Gynecological pelvic cavity auxiliary ultrasound report automatic generation method

Abstract

The invention relates to the technical field of ultrasonic report generation, in particular to a gynecological pelvic cavity auxiliary ultrasonic report automatic generation method, which comprises the following steps of: performing edge detection on an acquired gynecological pelvic cavity ultrasonic image, and screening out a closed edge with preliminarily uniform gray scale; determining the pelvic effusion possibility corresponding to each target boundary; screening out candidate boundaries of pelvic effusion; determining the region consistency corresponding to each pelvic effusion candidate boundary; according to the consistency of all the regions, screening out a target region representing real pelvic effusion from the regions defined by all the candidate boundaries of the pelvic effusion; according to the area of the target area and the gray level confusion degree in the target area, determining the pathological possibility corresponding to the target area; and generating a gynecological pelvic assisted ultrasound report based on the pathologic likelihood. According to the method and the device, automatic generation of the gynecological pelvic cavity auxiliary ultrasonic report is realized, so that clinical doctors can be assisted in final assistance of the gynecological pelvic cavity, and the efficiency of generating the gynecological pelvic cavity auxiliary ultrasonic report is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic report generation, and particularly relates to a method for automatically generating a gynecological pelvic assisted ultrasonic report. Background Art

[0002] With the development of technology, ultrasonic images are more and more widely used, and a large number of ultrasonic images are often generated. For example, clinicians often observe gynecological pelvic ultrasonic images and write gynecological pelvic assisted ultrasonic reports, which is a time-consuming task and requires rich professional knowledge. However, with the increase in gynecological pelvic ultrasonic images, when generating gynecological pelvic assisted ultrasonic reports, if only relying on clinicians to observe gynecological pelvic ultrasonic images and write gynecological pelvic assisted ultrasonic reports by clinicians, it may lead to some gynecological pelvic ultrasonic images being difficult to be observed by clinicians in time, resulting in poor efficiency in generating gynecological pelvic assisted ultrasonic reports. Summary of the Invention

[0003] In order to solve the technical problem of poor efficiency in generating gynecological pelvic assisted ultrasonic reports, the present invention proposes a method for automatically generating a gynecological pelvic assisted ultrasonic report.

[0004] In a first aspect, the present invention provides a method for automatically generating a gynecological pelvic assisted ultrasonic report, the method comprising: Obtaining a gynecological pelvic ultrasonic image, performing edge detection on the gynecological pelvic ultrasonic image, screening out closed edges from the detected edges, and screening out closed edges with initially uniform gray levels from all the closed edges as target boundaries; Determining the possibility of pelvic effusion corresponding to each target boundary according to the change between the chain code values corresponding to all edge pixel points on each target boundary and the gray level distribution on each target boundary; Screening out candidate boundaries of pelvic effusion from all the target boundaries according to all the possibilities of pelvic effusion; Determining the regional consistency corresponding to each candidate boundary of pelvic effusion according to the area of the region enclosed by each candidate boundary of pelvic effusion and the gray level distribution within the region enclosed by each candidate boundary of pelvic effusion; Screening out a target region representing real pelvic effusion from the regions enclosed by all the candidate boundaries of pelvic effusion according to all the regional consistencies; Determining the pathological possibility corresponding to the target region according to the area of the target region and the degree of gray level chaos therein; Generating a gynecological pelvic assisted ultrasonic report based on the pathological possibility.

[0005] Combined with the above first aspect, in a possible implementation manner, the screening out closed edges with initially uniform gray levels from all the closed edges as target boundaries includes: Cluster all the pixel points in the gynecological pelvic ultrasound image according to the gray values corresponding to all the pixel points in the gynecological pelvic ultrasound image, to obtain target clustering clusters; If at least a preset proportion of the pixel points within the region enclosed by the closed edge belong to the same target clustering cluster, then determine the closed edge as the target boundary.

[0006] Combined with the first aspect above, in a possible implementation manner, the determining the pelvic effusion possibility corresponding to each target boundary according to the change between the chain code values corresponding to all the edge pixel points on each target boundary and the gray distribution on each target boundary includes: Determine the chain code change factor corresponding to each edge pixel point on each target boundary according to the change between the chain code values corresponding to each edge pixel point on each target boundary and the two edge pixel points adjacent to its position. Determine the pelvic effusion possibility corresponding to each target boundary according to the chain code change factors corresponding to all the edge pixel points on each target boundary and the difference between the gray value corresponding to each edge pixel point on each target boundary and its gray mean value.

