Quality control processing method, device, equipment and storage medium for ultrasonic image and text data

Abstract

The present application discloses a quality control processing method, device, equipment and storage medium for ultrasonic image and text data. The method obtains characteristic information of an ultrasonic image, determines the quality evaluation result of the ultrasonic image based on the comparison result of a first quality control standard and the characteristic information; determines the quality evaluation result of text data based on a second quality control standard; and generates a quality control data preview page based on the quality evaluation result of the ultrasonic image, the quality evaluation result of the text data, and the matching result of the characteristic information and the text data. When the ultrasonic image is evaluated by the characteristic information of the ultrasonic image and the text data is evaluated by the preset text quality control standard, the quality control result of the ultrasonic image and text data is obtained by matching the image characteristic information with the key information, so that the quality control content of the ultrasonic image and text data is more comprehensive, and the quality control data preview page can intuitively reflect all the quality control data, thereby improving the quality control effect of the ultrasonic image and text data.

Description

Technical Field

[0001] The present disclosure generally relates to the field of data processing technology, and in particular to a quality control processing method, device, equipment and storage medium for ultrasonic image and text data. Background Art

[0002] With the rapid and high-quality development of medical imaging technology, ultrasound image and text data, as the main manifestation of the work content of the imaging department of medical institutions, can be used to determine the target biological object. Among them, ultrasound image and text data can include ultrasound images and ultrasound text reports.

[0003] However, since existing ultrasound images are easily limited by image quality and noise interference, and the quality standardization of ultrasound text reports is poor, the quality of ultrasound image and text data is easily poor, which leads to errors in the determination results of the target biological object.

[0004] In this regard, the existing technology allows superior hospitals or quality control centers to monitor the quality of ultrasound image and text data of subordinate hospitals, but this manual quality control method can easily lead to problems such as low quality control efficiency of ultrasound image and text data, strong subjectivity of quality control standards, limited quality control content, and narrow quality control coverage.

[0005] Therefore, the poor quality control effect of ultrasound image data in the existing technology has become a problem that needs to be solved. Summary of the invention

[0006] In view of the above-mentioned defects or deficiencies in the prior art, it is desired to provide a quality control processing method, device, equipment and storage medium for ultrasonic image and text data. The method obtains the quality control results of the ultrasonic image and text data by matching the image feature information with the text data, while quality evaluating the ultrasonic image based on the feature information of the ultrasonic image and quality evaluating the text data based on the preset text quality control standard. This makes the quality control content of the ultrasonic image and text data more comprehensive, and the quality control data preview page generated based on different quality evaluation data can intuitively reflect the quality control data of the ultrasonic image and text data, thereby improving the quality control effect of the ultrasonic image and text data.

[0007] In a first aspect, the present invention provides a quality control processing method for ultrasound image and text data, wherein the ultrasound image and text data include ultrasound images and text data, and the method includes:

[0008] Acquire characteristic information of the ultrasound image, and determine the quality evaluation result of the ultrasound image based on the comparison result of the first quality control standard and the characteristic information; wherein the first quality control standard is a preset quality control standard of the ultrasound image;

[0009] Determining a quality evaluation result of the text data based on a second quality control standard; wherein the second quality control standard is a preset quality control standard of the text data;

[0010] A quality control data preview page is generated based on the quality evaluation results of the ultrasound image, the quality evaluation results of the text data, and the matching results of the feature information and the text data.

[0011] In a second aspect, a quality control processing device for ultrasonic graphic data is provided, wherein the ultrasonic graphic data includes ultrasonic images and text data, and the device includes:

[0012] A first processing unit, configured to obtain characteristic information of the ultrasound image, and determine a quality evaluation result of the ultrasound image based on a comparison result between the first quality control standard and the characteristic information; wherein the first quality control standard is a preset quality control standard of the ultrasound image;

[0013] A second processing unit, used to determine a quality evaluation result of the text data based on a second quality control standard; wherein the second quality control standard is a preset quality control standard of the text data;

[0014] The generating unit is used to generate a quality control data preview page based on the quality evaluation result of the ultrasound image, the quality evaluation result of the text data, and the matching result of the feature information and the text data.

[0015] In a third aspect, a computer device is provided, comprising 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 method described in any one of the first aspects is implemented.

[0016] According to a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, wherein when the program is executed by a processor, the method according to any one of the above-mentioned first aspects is implemented.

[0017] Compared with the prior art that uses manual quality control to achieve quality control of ultrasonic image and text data, the quality control processing method, device, equipment and storage medium of ultrasonic image and text data provided by the embodiment of the present application can, on the one hand, obtain the quality evaluation result of the ultrasonic image by comparing the characteristic information of the ultrasonic image with the preset quality control standard of the ultrasonic image, so as to achieve quality control of the ultrasonic image; the quality evaluation result of the text data can be determined by using the preset quality control standard of the text data, so as to achieve quality control of the text data; based on this, the quality control efficiency of the ultrasonic image and text data can be improved, and the application of the preset quality control standard can avoid the situation that the quality control result has errors due to the strong subjectivity of the quality control standard of the manual quality control method. On the other hand, the consistency comparison of the ultrasonic image and text data can be achieved by matching the characteristic information of the ultrasonic image with the text data, so that the quality control content of the ultrasonic image and text data is more comprehensive, and the quality control data preview page generated based on the characteristic information of the ultrasonic image, the quality evaluation result of the text data and the consistency comparison result of the ultrasonic image and text data can intuitively reflect the quality control data of the ultrasonic image and text data, thereby improving the quality control effect of the ultrasonic image and text data. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0019] Figure 1 This is a diagram of an implementation environment applicable to the embodiments of the present application;

[0020] Figure 2 A schematic diagram of a process flow of a quality control processing method for ultrasound image and text data provided in an embodiment of the present application;

[0021] Figure 3 A schematic diagram of determining a target area provided in an embodiment of the present application;

[0022] Figure 4 A schematic diagram of a process for determining a quality evaluation result of text data provided in an embodiment of the present application;

[0023] Figure 5 A flowchart of a method for determining a matching result between feature information and text data provided in an embodiment of the present application;

[0024] Figure 6 A schematic diagram of a quality control data preview page provided in an embodiment of the present application;

[0025] Figure 7 A schematic diagram of a flow chart of a method for acquiring characteristic information of an ultrasound image provided in an embodiment of the present application;

[0026] Figure 8 A block diagram of a quality control processing device for ultrasonic image and text data provided in an embodiment of the present application;

[0027] Fig. 9 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0029] It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments. In addition, the term "and / or" herein is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The terms "first" and "second" in the description and claims of the embodiments of the present application are used to distinguish different objects, rather than to describe a specific order of objects.

