determined target type of the target to be detected on the display interface, and this information is provided for the doctor's reference.
[0099] Optionally, the image analysis method provided in the embodiment of the present disclosure further includes the following steps:
[0100] (1) Calculating the first eigenvalue range and the second eigenvalue range corresponding to the target to be detected being an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively, based on multiple sample ultrasound images. The first eigenvalue range is used to characterize the boundary clarity corresponding to the target to be detected being an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively, and the second eigenvalue range is used to characterize the strength of the space-occupying effect corresponding to the target to be detected being an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively.
[0101] Specifically, the multiple sample ultrasound images need to include multiple sample ultrasound images corresponding to inflammatory nodules, multiple sample ultrasound images corresponding to benign nodules, and multiple sample ultrasound images corresponding to high-risk nodules. Exemplarily, there are 200 sample ultrasound images corresponding to each nodule type. By performing quantitative analysis on the sample ultrasound images under each nodule type, the first eigenvalue range and the second eigenvalue range corresponding to the nodule type can be obtained.
[0102] (2) determining the target type of the target to be detected according to the first eigenvalue and the first eigenvalue range and / or determining the target type of the target to be detected according to the second eigenvalue and the second eigenvalue range.
[0103] Exemplarily, when diagnosis is performed based on two quantitative indicators, the first eigenvalue and the second eigenvalue, the diagnostic process can be: if the first eigenvalue is within the first eigenvalue range where the inflammatory nodule is located and the second eigenvalue is within the second eigenvalue range where the inflammatory nodule is located, it can be preliminarily determined that the target type of the target to be detected is an inflammatory nodule.
[0104] Similarly, if the first eigenvalue is within the first eigenvalue range of benign nodules and the second eigenvalue is within the second eigenvalue range of benign nodules, it can be preliminarily determined that the target type of the target to be detected is a benign nodule. If the first eigenvalue is within the first eigenvalue range of high-risk nodules and the second eigenvalue is within the second eigenvalue range of high-risk nodules, it can be preliminarily determined that the target type of the target to be detected is a high-risk nodule, and further detection can be performed in combination with puncture detection.
[0105] It is understandable that the embodiments of the present disclosure only disclose the technical solution of diagnosis through two quantitative indicators, namely, boundary clarity and space-occupying effect. In actual scenarios, when doctors give the diagnosis results of the target to be detected, they not only need to use the two indicators of boundary clarity and space-occupying effect, but also need to confirm it in combination with other information in the ultrasound image, such as the shape and size of the target to be detected, echo, internal structure, whether there is protruding capsule growth, acoustic halo, blood flow, whether there is calcification, whether the aspect ratio is greater than 1, hardness, and whether there are abnormally swollen lymph nodes in the neck. Compared with the diagnosis based on doctor's experience in the related art, the present disclosure quantitatively analyzes some important key indicators (boundary clarity and space-occupying effect), and uses a large amount of sample data to derive the reference range of indicators when the target to be detected is of different types, so that the diagnosis result is more objective and more accurate.
[0106] After preliminary verification, the image analysis method of the disclosed embodiment was used to compare the ultrasonic image quantitative analysis of the targets to be detected with unclear positive and unclear negative boundaries, P < 0.001, with significant differences and an accuracy rate of more than 60%. The ultrasonic image quantitative analysis method in the disclosed embodiment can significantly improve the sensitivity and accuracy of the diagnosis of risk nodules (grade 4 and above), while reducing unnecessary medical interventions, which is of great significance for clinical diagnosis and treatment.
[0107] Among them, P is a commonly used indicator in statistics, namely P value (Probability value or Probability of error), which is used to indicate the probability of observing sample data or more extreme situations under the premise that the null hypothesis is true. P < 0.001 means that the observed data difference is highly significant in statistics, and the possibility that the difference is caused by accidental factors (such as sampling error) can almost be ruled out. There is a real difference between the two groups or phenomena under study.
[0108] Based on the same inventive concept, the second aspect of the present disclosure provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the image analysis method described above are implemented.
[0109] Exemplarily, when the program is executed by a processor, the following steps are implemented:
[0110] Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected;
[0111] Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected;
[0112] The first characteristic value and / or the second characteristic value are displayed.
[0113] In the specific implementation process, the computer storage medium may include: Universal Serial Bus Flash Drive (USB), mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and other storage media that can store program codes.
[0114] Based on the same inventive concept, the third aspect of the present disclosure provides a computer program product, including a computer program, which implements the steps of the above-mentioned image analysis method when executed by a processor. Since the principle of solving the problem by the above-mentioned computer program is similar to the principle of the image analysis method, the implementation of the above-mentioned computer program can refer to the implementation of the image analysis method, and the repeated parts will not be repeated.