[0007] Combined with the first aspect above, in a possible implementation manner, the determining the chain code change factor corresponding to each edge pixel point on each target boundary according to the change between the chain code values corresponding to each edge pixel point on each target boundary and the two edge pixel points adjacent to its position includes: Determine any one target boundary as the marked boundary, determine any one edge pixel point on the marked boundary as the marked pixel point, and respectively determine the two edge pixel points adjacent to the marked pixel point on the marked boundary as the first reference point and the second reference point; Determine the absolute value of the difference between the chain code value corresponding to the marked pixel point and the chain code value corresponding to the first reference point as the first chain code difference corresponding to the marked pixel point; Determine the absolute value of the difference between the chain code value corresponding to the marked pixel point and the chain code value corresponding to the second reference point as the second chain code difference corresponding to the marked pixel point; Determine the mean value of the first chain code difference and the second chain code difference corresponding to the marked pixel point as the chain code change factor corresponding to the marked pixel point.

[0008] Combined with the first aspect above, in a possible implementation manner, the formula for the pelvic effusion possibility corresponding to the target boundary is: ; where is the pelvic effusion possibility corresponding to the th target boundary; is the serial number of the target boundary; is the normalization function; is the i number of upper edge pixels of the is the i serial number of the upper edge pixel of the is the natural exponential function; is the absolute value function; is the i gray value corresponding to the th edge pixel on the is the i average value of the gray values corresponding to all edge pixels on the is the i gray value corresponding to the th edge pixel on the

[0009] Combined with the first aspect above, in a possible implementation manner, the screening of pelvic effusion candidate boundaries from all target boundaries according to all pelvic effusion possibilities includes: If the pelvic effusion possibility corresponding to the target boundary is greater than the preset pelvic effusion threshold, then the target boundary is determined as the pelvic effusion candidate boundary.

[0010] Combined with the first aspect above, in a possible implementation manner, the formula for the regional consistency corresponding to the pelvic effusion candidate boundary is: ; where is the j regional consistency corresponding to the j th pelvic effusion candidate boundary; is the normalization function; is the j standard deviation of the gray values corresponding to all pixels within the region enclosed by the is the j area of the region enclosed by the is the j average value of the gray values corresponding to all pixels within the region enclosed by the

[0011] Combined with the first aspect above, in a possible implementation manner, the screening of the target region representing the true pelvic effusion from the regions enclosed by all pelvic effusion candidate boundaries includes: Screen out the pelvic effusion candidate boundary with the largest corresponding regional consistency from all pelvic effusion candidate boundaries as the temporary boundary; If the regional consistency corresponding to the temporary boundary is greater than a preset consistency threshold, the area enclosed by the temporary boundary is determined as the target area.

[0012] Combined with the first aspect above, in a possible implementation, the determining the pathological possibility corresponding to the target area according to the area of the target area and the degree of gray-scale chaos therein includes: Determining the information entropy of the gray-scale values corresponding to all pixel points within the target area as the degree of gray-scale chaos corresponding to the target area; Determining the pathological possibility corresponding to the target area according to the area and the degree of gray-scale chaos of the target area, as well as the standard deviation and the mean value of the gray-scale values corresponding to all pixel points within the target area.

[0013] Combined with the first aspect above, in a possible implementation, the formula corresponding to the pathological possibility of the target area is: ; wherein, A is the pathological possibility corresponding to the target area; is a normalization function; v is the area of the target area; H is the degree of gray-scale chaos corresponding to the target area; is the standard deviation of the gray-scale values corresponding to all pixel points within the target area; is the mean value of the gray-scale values corresponding to all pixel points within the target area.