[0030] First, the terms involved in this application are explained as follows:

[0031] (1) Picture archiving and communication system (PACS): a system used in the imaging department of medical institutions, mainly used to save various medical images in a digital way through various interfaces; the interfaces may include analog interfaces, DICOM interfaces, network interfaces, etc.; medical images may include images produced by MRI, CT, ultrasound, X-ray machines, infrared devices, microscopes, etc.;

[0032] (2) Target biological object: It is the scanning object of the ultrasound device, which can be a biological object suspected of having abnormal conditions, such as a lesion in the human body;

[0033] (3) BI-RADS grade: It is a grading standard for evaluating breast diseases, which includes categories 1, 2, 3, 4a, 4b, 4c, 5 and 6.

[0034] With the rapid and high-quality development of medical imaging technology, ultrasound image and text data, as the main manifestation of the work content of the imaging department of medical institutions, can be used to determine the target biological object. Among them, ultrasound image and text data can include ultrasound images and ultrasound text reports.

[0035] Specifically, an ultrasonic device can be used to scan biological objects, obtain an ultrasonic image based on the scanning results of the biological objects, and complete the writing of an ultrasonic text report in combination with a text report template and the specific content of the ultrasonic image; wherein the ultrasonic text report can be used to record the scanning results and specific characteristic information of abnormal biological objects.

[0036] However, in the prior art, ultrasound images are subject to image quality limitations and noise interference, which can easily lead to poor display effects of ultrasound images, resulting in certain errors in the determination results of characteristic information of ultrasound images; and in the process of writing ultrasound text reports, due to differences between scanned ultrasound images and retained ultrasound images, erroneous analysis of retained ultrasound images, logical contradictions in the written text content, typos and other problems, the quality standardization of ultrasound text reports is poor, and the writing efficiency is poor due to the large amount of written content.

[0037] Based on the above defects, existing ultrasound image and text data still have problems such as poor quality, long quality control time, poor quality control efficiency, and over-reliance on human judgment and experience. Such quality control problems can easily lead to certain errors in the determination results of target biological objects (for example, abnormal biological objects).

[0038] In this regard, the existing technology allows superior hospitals or quality control centers to monitor the quality of ultrasound image and text data of subordinate hospitals, but this manual quality control method is prone to problems such as low quality control efficiency of ultrasound image and text data, strong subjectivity of quality control standards, limited quality control content, and narrow quality control coverage. These problems make it difficult to provide timely, effective and comprehensive feedback on the quality control results of ultrasound image and text data to the relevant staff who generate the ultrasound image and text data, thereby hindering the goal of improving the accuracy and standardization of ultrasound scanning results.

[0039] Therefore, the poor quality control effect of ultrasound image data in the existing technology has become a problem that needs to be solved.

[0040] Based on this, an embodiment of the present application provides a quality control processing method, device, equipment and storage medium for ultrasonic image and text data. The method obtains the quality control results of the ultrasonic image and text data by matching the image feature information with the text data, while quality evaluating the ultrasonic image based on the feature information of the ultrasonic image and quality evaluating the text data based on the preset text quality control standard. This makes the quality control content of the ultrasonic image and text data more comprehensive, and the quality control data preview page generated based on different quality evaluation data can intuitively reflect the quality control data of the ultrasonic image and text data, thereby improving the quality control effect of the ultrasonic image and text data.

[0041] Figure 1 This is an implementation environment architecture diagram applicable to the embodiments of this application. Figure 1 As shown, the implementation environment may be a quality control processing system 10. Specifically, the quality control processing system 10 may include an image quality control module 101, a text quality control module 102, an image-text matching module 103, and a quality control data generation module 104. Exemplarily, the quality control processing system 10 may be deployed on a computer device.

[0042] Exemplarily, the image quality control module 101 can be used to compare the characteristic information of the ultrasound image using a preset image quality control standard, so as to determine the quality evaluation result of the ultrasound image based on the comparison result between the quality control standard and the ultrasound image, wherein the quality evaluation result of the ultrasound image can reflect the degree of standardization of the ultrasound image; specifically, the ultrasound image can be generated by an ultrasound device and directly imported into the image quality control module 101, or it can be extracted from the PACS system by the quality control processing system 10 and transmitted to the image quality control module 101. There is no specific limitation on the method of obtaining the ultrasound image here.

[0043] The text quality control module 102 can be used to determine the completeness and correctness of key information of the text data using preset text quality control standards, so as to determine the quality evaluation results of the text data based on the completeness and correctness of the key information, wherein the quality evaluation results of the text data can reflect the standardization and normativeness of the text data; specifically, the text data can be extracted from the PACS system by the quality control processing system 10 and transmitted to the text quality control module 102; for example, the text data includes but is not limited to electronic pictures, PDF documents, and picture files taken with a camera, etc.

[0044] The image-text matching module 103 may be used to determine a matching result between feature information of an ultrasound image and text data, so as to determine the consistency and matching degree between the ultrasound image and the text data based on the matching result.

[0045] The quality control data generation module 104 can be used to obtain different quality evaluation data generated by the image quality control module 101, the text quality control module 102 and the image-text matching module 103, and generate a quality control data preview page by summarizing the different quality evaluation data; wherein the quality control data preview page can be presented as a quality control report.

[0046] In the specific implementation, refer to Figure 1 The ultrasound image can be transmitted to the image quality control module 101, so that the image quality control module 101 generates a quality evaluation result of the ultrasound image, and transmits the quality evaluation result of the ultrasound image and the characteristic information of the ultrasound image to the image-text matching module 103 and the quality control data generation module 104 respectively.

[0047] The ultrasonic text data is transmitted to the text quality control module 102 so that the text quality control module 102 generates a quality evaluation result of the ultrasonic text data, and the quality evaluation result of the ultrasonic text data and the key information of the ultrasonic text data are transmitted to the image-text matching module 103 and the quality control data generation module 104 respectively.

[0048] Then, the image-text matching module 103 determines the matching result between the ultrasound image and the ultrasound text data based on the consistency comparison of the characteristic information of the ultrasound image and the key information of the ultrasound text data, and transmits the matching result between the ultrasound image and the ultrasound text data to the quality control data generation module 104 .

[0049] Finally, the quality control data generation module 104 summarizes the different quality evaluation data transmitted by the image quality control module 101, the text quality control module 102 and the image-text matching module 103 to form a quality control report for the ultrasound text data.

[0050] For example, Figure 2 Schematic diagram of a process flow of a quality control processing method for ultrasound image data provided in an embodiment of the present application. Figure 2 As shown, the method may include the following steps:

[0051] Step S201, acquiring characteristic information of an ultrasound image, and determining a quality evaluation result of the ultrasound image based on a comparison result between a first quality control standard and the characteristic information; wherein the first quality control standard is a preset quality control standard for ultrasound images.

[0052] Exemplarily, the ultrasound graphic data includes ultrasound images and text data; wherein the ultrasound image can be obtained by scanning with an ultrasound device, and the text data can be an ultrasound text report written based on a text report template combined with the specific content of the ultrasound image.

[0053] Compared with the prior art that uses image clarity as a quality evaluation standard, the embodiment of the present application utilizes a first quality control standard corresponding to the characteristic information of the ultrasound image to perform quality evaluation on the ultrasound image, thereby making the quality evaluation result of the ultrasound image more comprehensive and improving the quality control effect of the ultrasound image.