[0115] Exemplarily, when the computer program is executed by a processor, the following steps are implemented:
[0116] Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected;
[0117] Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected;
[0118] The first characteristic value and / or the second characteristic value are displayed.
[0119] The computer program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more conductors, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.
[0120] Obviously, the above embodiments of the present disclosure are merely examples for clearly illustrating the present disclosure, and are not limitations on the implementation methods of the present disclosure. For ordinary technicians in this field, other different forms of changes or modifications can be made on the basis of the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present disclosure are still within the protection scope of the present disclosure.
Claims
1. A computer device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the following steps are implemented: Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected; Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected; The first characteristic value and / or the second characteristic value is displayed.
2. The computer device according to claim 1, characterized in that The step of calculating the first eigenvalue according to the ultrasonic image to be detected comprises: Acquire the grayscale value of the target feature point at the boundary line of the target to be detected in the ultrasonic image to be detected, wherein the target feature point includes a first feature point located inside the boundary line of the target to be detected and close to the boundary line, and a second feature point located outside the boundary line of the target to be detected and adjacent to the first feature point; The ratio of the grayscale value of the second feature point to the grayscale value of the first feature point is used as the first feature value, and the value of the first feature value is positively correlated with the boundary clarity of the target to be detected.
3. The computer device according to claim 2, characterized in that The target feature points include a plurality of feature point pairs, each feature point pair includes one of the first feature points and one of the second feature points, and for any feature point pair, a ratio of a grayscale value of the second feature point to a grayscale value of the first feature point is recorded as a first ratio; The step of calculating the first eigenvalue according to the ultrasonic image to be detected also includes: For multiple feature point pairs, an average of the first ratios of the multiple feature point pairs is calculated, and the average is used as the first feature value.
4. The computer device according to claim 1, characterized in that The step of calculating the second eigenvalue according to the ultrasonic image to be detected comprises: Acquire the gray value of each point in the first region where the target to be detected is located in the ultrasonic image to be detected; The second eigenvalue is calculated based on the grayscale value of each point in the first area and a preset formula, and the preset formula is expressed as: O=(Lmax-L) / L, wherein O represents the second eigenvalue, Lmax represents the maximum grayscale value of the first area, and L represents the average grayscale value of each point in the first area.
5. The computer device according to claim 4, characterized in that When the processor executes the program, the following steps are also implemented: A three-dimensional grayscale distribution graph is generated according to the grayscale values of each point in the first area, wherein any point in the three-dimensional grayscale distribution graph includes a first coordinate, a second coordinate and a third coordinate, wherein the first coordinate and the second coordinate represent the position of the point in the first area, and the third coordinate represents the grayscale value of the point.
6. The computer device according to claim 1, characterized in that When the processor executes the program, the following steps are also implemented: The target type of the target to be detected is determined according to the first characteristic value and / or the second characteristic value.
7. The computer device according to claim 1, characterized in that When the processor executes the program, the following steps are also implemented: Calculating, according to a plurality of sample ultrasound images, a first eigenvalue range and a second eigenvalue range corresponding to when the target to be detected is an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively; the first eigenvalue range is used to characterize the boundary clarity corresponding to when the target to be detected is an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively, and the second eigenvalue range is used to characterize the strength of the space-occupying effect corresponding to when the target to be detected is an inflammatory nodule, a benign nodule, and a high-risk nodule, respectively; The target of the target to be detected is determined according to the first eigenvalue and the first eigenvalue range, and / or the target type of the target to be detected is determined according to the second eigenvalue and the second eigenvalue range.
8. The computer device according to claim 1, characterized in that When the processor executes the program, the following steps are also implemented: Determining the target feature point in response to a first user operation, where the first user operation is an operation by a user to mark the target feature point in the ultrasonic image to be detected; and / or The first area is determined in response to a second user operation, where the second user operation is an operation of a user marking a first area where the object to be detected is located in the ultrasonic image to be detected.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, the following steps are implemented: Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected; Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected; The first characteristic value and / or the second characteristic value is displayed.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the following steps are implemented: Acquire an ultrasonic image to be detected, wherein the ultrasonic image to be detected includes a target to be detected, and the target to be detected is an area with abnormal grayscale in the ultrasonic image to be detected; Calculate a first eigenvalue and / or a second eigenvalue according to the ultrasonic image to be detected, wherein the first eigenvalue is a grayscale difference value of target feature points inside and outside the boundary line of the target to be detected, which is used to characterize the boundary clarity of the target to be detected, and the second eigenvalue is a grayscale difference value of a maximum grayscale value and an average grayscale value of a first region where the target to be detected is located, which is used to characterize the strength of the space effect of the target to be detected; The first characteristic value and / or the second characteristic value is displayed.