[0014] In a second aspect, the present invention provides a system for automatically generating a gynecological pelvic assisted ultrasound report, the system comprising: An acquisition and screening module, configured to acquire a gynecological pelvic ultrasound image, perform edge detection on the gynecological pelvic ultrasound image, and screen out closed edges from the detected edges, and screen out closed edges with initially uniform gray-scale from all the closed edges as target boundaries; A pelvic effusion possibility determination module, configured to determine the pelvic effusion possibility corresponding to each target boundary according to the change between the chain code values corresponding to all edge pixel points on each target boundary and the gray-scale distribution on each target boundary; A pelvic effusion candidate boundary screening module, configured to screen out pelvic effusion candidate boundaries from all the target boundaries according to all the pelvic effusion possibilities; A regional consistency determination module, configured to determine the regional consistency corresponding to each pelvic effusion candidate boundary according to the area of the area enclosed by each pelvic effusion candidate boundary and the gray-scale distribution within the area enclosed by each pelvic effusion candidate boundary; A target area screening module, configured to screen out a target area representing true pelvic effusion from the areas enclosed by all candidate boundaries of pelvic effusion according to the consistency of all areas; A pathological possibility determination module, configured to determine the pathological possibility corresponding to the target area according to the area of the target area and the degree of gray-scale disorder therein; A gynecological pelvic assisted ultrasound report generation module, configured to generate a gynecological pelvic assisted ultrasound report based on the pathological possibility.

[0015] In a third aspect, a server is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the method in the first aspect or any possible implementation manner of the first aspect.

[0016] In a fourth aspect, a computer program product is provided, which includes: computer program code, when the computer program code runs on a computer, enabling the computer to execute the method in the first aspect or any possible implementation manner of the first aspect.

[0017] In a fifth aspect, a computer-readable storage medium is provided, which stores computer program code, when the computer program code runs on a computer, enabling the computer to execute the method in the first aspect or any possible implementation manner of the first aspect.

[0018] The present invention has the following beneficial effects: The automatic generation method of a gynecological pelvic assisted ultrasound report of the present invention realizes the automatic generation of a gynecological pelvic assisted ultrasound report, thereby assisting clinicians in the final assistance for gynecological pelvic, solving the technical problem of poor efficiency in generating a gynecological pelvic assisted ultrasound report, and thus improving the efficiency of generating a gynecological pelvic assisted ultrasound report. Compared with only relying on clinicians to generate a gynecological pelvic assisted ultrasound report, the present invention analyzes gynecological pelvic ultrasound images, quantifies multiple features related to the pathological conditions of pelvic effusion, such as the possibility of pelvic effusion, regional consistency, and pathological possibility, and realizes the automatic generation of a gynecological pelvic assisted ultrasound report based on the pathological possibility, thereby improving the efficiency of generating a gynecological pelvic assisted ultrasound report and assisting clinicians in the final assistance for gynecological pelvic to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0020] Figure 1 It is a flowchart of a method for automatically generating a gynecological pelvic assisted ultrasound report according to the present invention; Figure 2 It is a schematic diagram of the composition structure of a system for automatically generating a gynecological pelvic assisted ultrasound report according to the present invention; Figure 3 It is a schematic diagram of the structure of a computer device according to the present invention. Detailed Embodiments

[0021] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, will detail the specific embodiments, structures, features, and effects of the technical solutions proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0023] Reference Figure 1 , which shows the flow of some embodiments of a method for automatically generating a gynecological pelvic assisted ultrasound report according to the present invention. The method for automatically generating a gynecological pelvic assisted ultrasound report includes the following steps: Step S1, obtain a gynecological pelvic ultrasound image, perform edge detection on the gynecological pelvic ultrasound image, and select closed edges from the detected edges. Then, select closed edges with initially uniform gray levels from all the closed edges as the target boundaries.

[0024] Among them, pelvic ultrasound is a commonly used medical examination method, which mainly examines the uterine condition, ovarian condition, pelvic effusion, etc. The gynecological pelvic ultrasound image is an ultrasound image obtained during the pelvic ultrasound process.

[0025] As an example, this step may include the following steps: The first step is to obtain a gynecological pelvic ultrasound image.

[0026] For example, a gynecological pelvic ultrasound image can be collected through an ultrasound probe.

[0027] It should be noted that after a patient completes an ultrasound examination in the hospital, examination information such as ultrasound images and the corresponding patient information are often stored in the hospital's database. When it is necessary to issue an examination report for the patient's ultrasound examination, the system often first reads the patient identification information, such as the patient's name, registration time, and department for consultation. Based on the above patient identification information, the corresponding ultrasound image information is read from the database. It should be noted that it is often necessary to determine whether the current ultrasound image has been analyzed and an auxiliary report has been issued. If not, the corresponding ultrasound image is retrieved for subsequent analysis.

[0028] In the second step, edge detection is performed on the gynecological pelvic ultrasound image.

[0029] For example, the Otsu threshold method can be used to perform binary segmentation on the gynecological pelvic ultrasound image to obtain a binary image. Then, through the Sobel operator, edge detection is performed on the above binary image to obtain an edge image. Among them, the edges in the edge image, that is, the corresponding edges in the gynecological pelvic ultrasound image, thus realizing the edge detection of the gynecological pelvic ultrasound image.