[0054] In a possible implementation, the ultrasound image may be divided into at least one target region based on preset region characteristics, so that feature information of the ultrasound image may be determined based on the region characteristics of each target region.

[0055] Exemplarily, the preset region characteristics for dividing the ultrasound image may be determined according to the display information of each region of the ultrasound image.

[0056] Specifically, when the display information of a certain area of ​​the ultrasound image is an image of a scanned object of the ultrasound device, the area characteristic of the area is a display characteristic (that is, the function of the area is to display the scanned object).

[0057] When the display information of a certain area of ​​the ultrasound image is the position information of the scanned object of the ultrasound device, the area characteristic of the area is an identification characteristic (that is, the function of the area is to identify the scanned object).

[0058] When the displayed information of a certain area of ​​the ultrasound image is a scale representing the actual physical distance, the regional characteristic of the area is a proportional characteristic (that is, the function of the area is the ratio of the size of the scanned object).

[0059] Based on this, the preset area characteristics include display characteristics, identification characteristics and ratio characteristics.

[0060] In a possible implementation, the quality assessment result of the ultrasound image may be determined based on a comparison result between the first quality control standard and the characteristic information.

[0061] Exemplarily, when the first quality control standard is represented by multiple standard regulations, the quality evaluation result of the ultrasound image can be determined based on the number of standard regulations satisfied by the feature information and the total number of standard regulations. The quality evaluation result of the ultrasound image can be represented by a score or text.

[0062] Step S202, determining a quality evaluation result of the text data based on a second quality control standard; wherein the second quality control standard is a preset quality control standard of the text data.

[0063] In one possible implementation, based on the second quality control standard, the completeness of key information in the text data and the accuracy of the text content in the text data can be determined, so as to determine the quality evaluation result of the text data based on the completeness of key information in the text data and the accuracy of the text content in the text data.

[0064] Exemplarily, the second quality control standard may include examples of key information and a criterion for judging the accuracy of text content, so as to determine the completeness of the key information based on the examples of key information and determine the accuracy of the text content based on the criterion for judging the accuracy of text content.

[0065] Exemplarily, the quality assessment result of the text data may be expressed as at least one of a score and text.

[0066] Step S203: generating a quality control data preview page based on the quality assessment results of the ultrasound image, the quality assessment results of the text data, and the matching results of the feature information and the text data.

[0067] In a possible implementation, the matching result between the feature information and the text data may be determined according to the consistency between the feature information and the description content of the text data.

[0068] Exemplarily, the abnormality type, abnormality position, abnormality size, abnormality level and other information of the target biological object obtained based on the feature information and based on the text data can be compared to obtain a matching result between the feature information and the text data based on the comparison result.

[0069] Compared with the prior art that uses manual quality control to achieve quality control of ultrasonic image and text data, the quality control processing method for ultrasonic image and text data provided in the embodiment of the present application can, on the one hand, obtain the quality evaluation result of the ultrasonic image by comparing the characteristic information of the ultrasonic image with the preset quality control standard of the ultrasonic image, so as to achieve quality control of the ultrasonic image; the quality evaluation result of the text data can be determined by using the preset quality control standard of the text data, so as to achieve quality control of the text data; based on this, the quality control efficiency of the ultrasonic image and text data can be improved, and the application of the preset quality control standard can avoid the situation that the quality control result has errors due to the strong subjectivity of the quality control standard of the manual quality control method. On the other hand, the consistency comparison of the ultrasonic image and text data can be achieved by matching the characteristic information of the ultrasonic image with the text data, so that the quality control content of the ultrasonic image and text data is more comprehensive, and the quality control data preview page generated based on the characteristic information of the ultrasonic image, the quality evaluation result of the text data and the consistency comparison result of the ultrasonic image and text data can intuitively reflect the quality control data of the ultrasonic image and text data, thereby improving the quality control effect of the ultrasonic image and text data.

[0070] In another embodiment of the present application, a specific method for determining the target area in the ultrasound image is also provided.

[0071] In a possible implementation manner, at least one target region of the ultrasound image may be acquired based on preset region characteristics.

[0072] Exemplarily, before dividing the ultrasound image to obtain at least one target region, based on step S201, the preset region characteristics for dividing the ultrasound image can be formed by determining the display information of each region of the historical ultrasound image.

[0073] It should be noted that the above historical ultrasound images are specifically historical ultrasound images obtained by scanning with the same ultrasound equipment or the same type of ultrasound equipment as the ultrasound image currently requiring quality assessment.

[0074] In a possible implementation, a region division model is trained based on preset region characteristics, so as to output at least one target region of an ultrasound image using the region division model.

[0075] Exemplarily, the region segmentation model is a target detection model obtained based on a deep learning algorithm, such as Faster-RCNN, YOLO v8, DETR and other models.

[0076] Exemplarily, a region of interest (ROI) for an ultrasound image may be obtained using preset region characteristics, so that a region segmentation model may be trained using regional images corresponding to each region of interest.

[0077] Specifically, a preset region characteristic may form at least one region of interest, and a region of interest corresponds to a target region.

[0078] For example, when the preset area characteristic is a display characteristic, the area of ​​interest corresponds to a display area; when the preset area characteristic is an identification characteristic, the area of ​​interest corresponds to an image identification area and a text identification area; when the preset area characteristic is a proportional characteristic, the area of ​​interest corresponds to a proportional area.

[0079] Exemplarily, the target area output by the area division model can be represented in the form of a data tuple.

[0080] Specifically, each target area corresponds to a data tuple (x, y, h, w, c, s), where x and y are used to represent the center point coordinates of the target area, h and w are used to represent the height and width of the target area, and c and s are used to represent the regional characteristics and confidence of the target area.

[0081] Based on this, the ultrasound image may be divided based on the data tuple corresponding to each target region to obtain at least one target region of the ultrasound image.

[0082] It should be noted that when the confidence s in a data tuple is too low, it means that there is a certain error in the regional characteristics of the target area corresponding to the data tuple. In this case, the target area can be ignored.

[0083] Exemplarily, based on the above example of the region of interest, the target region includes at least one of a display region, an image identification region, a text identification region, and a ratio region.

[0084] Specifically, the display area refers to the area where the image of the target biological object is displayed; the image identification area refers to the area where the position of the target biological object is identified in the form of an image, such as the probe identification position of an ultrasound device; the text identification area refers to the area where the position of the target biological object is identified in the form of text, such as the left and right directions and the clock direction; the scale area refers to the area that can provide a reference for the actual size of the target biological object (i.e., the scale area that represents the actual physical distance).

[0085] It should be noted that the image identification area and the text identification area usually do not exist in the same ultrasound image.

[0086] For example, Figure 3 is a schematic diagram of determining a target area provided in an embodiment of the present application, such as Figure 3 As shown, the ultrasound image can be divided into a display area A, an image mark area B, a text mark area C and a ratio area D using the area division model.

[0087] In another embodiment of the present application, a method for acquiring characteristic information of an ultrasound image is also provided.