[0030] In the third step, closed edges are screened out from the detected edges.

[0031] Among them, the closed edge can be the boundary enclosing a closed area.

[0032] In the fourth step, screening out the closed edges with preliminarily uniform gray levels from all the closed edges and taking them as the target boundaries may include the following sub-steps: In the first sub-step, according to the gray values corresponding to all the pixel points in the above gynecological pelvic ultrasound image, all the pixel points in the above gynecological pelvic ultrasound image are clustered to obtain target clustering clusters.

[0033] For example, according to the gray values corresponding to all the pixel points in the gynecological pelvic ultrasound image, the K-means clustering algorithm can be used to cluster the pixel points with similar gray values in the gynecological pelvic ultrasound image into the same clustering cluster, and each clustering cluster obtained at this time is taken as the target clustering cluster. Among them, when using K-means clustering, the selection of the K value and the initial point can be determined by the elbow method.

[0034] In the second sub-step, if at least a preset proportion of the pixel points within the area enclosed by the closed edge belong to the same target clustering cluster, the closed edge is determined as the target boundary.

[0035] Among them, the preset proportion can be a pre-set percentage, which can be equal to 80%.

[0036] For example, if the preset proportion is equal to 80%, and 85% of the pixel points within the region enclosed by a certain closed edge belong to the same target clustering cluster, then this closed edge can be determined as the target boundary.

[0037] Optionally, the method for obtaining the target boundary can also be: performing binary segmentation on the gynecological pelvic ultrasound image using the Otsu threshold method to obtain the binary image of the gynecological pelvic ultrasound image; then performing edge detection on the binary image through the Sobel operator to obtain the edge image of the binary image; according to the gray values corresponding to all pixel points in the gynecological pelvic ultrasound image, clustering the pixel points in the gynecological pelvic ultrasound image through the K-means clustering algorithm to obtain the clustering image; after completing the clustering, using the edge image as a mask to cover the clustering image so as to obtain each different region in the gynecological pelvic ultrasound image and its corresponding edge line, and at this time, the obtained edge line is the target boundary.

[0038] It should be noted that pelvic fluid often appears as an anechoic region as a whole on the ultrasound image, that is, it appears as a black region as a whole. This is because the liquid does not reflect ultrasonic waves, so it appears as anechoic as a whole. And the liquid usually accumulates in the low-lying parts of the pelvis, such as the rectouterine pouch or the vesicouterine pouch. Therefore, pelvic fluid often appears as an aggregated black region as a whole in the ultrasound image, and it is often a closed and relatively smooth-connected domain with an edge. However, there are often many dot-like echoes in other parts of the ultrasound image. Therefore, through steps such as edge detection and clustering in the embodiments of the present invention, these dot-like echoes are separated from the fluid accumulation region, that is, a target boundary can represent the pelvic fluid boundary or a dot-like echo boundary.

[0039] Step S2, according to the change between the chain code values corresponding to all edge pixels on each target boundary, and the gray distribution on each target boundary, determine the possibility of pelvic fluid corresponding to each target boundary.

[0040] Among them, the chain code value corresponding to the edge pixel can be represented by the 8-connected chain code value of the edge pixel.

[0041] As an example, this step may include the following steps: The first step, according to the change between the chain code values corresponding to each edge pixel on each target boundary and the two edge pixels adjacent to its position, determining the chain code change factor corresponding to each edge pixel on each target boundary may include the following sub-steps: The first sub-step, determining any one target boundary as the marked boundary, determining any one edge pixel on the above-mentioned marked boundary as the marked pixel, and respectively determining the two edge pixels adjacent to the marked pixel on the above-mentioned marked boundary as the first reference point and the second reference point.

[0042] The second sub-step is to determine the absolute value of the difference between the chain code value corresponding to the above-mentioned marked pixel point and the chain code value corresponding to the above-mentioned first reference point as the first chain code difference corresponding to the above-mentioned marked pixel point.

[0043] The third sub-step is to determine the absolute value of the difference between the chain code value corresponding to the above-mentioned marked pixel point and the chain code value corresponding to the above-mentioned second reference point as the second chain code difference corresponding to the above-mentioned marked pixel point.

[0044] The fourth sub-step is to determine the average value of the first chain code difference and the second chain code difference corresponding to the above-mentioned marked pixel point as the chain code change factor corresponding to the above-mentioned marked pixel point.

[0045] The second step is to determine the pelvic effusion possibility corresponding to each target boundary according to the chain code change factors corresponding to all edge pixel points on each target boundary and the difference between the gray value corresponding to each edge pixel point on each target boundary and its gray mean value.

[0046] For example, the formula for determining the pelvic effusion possibility corresponding to the target boundary can be: ; where is the pelvic effusion possibility corresponding to the th target boundary; is the serial number of the target boundary; is the normalization function; is the i th number of edge pixel points on the th target boundary; i is the serial number of the th edge pixel point on the th target boundary; is the i th natural exponential function; is the absolute value function; is the i th gray value corresponding to the th edge pixel point on the i th target boundary; is the mean value of the gray values corresponding to all edge pixel points on the

[0047] It should be noted that in actual situations, the boundary of pelvic effusion often presents the characteristics of being smooth and continuous. The change directions between adjacent edge pixels on it are often relatively similar. That is to say, the change degree of the chain code between adjacent edge pixels on the boundary of pelvic effusion is often relatively small. While the shape change of the boundary of punctate echoes is often relatively uneven, and the change directions between adjacent edge pixels on it may not be very similar. That is, the change degree of the chain code between adjacent edge pixels on the boundary of punctate echoes may be relatively large. Therefore, when is smaller, it often indicates that the i th target boundary has a smaller change degree of the chain code between the th edge pixel and its adjacent edge pixel. It often indicates that the th edge pixel belongs to a target boundary that is more likely to be the boundary of pelvic effusion. When is smaller, it often indicates that the i th edge pixel on the th target boundary is more similar to the average gray level of the i th target boundary. It often indicates that the gray level change on the i th target boundary is relatively more uniform. It often indicates that the change between edge pixels on the i th target boundary is relatively smoother. It often indicates that the th edge pixel belongs to a target boundary that is more likely to be the boundary of pelvic effusion. Therefore, when is larger, it often indicates that the i th target boundary is more likely to be the boundary of pelvic effusion.

[0048] Step S3: According to all the possibilities of pelvic effusion, screen out the candidate boundaries of pelvic effusion from all the target boundaries.

[0049] As an example, if the possibility of pelvic effusion corresponding to the target boundary is greater than the preset pelvic effusion threshold, the target boundary can be determined as the candidate boundary of pelvic effusion. Among them, the preset pelvic effusion threshold can be a threshold set in advance, and it can be 0.8.

[0050] It should be noted that when the possibility of pelvic effusion corresponding to the target boundary is greater, it often indicates that the target boundary is more likely to be the boundary of pelvic effusion. Therefore, the candidate boundary of pelvic effusion represents the target boundary that is preliminarily screened out and may be the boundary of pelvic effusion.

[0051] Step S4: According to the area of the region enclosed by each candidate boundary of pelvic effusion and the gray level distribution within the region enclosed by each candidate boundary of pelvic effusion, determine the regional consistency corresponding to each candidate boundary of pelvic effusion.

[0052] As an example, the formula for determining the regional consistency corresponding to the candidate boundary of pelvic effusion can be: ; wherein, is the regional consistency corresponding to the j th pelvic fluid candidate boundary. j is the serial number of the pelvic fluid candidate boundary. is the normalization function. is the j th standard deviation of the gray values corresponding to all pixel points within the region enclosed by the th pelvic fluid candidate boundary. j is the area of the region enclosed by the th pelvic fluid candidate boundary. j is the mean value of the gray values corresponding to all pixel points within the region enclosed by the

[0053] It should be noted that in actual situations, pelvic fluid often appears as an anechoic region on ultrasound images, that is, it appears as a black region as a whole, and its gray value changes are often relatively uniform, and its overall gray value is relatively low. Compared with punctate echoes, the area of pelvic fluid is often larger. When is smaller, it often indicates that the gray value change within the region enclosed by the j th pelvic fluid candidate boundary is relatively more uniform. When is larger, it often indicates that the area of the region enclosed by the j th pelvic fluid candidate boundary is relatively larger. When is smaller, it often indicates that the overall gray value of the region enclosed by the j th pelvic fluid candidate boundary is lower. Therefore, when is larger, it often indicates that the region enclosed by the j th pelvic fluid candidate boundary is more likely to be the pelvic fluid region.