[0088] In a possible implementation manner, feature information of the ultrasound image may be determined based on regional characteristics of at least one target area.

[0089] Exemplarily, basic image information corresponding to the target area is determined based on the regional characteristics of the target area, so that a classification result of the ultrasound image is determined by fusing the basic image information, and an abnormality level of the target biological object in the ultrasound image is determined based on the classification result of the ultrasound image, thereby the characteristic information of the ultrasound image is constituted by the basic image information of the ultrasound image, the classification result and the abnormality level of the target biological object.

[0090] In a possible implementation manner, for any target area, basic image information of an ultrasound image corresponding to the target area may be determined based on the regional characteristics of the target area.

[0091] The following is an introduction to the basic image information that can be determined based on different target areas:

[0092] First, when the target area is a display area, the area characteristics of the target area may be determined as display characteristics, that is, used to display the image of the target biological object.

[0093] Based on this, the ultrasound mode of the ultrasound image can be determined based on the image information of the target biological object, wherein the ultrasound mode includes at least one of a B-mode ultrasound mode, a color Doppler mode, a power Doppler mode, a shear wave elastic imaging mode, a strain wave elastic imaging mode, and a contrast imaging mode.

[0094] In one possible implementation, the image of the display area may be input into a pattern classification model to output an ultrasound pattern of the ultrasound image using the pattern classification model.

[0095] Exemplarily, the pattern classification model is a classification model obtained based on a deep learning algorithm, such as ResNet18, MobileNet v3, Vision Transformer (ViT) and other models.

[0096] Specifically, the pattern classification model may be trained using ultrasound images of known patterns, so that the pattern classification model receives an image of a display area as input and outputs an ultrasound pattern of the ultrasound image.

[0097] Secondly, when the target area is an image identification area or a text identification area, the area characteristics of the target area can be determined as identification characteristics, that is, used to identify the target biological object and its position.

[0098] Based on this, the name and coordinates of the target biological object in the ultrasound image (ie, the object scanned by the ultrasound device and its specific coordinates) can be determined based on the image of the image identification area or the text identification area.

[0099] In a possible implementation, based on the candidate body landmark set, an image retrieval method may be used to determine the name and coordinates of the target biological object.

[0100] Exemplarily, the candidate body marker set may include multiple identification regions, wherein the relevant information of the target biological object corresponding to each identification region is in a known state.

[0101] Specifically, the matching algorithm provided by the image retrieval method can be used to match the image of the image identification area / text identification area with the candidate body mark set to determine the identification area in the candidate body mark set that has the highest image matching degree with the image identification area / text identification area, and the name and coordinates of the target biological object corresponding to the identification area are used as the name and coordinates of the target biological object in the ultrasound image.

[0102] The matching algorithm provided by the image retrieval method may be a cosine similarity algorithm or a structural similarity algorithm.

[0103] For example, refer to Figure 3 By matching the corresponding images of the image identification area / text identification area, it can be determined that the name of the target biological object of the ultrasound image is the left breast and the coordinates are the three o'clock direction.

[0104] Optionally, the image of the image identification area / text identification area can be input into a multi-task classification model to output the name and coordinates of the target biological object in the ultrasound image using the multi-task classification model.

[0105] Exemplarily, the multi-task classification model is a classification model obtained based on a deep learning algorithm, such as ResNet18, MobileNet v3, Vision Transformer (ViT) and other models.

[0106] Specifically, the multi-task classification model can be trained using images of identification areas of known target biological object related information, so that the multi-task classification model receives input images of image identification areas / text identification areas and outputs the name and coordinates of the target biological object in the ultrasound image.

[0107] The multi-task classification model includes two classification output terminals, one of which is used to output the name of the target biological object in the ultrasound image, and the other classification output terminal is used to output the coordinates of the target biological object in the ultrasound image.

[0108] Thirdly, when the target region is a proportional region, the region characteristic of the target region can be determined as a proportional characteristic, that is, it can provide a reference for the actual size of the target biological object.

[0109] Based on this, the actual physical distance of a single pixel in the ultrasound image can be determined based on the image of the proportional area, thereby determining the actual size of the target biological object based on the actual physical distance of the single pixel.

[0110] In one possible implementation, based on the candidate scale set, an image retrieval method may be used to determine the actual physical distance of a single pixel in the ultrasound image.

[0111] Exemplarily, the candidate scale set may include multiple scale areas, wherein the actual physical distance between two scale points in each scale area is known.

[0112] Specifically, the matching algorithm provided by the image retrieval method can be used to match the image of the proportional area with the candidate scale set to determine the scale area in the candidate scale set with the highest matching degree with the image of the proportional area, and the actual physical distance between the two scale points in the scale area is used as the actual physical distance between the two scale points in the proportional area.

[0113] Based on the determined actual physical distance between two scale points in the proportional area, the actual physical distance of a single pixel in the ultrasound image is determined by extracting the actual scale points in the proportional area.

[0114] For example, the number of scale points in a single pixel may be determined, so that the product of the number of scale points in a single pixel and the actual physical distance between two scale points is determined as the actual physical distance of the single pixel in the ultrasound image.

[0115] Optionally, the image of the proportional region may be input into a regression model to output the actual physical distance of a single pixel in the ultrasound image using the regression model.

[0116] Exemplarily, the regression model is a regression model obtained based on a deep learning algorithm, such as ResNet18, MobileNet v3, Vision Transformer (ViT) and other models.

[0117] Specifically, the regression model may be trained using an ultrasound image with a known actual physical distance of a single pixel, so that the regression model receives an image of an input proportional region and outputs the actual physical distance of a single pixel in the ultrasound image.

[0118] Fourthly, when the target area is a display area, abnormal information of the target biological object may be determined based on the image of the target biological object represented by the display area.

[0119] In a possible implementation, the image of the display area may be input into an anomaly detection model, so as to output the anomaly information of the target biological object using the anomaly detection model.

[0120] Exemplarily, the anomaly detection model is a target detection model obtained based on a deep learning algorithm, such as a MASK-RCNN model.

[0121] Specifically, the anomaly detection model can be trained using an image of a display area of ​​known abnormal information of a target biological object, so that the anomaly detection model receives an input image of the display area and outputs abnormal information of the target biological object in the display area; wherein the abnormal information of the target biological object includes abnormal position coordinates, abnormal area image, and abnormal type.

[0122] Based on this, when the target biological object is a diseased part of the human body, the above-mentioned abnormal position coordinates are the precise coordinate positions of the diseased part, the abnormal area image is the segmentation mask of the diseased part, and the abnormal type is the diseased type (for example, cyst, nodule, hyperplasia).

[0123] In a possible implementation, basic image information of an ultrasound image corresponding to at least one target region is fused, and the ultrasound image is classified based on the target biological object corresponding to the fusion result to determine a classification result of the ultrasound image.

[0124] Exemplarily, for each ultrasound image, the fusion of basic image information is achieved by adding label information to the image; wherein all label information added to the image is the fusion result, including at least one corresponding content of the basic image information.

[0125] Correspondingly, the basic image information includes the ultrasound mode of the ultrasound image, the name and position of the target biological object in the ultrasound image, the actual physical distance of a single pixel in the ultrasound image, and abnormal information of the target biological object.

[0126] For example, in a certain ultrasound image, first-level label information including the name and location of the target biological object is added, and then second-level label information including the ultrasound mode is added to achieve the fusion of basic image information of the ultrasound image corresponding to the display area and the identification area.

[0127] Based on the above two-level label information, refer to Figure 3 , it can be determined that the label information added to the ultrasound image (ie, the fusion result of the basic image information) is: left breast, three o'clock, B-type ultrasound mode.

[0128] Exemplarily, after obtaining the fusion result of each ultrasound image, the classification result of each ultrasound image may be obtained according to the target biological object corresponding to each fusion result.

[0129] Specifically, ultrasound images with the same target biological object in the fusion result may be determined as ultrasound images of the same type. Further, the association relationship of ultrasound images of the same type may be determined based on difference information in the fusion results corresponding to the ultrasound images of the same type.

[0130] For example, when there are differences in the coordinates of the target biological object in the fusion results corresponding to a certain type of ultrasound images, it can be determined that the ultrasound images of this type are B-type ultrasound mode images of the same target biological object at different angles; when there are differences in the ultrasound modes in the fusion results corresponding to a certain type of ultrasound images, it can be determined that the ultrasound images of this type are different ultrasound mode images of the same target biological object at the same angle.

[0131] In a possible implementation, based on the classification result of the ultrasound image, the abnormality level of the target biological object represented by the ultrasound image may be determined.

[0132] Exemplarily, for each ultrasound image in any category, the display area in the ultrasound image is obtained, and features of the display area are extracted; then the features of the display area in any category are fused to determine the abnormality level of the target biological object represented by the ultrasound image based on the feature fusion result.

[0133] Specifically, the images of the display areas of all ultrasound images of the same category may be input into the grade division model, so as to use the grade division model to output the abnormality grade of the target biological object corresponding to the ultrasound images of the category.

[0134] Exemplarily, the hierarchical classification model is a classification model composed of a classification module and a feature fusion module obtained based on a deep learning algorithm, wherein the architecture of the classification module is, for example, ResNet18, MobileNet v3, ViT, and other architectures.

[0135] Correspondingly, after the images of the display areas of all ultrasound images of the same category are input into the grade division model, the classification module in the grade division model can be used to determine the feature information of each display area image, and then the feature fusion module in the grade division model can be used to fuse the feature information of all display area images determined by the classification module, and the abnormality level of the target biological object can be determined based on the fusion result.

[0136] Specifically, the level classification model can be trained using images of display areas of known abnormality levels of target biological objects and preset abnormality levels; wherein the feature fusion module in the level classification model can be represented in the form of feature averaging or self-attention.

[0137] For example, when the target biological object is a single breast, the preset abnormality level used to train the grading model is the BI-RADS level, based on which the grading model can be called a BI-RADS classification model.

[0138] In another embodiment of the present application, a specific method for determining the quality assessment result of the ultrasound image is also provided.

[0139] In a possible implementation, the quality assessment result of the ultrasound image may be determined based on the comparison result between the first quality control standard and the characteristic information.

[0140] Exemplarily, the characteristic information of the ultrasound image may be constituted by the basic image information of the ultrasound image obtained in the previous embodiment, the classification result of the ultrasound image, and the abnormality level of the target biological object.

[0141] Exemplarily, the first quality control standard may include at least one control standard for the position of the target biological object in the ultrasound image, the number of ultrasound images corresponding to the target biological object, the identification area of ​​the ultrasound image, and the ultrasound mode of the ultrasound image.

[0142] Exemplarily, the control criterion for the position of the target biological object in the ultrasound image is: the abnormal position of the target biological object is located at the center of the ultrasound image; correspondingly, it can be determined whether the ultrasound image meets this control criterion based on the abnormal area image of the target biological object determined by the display area (for example, the segmentation mask of the lesion site).

[0143] Specifically, based on the distance between the abnormal region of the target biological object and the border of the ultrasonic image, it is determined whether the ultrasonic image meets this control standard. For example, when the distance between the abnormal region of the target biological object and the left and right borders of the ultrasonic image does not exceed 10% of the width of the ultrasonic image, and the distance between the abnormal region of the target biological object and the upper and lower borders of the ultrasonic image does not exceed 10% of the height of the ultrasonic image, it is determined that the ultrasonic image meets this control standard.

[0144] The control standard for the number of ultrasound images corresponding to the target biological object is: target biological objects with different abnormality levels correspond to different numbers of images of abnormal regions of the target biological object;

[0145] Correspondingly, whether the ultrasound image meets this control standard can be determined based on the abnormality level of the target biological object determined by fusing the basic image information of the ultrasound image, and the abnormal area image of the target biological object determined by the display area (for example, the segmentation mask of the lesion site).

[0146] For example, when the target biological object is a single breast, target biological objects with abnormality levels BI-RADS4a and above correspond to at least two ultrasound images, and target biological objects with abnormality levels BI-RADS3 and below correspond to at least one ultrasound image.

[0147] The control criterion for the identification area of ​​the ultrasound image is: whether at least one of a text identification area and an image identification area exists in the ultrasound image.

[0148] The control standard for the ultrasound mode of the ultrasound image is: an ultrasound image whose ultrasound mode is a color Doppler mode is required to be included, and the blood flow sampling frame of the image completely includes the abnormal area of ​​the target biological object.

[0149] Correspondingly, whether the ultrasound image satisfies this control criterion may be determined based on the ultrasound mode of the ultrasound image determined by the display area and the abnormal position coordinates / abnormal region image of the target biological object.

[0150] In another embodiment of the present application, a specific method for determining the quality evaluation result of text data is also provided.

[0151] In one possible implementation, Figure 4 is a flow chart of determining the quality evaluation result of text data provided by the embodiment of the present application, such as Figure 4 As shown, the method comprises the following steps:

[0152] Step S401, extracting text content from text data.

[0153] Exemplarily, the text data may be represented in an image form or a file form; wherein the image form may correspond to an electronic image or a camera photo, and the file form may correspond to a PDF format file.

[0154] Correspondingly, when the text data is represented in image form, the OCR text recognition model can be used to extract the text content in the text data.

[0155] Specifically, by inputting text data in the form of an image into the OCR text recognition model, the text content in the image and the coordinates corresponding to each character can be output, and then based on the coordinates corresponding to each character, similar characters can be merged into a complete paragraph, thereby completing the extraction of the text content.

[0156] Correspondingly, when the text data is represented in a file format, the text content in the text data can be extracted using a PDF text extraction program.

[0157] Step S402: Determine the quality evaluation result of the text data using the second quality control standard.

[0158] In a possible implementation, the completeness of key information in the text data and the accuracy of text content in the text data can be determined based on the second quality control standard to determine the quality evaluation result of the text data based on the completeness and accuracy.