[0054] Step S5, according to all regional consistencies, screen out the target region representing the true pelvic fluid from the regions enclosed by all pelvic fluid candidate boundaries.

[0055] As an example, this step may include the following steps: The first step is to screen out the pelvic fluid candidate boundary with the largest corresponding regional consistency from all pelvic fluid candidate boundaries as the temporary boundary.

[0056] The second step is that if the regional consistency corresponding to the above temporary boundary is greater than the preset consistency threshold, then the region enclosed by the above temporary boundary is determined as the target region.

[0057] Among them, the preset consistency threshold can be a threshold set in advance, and it can be equal to 0.7.

[0058] Step S6: Determine the pathological possibility corresponding to the target area according to the area of the target area and the degree of gray-scale chaos therein.

[0059] As an example, this step may include the following steps: First step: Determine the entropy of the gray-scale values corresponding to all pixel points within the above-mentioned target area as the degree of gray-scale chaos corresponding to the above-mentioned target area.

[0060] Second step: Determine the pathological possibility corresponding to the target area according to the area and gray-scale chaos degree of the above-mentioned target area, as well as the standard deviation and mean value of the gray-scale values corresponding to all pixel points within the target area.

[0061] For example, the formula for determining the pathological possibility corresponding to the target area may be: ; where A is the pathological possibility corresponding to the target area. is the normalization function. v is the area of the target area. is the degree of gray-scale chaos corresponding to the target area. is the standard deviation of the gray-scale values corresponding to all pixel points within the target area. is the mean value of the gray-scale values corresponding to all pixel points within the target area.

[0062] It should be noted that in actual situations, compared with physiological pelvic fluid, the area of pathological pelvic fluid is often larger. For example, pathological fluid sometimes can fill the entire pelvic cavity. And pathological pelvic fluid often contains a certain amount of blood or pus, etc., which may cause irregular echoes or layering phenomena inside the fluid; if the fluid contains pus or gas, it may show bubble-like echoes in the ultrasound image, which are manifested as bright spots or bubble-shaped shadows. Therefore, the degree of gray-scale uniformity of pathological pelvic fluid reflected in the ultrasound image is often lower than that of physiological pelvic fluid; similarly, the gray-scale value of pathological pelvic fluid reflected in the ultrasound image is often slightly higher than that of physiological pelvic fluid. When v is larger, it often indicates that the area of the target area representing the real pelvic fluid is larger. When H and are smaller, it often indicates that the gray-scale within the target area representing the real pelvic fluid is more uniform. When is smaller, it often indicates that the gray-scale within the target area representing the real pelvic fluid is smaller. Therefore, when A is larger, it often indicates that the target area is more likely to represent pathological pelvic fluid.

[0063] Step S7: Generate a gynecological pelvic assisted ultrasound report based on the pathological possibility.

[0064] As an example, by analyzing and calculating the ultrasound image, the pelvic effusion area therein was extracted, and the possibility that the pelvic effusion area is pathological was roughly calculated. Next, based on the pathological possibility calculated above, combined with the keywords of the ultrasound image output by the Transformer, the quality of the ultrasound report can be improved to assist clinicians in providing a more accurate and reliable report. The structure of the ultrasound report keyword generation model based on the Transformer proposed in the embodiment of the present invention can be divided into two parts. The first part is a shared convolutional neural network, which, as an extractor, extracts the visual features of the above-mentioned ultrasound image, and its specific feature calculation method is the pathological possibility calculation method; the second part is a text encoder, which selects the Transformer as the backbone network, including a Transformer encoder and a decoder, encodes the text feature keywords into vector representations, and assists in generating a gynecological pelvic assisted ultrasound report.

[0065] Optionally, based on the pathological possibility, the method for generating a gynecological pelvic assisted ultrasound report can also be: if the pathological possibility is greater than a preset pathological threshold, it can be preliminarily determined that the target area is pathological pelvic effusion, and a prompt message indicating the existence of pathological pelvic effusion is generated, which constitutes a gynecological pelvic assisted ultrasound report. Among them, the preset pathological threshold can be a threshold set in advance, and it can be 0.8.