[0159] Exemplarily, the second quality control standard includes an example of key information and a judgment standard for the accuracy of the text content. In this case, the large language model can be trained based on the second quality control standard so that when the text content extracted in step S401 is input into the large language model, the large language model can output the integrity detection result of the key information in the text content and the accuracy detection result of the text content. The large language model is a language model based on the GPT form, such as the Lamma2 model.

[0160] Specifically, the text content is input into the large language model, and the large language model can output whether the text content contains certain key information based on the comparison between the key information contained in the text content and the examples of the key information in the second quality control standard; at the same time, the large language model can determine the accuracy of the text content based on the judgment criteria of the text content and the accuracy of the text content in the second quality control standard.

[0161] It should be noted that since the large language model is a language model based on the GPT format, when inputting text content into the large language model, it is also necessary to input pre-set prompt words (for example, specific questions) into the large language model so that the large language model can output integrity detection results and / or correctness detection results.

[0162] Exemplarily, when the target biological object is a diseased part of the human body, examples of key information in the second quality control standard may include: basic information of the patient, the scanning object of the ultrasound equipment, basic information of the scanning doctor, text content of the abnormal area, and conclusive text content.

[0163] Exemplarily, the basic patient information includes the patient's name, gender, age, etc.; the scanning object of the ultrasound device corresponds to the name of the target biological object (for example, bilateral breasts, lymph nodes of bilateral breasts, etc.); the basic information of the scanning doctor corresponds to the signature of the scanning doctor.

[0164] Target biological objects of different abnormal levels correspond to different text contents of the abnormal area, and the text content of the abnormal area includes the position, size, shape, growth direction, edge, echo, rear echo, calcification description, lymph node description, blood flow, etc. of the abnormal area.

[0165] Specifically, when the target biological object is a single breast, the text content of the abnormal area with abnormality level BI-RADS 3 or above includes the location (for example, (left and right, clock direction, distance from the nipple)), size, shape, growth direction, edge, echo, posterior echo, calcification description, lymph node description, and blood flow of the abnormal area; the text content of the abnormal area with abnormality level BI-RADS 2 or below includes the location, size, shape, echo, and blood flow of the abnormal area.

[0166] The conclusive text content includes the number and location of abnormal areas and the corresponding relationship with the preset abnormal level. For example, when the target biological object is a single breast, the conclusive text content corresponds to the number and location of the lesion site and the corresponding relationship with the abnormal level BI-RADS.

[0167] Exemplary criteria for determining the accuracy of text content may include: (1) whether the text content describing the same abnormal area (e.g., lesion site) is contradictory or ambiguous; (2) whether the text content describing the same abnormal area (e.g., lesion site) contains typos, missing words, or failure to delete preset templates; (3) whether the text content describing the same abnormal area is consistent with the conclusive text content; (4) whether the conclusive text content contains typos, missing words, or failure to delete preset templates.

[0168] In one possible implementation, the completeness of key information in the text data can be determined based on the integrity detection result of the key information in the text content output by the large language model, and the accuracy of the text content in the text data can be determined based on the accuracy detection result of the text content output by the large language model.

[0169] Exemplarily, the completeness / accuracy detection result of the large language model output can be expressed as a response to a specific question input into the large language model, which response can include an affirmative response, a negative response, and specific information determined based on the affirmative response; wherein the affirmative response and the negative response can be represented by "yes" and "no", respectively.

[0170] Specifically, the questions input into the large language model are, for example: (1) whether the text content includes the key information "basic information of the patient"; (2) whether the text content describing the abnormal area contains typos, missing words, or failure to delete preset templates; (3) whether the text content includes the key information "the size of the abnormal area".

[0171] Correspondingly, the detection results of the output large language model are, for example: (1) "yes" and the patient's name, gender, age, etc., or "no"; (2) "yes" and the specific location where the error occurred, or "no"; (3) "yes" and the location coordinates and size of the abnormal area, for example, the location coordinates are at 3 o'clock on the left breast, and the size is 13x7mm.

[0172] Specifically, the degree of completeness or accuracy can be determined based on the ratio of the number of affirmative answers output by the large language model to the number of questions input to the large language model, that is, the degree of completeness or accuracy is expressed as a specific value a%.

[0173] Based on this, when the quality assessment results of text data are expressed as quality assessment scores and quality assessment text, the completeness of the determined key information and the accuracy of the text content in the text data can be weighted to obtain the quality assessment score of the text data, and the quality assessment text can be determined based on the detection results output by the large language model.

[0174] In another embodiment of the present application, a specific method for determining the matching result between the characteristic information of the ultrasound image and the text data is also provided.

[0175] In a possible implementation, the abnormality type, abnormality position, abnormality size, abnormality level and other information of the target biological object obtained based on the feature information and based on the text data can be compared to obtain a matching result between the feature information and the text data based on the comparison result.

[0176] For example, Figure 5 FIG. 1 is a flow chart of a method for determining a matching result between feature information and text data provided in an embodiment of the present application. Figure 5 As shown, the method comprises the following steps:

[0177] Step S501, determining a matching result between the abnormal size of the target biological object represented by the feature information and the abnormal size of the target biological object recorded in the text data.

[0178] First, the large language model described in the previous embodiment may be used to determine whether the text content of the text data records the text content of the abnormal area of ​​the target biological object.

[0179] Correspondingly, if it exists, the abnormal type of the abnormal area, the coordinates of the abnormal area, the size of the abnormal area and other information are determined according to the text content of the abnormal area; otherwise, the matching operation is terminated.

[0180] Secondly, based on the coordinates of the abnormal area determined by the text content of the abnormal area, it is determined whether the abnormal type of the abnormal area corresponding to the coordinates in the ultrasound image is consistent with the abnormal type of the abnormal area determined according to the text content of the abnormal area.

[0181] If the abnormality types of the abnormal regions are consistent, the major axis and minor axis (the size of the abnormal region, i.e., the abnormal size of the corresponding target biological object) of the abnormal region in the ultrasound image are determined based on the characteristic information; otherwise, the matching operation is terminated.

[0182] Finally, the size of the abnormal area recorded in the text content is compared with the size of the abnormal area determined by the feature information for consistency. If the error between the two is less than 10%, it is determined to be a match, otherwise it is not a match.

[0183] For example, when the major and minor axes of the abnormal area cannot be determined based on the characteristic information of the ultrasound image, a prompt message may be generated: the characteristic information of the ultrasound image is incomplete and the size of the abnormal area cannot be calculated; when the size of the abnormal area cannot be determined based on the text content in the text data, a prompt message may be generated: the text data record is incomplete and the size of the abnormal area is not recorded.

[0184] It should be noted that the matching result includes the above-mentioned matching conclusion and / or prompt information.

[0185] Step S502, determining a matching result between the abnormal information of the target biological object represented by the feature information and the abnormal information of the target biological object recorded in the text data.

[0186] First, the large language model described in the previous embodiment may be used to determine the text content of the abnormal region of the target biological object recorded in the text data.

[0187] For example, the text content includes the location of the abnormal area (three o'clock on the left breast), shape (oval), edge (smooth edges), growth direction (horizontal), interior (anechoic), posterior echo (none), calcification (none), blood flow (none), and lymph nodes (no metastasis).

[0188] Secondly, based on the coordinates of the abnormal area determined by the text content of the abnormal area, it is determined whether the area corresponding to the coordinates in the ultrasound image is the abnormal area.

[0189] If it is an abnormal area, the major axis and the minor axis of the abnormal area in the ultrasound image are determined based on the characteristic information; otherwise, the matching operation is terminated.

[0190] Then, based on the characteristic information of the ultrasound image, the position coordinates, shape, edge information, growth direction, echo, rear echo, calcification, blood flow, lymph nodes and other information of the abnormal area are determined.

[0191] Specifically, the position coordinates of the abnormal area are determined based on the above-mentioned text identification area or image identification area; the shape of the abnormal area is determined based on the intersection-over-union (IOU) of the abnormal area image and the fitted ellipse: if the IOU value is greater than a preset threshold, the shape is determined to be an ellipse, and the shape is further judged to be a circle or an ellipse based on the major and minor axes of the abnormal area; if the IOU value is less than the preset threshold, the shape is determined to be irregular; the corner point of the abnormal area is determined by using an image processing algorithm: if the angle of the corner point is less than a certain threshold, the edge of the abnormal area image is determined to be angled, otherwise the edge of the abnormal area image is smooth; the growth direction of the abnormal area is determined based on the angle between the major axis of the abnormal area and the horizontal line: if the angle is greater than a certain threshold, the growth direction is determined to be a vertical direction, otherwise the growth direction is determined to be a horizontal direction; the echo of the abnormal area is determined based on the distribution of pixel values ​​in the abnormal area image: if the pixel value is greater than a certain threshold, it is determined that there is an echo in the abnormal area, otherwise it is determined that there is no echo.

[0192] Optionally, the abnormal region image is input into the classification model, so that the classification model outputs the position coordinates, shape, edge information, growth direction, echo, rear echo, calcification, blood flow, lymph node and other information of the abnormal region.

[0193] Specifically, the classification model includes multiple classification output terminals corresponding to the number of information of abnormal areas that need to be determined, and each classification output terminal can output corresponding abnormal area information; for example, the classification model is a classification model obtained based on a deep learning algorithm, such as ResNet18, MobileNet v3, Vision Transformer (ViT) and other models.

[0194] Finally, the information of the abnormal area recorded in the text content is matched with the information of the abnormal area determined by the feature information, and the information that is inconsistent with the match is recorded.

[0195] Exemplarily, when a certain abnormal area cannot be determined based on the characteristic information of the ultrasound image, a prompt message may be generated: the characteristic information of the ultrasound image is incomplete and the abnormal area information cannot be determined; when the abnormal area information cannot be determined based on the text content in the text data, a prompt message may be generated: the text data record is incomplete and the abnormal area information is not recorded.

[0196] It should be noted that the matching result includes the abnormal area information and / or prompt information of the above-mentioned matching inconsistency.

[0197] Step S503, determining a matching result between the abnormality level of the target biological object represented by the feature information and the abnormality level of the target biological object recorded in the text data.

[0198] First, the abnormality level of the target biological object in the ultrasound image may be determined based on feature information of the ultrasound image.

[0199] If the abnormality of the target biological object is not determined based on the characteristic information of the ultrasound image, the abnormality level of the target biological object is determined to be the lowest abnormality level.

[0200] Secondly, the abnormality level of the target biological object recorded in the text content is compared with the abnormality level of the target biological object determined by the feature information. If they are consistent, it is determined that the two match, otherwise an abnormal prompt message is generated.

[0201] It should be noted that the matching result includes the above-mentioned matching conclusion and / or abnormal prompt information.

[0202] In another embodiment of the present application, a specific method for generating a quality control data preview page is also provided.

[0203] Exemplarily, a quality control data preview page may be generated based on the quality assessment results of the ultrasound image, the quality assessment results of the text data, and the matching results of the feature information and the text data.

[0204] For example, Figure 6 is a schematic diagram of a quality control data preview page provided in an embodiment of the present application, such as Figure 6 As shown, the quality control data preview page is a breast ultrasound quality control report, that is, the target biological object is the breast; specifically, the quality assessment results of the ultrasound image correspond to the stored image score, the stored image integrity opinion, and the image quality opinion; the quality assessment results of the text data correspond to the report score, the report integrity opinion, and the report accuracy opinion; the matching results of the feature information and the text data correspond to the image-text consistency score and the image-text consistency opinion.

[0205] In addition, the quality control data preview page also includes the original ultrasound image, original text data and basic information of the quality control report; among them, the basic information of the quality control report includes the hospital name, examining doctor, quality control doctor, case name and quality control time.

[0206] In another embodiment of the present application, another method for acquiring characteristic information of an ultrasound image is provided.

[0207] For example, Figure 7 FIG. 1 is a flow chart of a method for acquiring characteristic information of an ultrasound image provided in an embodiment of the present application. Figure 7 As shown, the method comprises the following steps:

[0208] Step S701: using the region division model, determine the display region, the marking region and the proportional region of the ultrasound image.

[0209] Exemplarily, the identification area includes an image identification area and a text identification area.

[0210] In step S7021, the image of the display area is input into a pattern classification model to output an ultrasound pattern of the ultrasound image using the pattern classification model.

[0211] Step S7022: input the image of the marked area into the body landmark recognition module to determine the name and coordinates of the target biological object in the ultrasound image.

[0212] Exemplarily, the body landmark recognition module is configured with an image retrieval method or a multi-task classification model.

[0213] Step S7023, input the image of the proportional area into a spacing recognition module to determine the actual physical distance of a single pixel in the ultrasound image.

[0214] Exemplarily, the spacing recognition module is configured with an image retrieval method or a regression model.

[0215] Step S7031, determining a correspondence between the ultrasound image and the target biological object based on the ultrasound mode of the ultrasound image and the name and coordinates of the target biological object in the ultrasound image.

[0216] Step S7032, determining the abnormal region image based on the identified region and the name and coordinates of the target biological object in the ultrasound image.

[0217] Step S704: determining the abnormality level of the target biological object based on the correspondence between the ultrasonic image and the target biological object.

[0218] In another embodiment of the present application, a quality control processing device for ultrasonic image and text data is also provided. Figure 8 Schematic diagram of a block diagram of a quality control processing device for ultrasound image and text data according to an embodiment of the present application. Figure 8 The device includes a first processing unit 801, a second processing unit 802 and a generating unit 803.