[0066] Reference Figure 2 , based on the same inventive concept as the above method embodiment, the present invention provides a gynecological pelvic assisted ultrasound report automatic generation system, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the above computer program is executed by the processor, the steps of a method for automatically generating a gynecological pelvic assisted ultrasound report are implemented, which may specifically include: An acquisition and detection screening module 201, configured to acquire a gynecological pelvic ultrasound image, perform edge detection on the gynecological pelvic ultrasound image, and screen out closed edges from the detected edges, and screen out closed edges with initially uniform gray levels from all closed edges as target boundaries; A pelvic effusion possibility determination module 202, configured to determine the pelvic effusion possibility corresponding to each target boundary according to the change between the chain code values corresponding to all edge pixels on each target boundary and the gray level distribution on each target boundary; A pelvic effusion candidate boundary screening module 203, configured to screen out pelvic effusion candidate boundaries from all target boundaries according to all pelvic effusion possibilities; A regional consistency determination module 204, configured to determine the regional consistency corresponding to each pelvic effusion candidate boundary according to the area of the region enclosed by each pelvic effusion candidate boundary and the gray level distribution within the region enclosed by each pelvic effusion candidate boundary; A target area screening module 205, configured to screen out a target area representing true pelvic fluid from the areas enclosed by all pelvic fluid candidate boundaries according to the consistency of all areas; A pathological possibility determination module 206, configured to determine the pathological possibility corresponding to the target area according to the area of the target area and the degree of gray-scale disorder therein; A gynecological pelvic assisted ultrasound report generation module 207, configured to generate a gynecological pelvic assisted ultrasound report based on the pathological possibility.

[0067] Figure 3 It is a schematic structural diagram of a computer device provided by an embodiment of the present invention. Exemplarily, as Figure 3 shown, the computer device 300 includes: a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. Wherein, when the processor 302 executes the computer program 303, the computer device can execute any one of the foregoing gynecological pelvic assisted ultrasound report automatic generation methods.

[0068] Based on the same inventive concept as the above method embodiment, the present invention provides a server, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes any one of the foregoing gynecological pelvic assisted ultrasound report automatic generation methods.

[0069] Based on the same inventive concept as the above method embodiment, the present invention provides a computer program product, which includes: computer program code. When the computer program code runs on a computer, the computer executes any one of the foregoing gynecological pelvic assisted ultrasound report automatic generation methods.

[0070] Based on the same inventive concept as the above method embodiment, the present invention provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer executes any one of the foregoing gynecological pelvic assisted ultrasound report automatic generation methods.

[0071] In summary, compared with only relying on clinicians to generate gynecological pelvic assisted ultrasound reports, the present invention analyzes gynecological pelvic ultrasound images, quantifies multiple features related to the pathological conditions of pelvic fluid, such as the possibility of pelvic fluid, regional consistency, and pathological possibility, and based on the pathological possibility, realizes the automatic generation of gynecological pelvic assisted ultrasound reports, thereby improving the efficiency of generating gynecological pelvic assisted ultrasound reports and can assist clinicians in the final assistance for gynecological pelvis to a certain extent.

[0072] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the various embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for automatically generating a gynecological pelvic auxiliary ultrasound report, characterized in that: The following steps are involved: Acquire a gynecological pelvic ultrasound image, perform edge detection on the gynecological pelvic ultrasound image, and select closed edges from the detected edges, and select closed edges with preliminary uniform grayscale from all closed edges as target boundaries; According to the changes between the chain code values ​​corresponding to all edge pixels on each target boundary and the grayscale distribution on each target boundary, the possibility of pelvic effusion corresponding to each target boundary is determined; According to all pelvic effusion possibilities, candidate boundaries of pelvic effusion are screened out from all target boundaries; Determine the consistency of the region corresponding to each pelvic effusion candidate boundary according to the area of ​​the region enclosed by each pelvic effusion candidate boundary and the grayscale distribution within the region enclosed by each pelvic effusion candidate boundary; According to the consistency of all regions, the target region representing the real pelvic effusion is screened out from the regions enclosed by all candidate boundaries of pelvic effusion; According to the area of ​​the target area and the grayscale disorder degree within it, the pathological possibility corresponding to the target area is determined; Generate a gynecological pelvic ultrasound report based on the possibility of pathology.

2. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 1, characterized in that: The step of selecting closed edges with initially uniform grayscale from all closed edges as target boundaries includes: Clustering all pixel points in the gynecological pelvic ultrasound image according to the grayscale values ​​corresponding to all pixel points in the gynecological pelvic ultrasound image to obtain a target cluster; If at least a preset percentage of pixels in the area enclosed by the closed edge belong to the same target cluster, the closed edge is determined as the target boundary.

3. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 1, characterized in that: Determining the possibility of pelvic effusion corresponding to each target boundary according to the changes between the chain code values ​​corresponding to all edge pixels on each target boundary and the grayscale distribution on each target boundary includes: According to the change between the chain code values ​​corresponding to each edge pixel point on each target boundary and the two edge pixel points adjacent to it, the chain code change factor corresponding to each edge pixel point on each target boundary is determined; According to the chain code change factors corresponding to all edge pixels on each target boundary and the difference between the grayscale values ​​corresponding to each edge pixel on each target boundary and their grayscale mean, the possibility of pelvic effusion corresponding to each target boundary is determined.

4. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 3, characterized in that: The determining of the chain code change factor corresponding to each edge pixel point on each target boundary according to the change between the chain code values ​​corresponding to each edge pixel point on each target boundary and two edge pixel points adjacent to each other at the same position comprises: Determine any target boundary as a marked boundary, determine any edge pixel point on the marked boundary as a marked pixel point, and determine two edge pixel points on the marked boundary adjacent to the marked pixel point as a first reference point and a second reference point, respectively; Determine the absolute value of the difference between the chain code value corresponding to the marked pixel point and the chain code value corresponding to the first reference point as the first chain code difference corresponding to the marked pixel point; Determine the absolute value of the difference between the chain code value corresponding to the marked pixel point and the chain code value corresponding to the second reference point as the second chain code difference corresponding to the marked pixel point; The average of the first chain code difference and the second chain code difference corresponding to the marked pixel point is determined as the chain code change factor corresponding to the marked pixel point.

5. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 3, characterized in that: The formula corresponding to the probability of pelvic effusion corresponding to the target boundary is: ;in, It is The probability of pelvic effusion corresponding to the target boundary; is the ordinal number of the target boundary; is the normalization function; It is i The number of edge pixels on the target boundary; It is i The serial number of the edge pixel points on the target boundary; is a natural exponential function; It is the absolute value function; It is i The target boundary Gray value corresponding to edge pixels; It is i The mean of the gray values ​​corresponding to all edge pixels on the target boundary; It is i The target boundary The chain code change factor corresponding to each edge pixel.

6. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 1, characterized in that: According to all pelvic effusion possibilities, candidate boundaries of pelvic effusion are screened out from all target boundaries, including: If the possibility of pelvic effusion corresponding to the target boundary is greater than a preset pelvic effusion threshold, the target boundary is determined as a candidate boundary for pelvic effusion.

7. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 1, characterized in that: The formula corresponding to the regional consistency of the candidate boundary of pelvic effusion is: ;in, It is j The consistency of the regions corresponding to the candidate boundaries of pelvic effusion; j is the serial number of the candidate boundary of the pelvic effusion; is the normalization function; It is j The standard deviation of the grayscale values ​​corresponding to all pixels in the area enclosed by the candidate boundary of pelvic effusion; It is j The area enclosed by the candidate boundaries of the pelvic effusion; It is j The mean of the gray values ​​corresponding to all pixels in the area enclosed by the candidate boundary of the pelvic effusion.

8. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 1, characterized in that: According to the consistency of all regions, the target region representing the real pelvic effusion is screened out from the regions enclosed by all candidate boundaries of pelvic effusion, including: Select the pelvic effusion candidate boundary with the greatest corresponding regional consistency from all pelvic effusion candidate boundaries as the temporary boundary; If the consistency of the area corresponding to the temporary boundary is greater than a preset consistency threshold, the area enclosed by the temporary boundary is determined as the target area.

9. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 1, characterized in that: Determining the pathological possibility corresponding to the target region according to the area of ​​the target region and the grayscale disorder degree therein includes: Determine the information entropy of the grayscale values ​​corresponding to all pixels in the target area as the grayscale disorder degree corresponding to the target area; The pathological possibility corresponding to the target area is determined according to the area and grayscale disorder degree of the target area, and the standard deviation and mean of the grayscale values ​​corresponding to all pixels in the target area.

10. The method for automatically generating a gynecological pelvic auxiliary ultrasound report according to claim 9, characterized in that: The formula corresponding to the pathological possibility of the target area is: ;in, A is the pathological possibility corresponding to the target area; is the normalization function; v is the area of ​​the target region; H is the grayscale chaos degree corresponding to the target area; It is the standard deviation of the grayscale values ​​corresponding to all pixels in the target area; It is the mean of the grayscale values ​​corresponding to all pixels in the target area.

Images