[0219] The first processing unit 801 is used to obtain characteristic information of the ultrasound image, and determine the quality evaluation result of the ultrasound image based on the comparison result of the first quality control standard and the characteristic information; wherein the first quality control standard is a preset quality control standard of the ultrasound image;

[0220] The second processing unit 802 is used to determine the quality evaluation result of the text data based on the second quality control standard; wherein the second quality control standard is a preset quality control standard of the text data;

[0221] The generating unit 803 is used to generate a quality control data preview page based on the quality evaluation result of the ultrasound image, the quality evaluation result of the text data, and the matching result of the feature information and the text data.

[0222] Compared with the prior art that uses manual quality control to achieve quality control of ultrasonic image and text data, the quality control processing device for ultrasonic image and text data provided by the embodiment of the present application can, on the one hand, obtain the quality evaluation result of the ultrasonic image by comparing the characteristic information of the ultrasonic image with the preset quality control standard of the ultrasonic image, so as to achieve quality control of the ultrasonic image; the quality evaluation result of the text data can be determined by using the preset quality control standard of the text data, so as to achieve quality control of the text data; based on this, the quality control efficiency of the ultrasonic image and text data can be improved, and the application of the preset quality control standard can avoid the situation that the quality control result has errors due to the strong subjectivity of the quality control standard of the manual quality control method. On the other hand, the consistency comparison of the ultrasonic image and text data can be achieved by matching the characteristic information of the ultrasonic image with the text data, so that the quality control content of the ultrasonic image and text data is more comprehensive, and the quality control data preview page generated based on the characteristic information of the ultrasonic image, the quality evaluation result of the text data and the consistency comparison result of the ultrasonic image and text data can intuitively reflect the quality control data of the ultrasonic image and text data, thereby improving the quality control effect of the ultrasonic image and text data.

[0223] Reference below Fig. 9 , Fig. 9 A schematic diagram of the structure of a computer device suitable for implementing the embodiments of the present application is shown. Fig. 9 As shown, the computer system 900 includes a central processing unit (CPU) 901, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 902 or the program loaded from the storage part 909 to the random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation instructions of the system are also stored. The CPU 901, the ROM 902, and the RAM 903 are connected to each other through a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0224] The following components are connected to the I / O interface 905: an input section 906 including a keyboard, a mouse, etc.; an output section 907 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 908 including a hard disk, etc.; and a communication section 909 including a network interface card such as a LAN card, a modem, etc. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the I / O interface 905 as needed. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 910 as needed so that a computer program read therefrom is installed into the storage section 908 as needed.

[0225] In particular, according to an embodiment of the present application, the above reference flow chart Figure 2The described process can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 909, and / or installed from a removable medium 911. When the computer program is executed by the central processing unit (CPU) 901, the above-mentioned functions defined in the system of the present application are executed.

[0226] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-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 of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer 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 of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0227] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operating instructions of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the aforementioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operating instruction, or can be implemented with a combination of dedicated hardware and computer instructions.

[0228] The units or modules involved in the embodiments described in the present application may be implemented by software or by hardware. The described units or modules may also be provided in a processor, for example, may be described as: a processor including a semantic extraction unit, a weight allocation unit, and a determination unit. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.

[0229] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the computer device described in the above embodiment, or may exist independently without being installed in the computer device. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the method described in the present application. For example, it can be executed Figure 2 The individual steps of the method are shown.

[0230] The present application embodiment provides a computer program product, which includes instructions. When the instructions are executed, the method described in the present application embodiment is executed. For example, Figure 2 The individual steps of the method are shown.

[0231] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) by each other to form a technical solution.

Claims

1. A quality control processing method for ultrasonic graphic data, wherein the ultrasonic graphic data includes ultrasonic images and text data, characterized in that: The method comprises: Acquire characteristic information of the ultrasound image, and determine a quality evaluation result of the ultrasound image based on a comparison result between a first quality control standard and the characteristic information; wherein the first quality control standard is a preset quality control standard of the ultrasound image; Determining a quality evaluation result of the text data based on a second quality control standard; wherein the second quality control standard is a preset quality control standard of the text data; Generate a quality control data preview page based on the quality evaluation result of the ultrasound image, the quality evaluation result of the text data, and the matching result of the feature information and the text data; The acquiring characteristic information of the ultrasound image comprises: Based on preset regional characteristics, acquiring at least one target region of the ultrasound image; For any target area, determining basic image information of an ultrasound image corresponding to the target area based on the regional characteristics of the target area; fusing basic image information of the ultrasound image corresponding to the at least one target area, and classifying the ultrasound image based on the target biological object corresponding to the fusion result to determine a classification result of the ultrasound image; The abnormality level of the target biological object represented by the ultrasonic image is determined based on the classification result of the ultrasonic image, so that the feature information of the ultrasonic image is composed of the basic image information of the ultrasonic image, the classification result of the ultrasonic image and the abnormality level of the target biological object.

2. The method according to claim 1, characterized in that The target area includes at least one of a display area, an image mark area, a text mark area, and a ratio area.

3. The method according to claim 1, characterized in that The determining, based on the classification result of the ultrasound image, the abnormality level of each type of target biological object represented by the ultrasound image comprises: For each ultrasound image in any category, obtaining a display area in the ultrasound image, and performing feature extraction on the display area; The features of the display area in any category are fused, and the abnormality level of the target biological object represented by the ultrasound image is determined based on the feature fusion result.

4. The method according to claim 1, characterized in that: The determining the quality evaluation result of the text data based on the second quality control standard includes: Based on the second quality control standard, determining the completeness of the key information in the text data and the accuracy of the text content in the text data; A quality assessment result of the text data is determined based on the completeness and the accuracy.

5. The method according to claim 1, characterized in that The first quality control standard includes at least one control standard for the position of the target biological object in the ultrasound image, the number of ultrasound images corresponding to the target biological object, the identification area of ​​the ultrasound image, and the ultrasound mode of the ultrasound image.

6. A quality control processing device for ultrasonic graphic data, wherein the ultrasonic graphic data includes ultrasonic images and text data, characterized in that: The device comprises: A first processing unit, configured to obtain characteristic information of the ultrasound image, and determine a quality evaluation result of the ultrasound image based on a comparison result between a first quality control standard and the characteristic information; wherein the first quality control standard is a preset quality control standard of the ultrasound image; A second processing unit, used to determine a quality evaluation result of the text data based on a second quality control standard; wherein the second quality control standard is a preset quality control standard of the text data; A generating unit, configured to generate a quality control data preview page based on the quality evaluation result of the ultrasound image, the quality evaluation result of the text data, and the matching result of the feature information and the text data; The first processing unit is specifically configured to acquire at least one target area of ​​the ultrasound image based on preset area characteristics; For any target area, determining basic image information of an ultrasound image corresponding to the target area based on the regional characteristics of the target area; fusing basic image information of the ultrasound image corresponding to the at least one target area, and classifying the ultrasound image based on the target biological object corresponding to the fusion result to determine a classification result of the ultrasound image; The abnormality level of the target biological object represented by the ultrasonic image is determined based on the classification result of the ultrasonic image, so that the feature information of the ultrasonic image is composed of the basic image information of the ultrasonic image, the classification result of the ultrasonic image and the abnormality level of the target biological object.

7. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images