Ultrasonic antenatal care report automatic generation method and system based on large model and image recognition and medium

Abstract

The invention aims to disclose an ultrasonic antenatal care report automatic generation method and system based on a large model and image recognition and a medium, and relates to the technical field of ultrasonic antenatal care report automatic processing, and the method comprises the following steps: S1, obtaining a plurality of antenatal care report text data; s2, performing incremental pre-training based on natural language processing by adopting a large language model to obtain a medical large model; s3, antenatal care time and fetal month parameters are input into the large medical model, an antenatal care report basic template is obtained, and a plurality of labels to be measured are arranged in the antenatal care report basic template; s4, freezing the ultrasonic image and carrying out screen capture on an area for displaying the measured value; s5, sending the screenshot obtained in the step S4 to a workstation; s6, carrying out OCR (Optical Character Recognition) on the screenshot to obtain a measurement value; s7, the step S4 to the step S6 are repeated until all the labels are replaced by the corresponding measurement values, and the ultrasonic antenatal care report is completed The method has the advantages that only the antenatal care time and the fetal month need to be input in the whole process, and the method is high in efficiency and accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic processing of ultrasonic prenatal examination reports, and in particular to a method, system and medium for automatically generating ultrasonic prenatal examination reports based on large models and image recognition. Background Art

[0002] Ultrasound prenatal examinations require immediate reports. Doctors must operate ultrasound equipment to examine patients, collect ultrasound images and perform measurements, enter the measurement values ​​into the report, and then edit the report. There are many steps, and if you are not careful, you will issue an erroneous report, which will bring medical risks. How to quickly and accurately automatically generate ultrasound prenatal examination reports is one of the technical problems that need to be solved urgently. Summary of the invention

[0003] The purpose of the present invention is to disclose a method, system and medium for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition.

[0004] The first inventive object of the present invention is to provide a method for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition.

[0005] The second inventive object of the present invention is to provide an automatic generation system for ultrasonic prenatal examination reports based on large models and image recognition.

[0006] A third object of the present invention is to provide a computer-readable storage medium.

[0007] To achieve the above first invention objective, the present invention provides a method for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition, comprising the following steps:

[0008] Step S1: Obtaining some prenatal examination report text data;

[0009] Step S2: using a large language model to perform incremental pre-training based on natural language processing on the prenatal examination report text data of step S1 to obtain a large medical model;

[0010] Step S3: inputting the prenatal examination time and fetal month parameters into the medical large model obtained in step S2 to obtain a basic template for a prenatal examination report, wherein the basic template for a prenatal examination report has a plurality of tags to be measured;

[0011] Step S4: Freeze the ultrasound image and measure the position corresponding to each label in step S3, the measured value will be displayed on the ultrasound image, and the area where the measured value is displayed will be captured;

[0012] Step S5: The screenshot obtained in step S4 is sent to the workstation, and the image classification model classifies the screenshot obtained in step S4 so that the screenshot corresponds to the label in step S3;

[0013] Step S6: performing OCR recognition on the screenshot to obtain a measurement value, and replacing the corresponding label with the measurement value;

[0014] Step S7: Repeat steps S4 to S6 until all labels are replaced by corresponding measurement values, and the ultrasound prenatal examination report is completed.

[0015] Preferably, the several tags to be determined include a BPD tag, a HC tag, an AC tag, and a FL tag.

[0016] Preferably, the prenatal examination report text data in step S1 is cleaned, and the BPD data is replaced with a BPD label, the HC data is replaced with a HC label, the AC data is replaced with an AC label, and the FL data is replaced with a FL label.

[0017] Preferably, the medical macromodel generates a basic form for ultrasound prenatal examination reports.

[0018] Preferably, the doctor checks the ultrasound prenatal examination report obtained in step S7 and sends it after confirmation.

[0019] Based on the same inventive principle, in order to achieve the above second inventive object, the present invention provides an automatic generation system of ultrasound prenatal examination report based on large model and image recognition, including an ultrasound examination unit and a PACS workstation;

[0020] The ultrasound examination unit is used to obtain a fetal ultrasound image and to generate a screenshot containing a measurement value, and the screenshot is sent to a PACS workstation;

[0021] The PACS workstation includes a large medical model unit, an OCR recognition unit and an image classification unit;

[0022] The medical large model unit generates a basic template for a prenatal examination report, wherein the basic template for a prenatal examination report has a plurality of tags to be measured;

[0023] The image classification unit classifies the screenshots so that the screenshots correspond to the labels to be measured;

[0024] The OCR recognition unit performs OCR recognition on the screenshot to obtain a measurement value, and the measurement value replaces the corresponding label;

[0025] The system executes the method for automatically generating ultrasonic prenatal examination reports based on large models and image recognition as described in the first invention.

[0026] Preferably, the several tags to be determined include a BPD tag, a HC tag, an AC tag, and a FL tag.

[0027] Preferably, a video signal line is provided between the ultrasonic inspection unit and the PACS workstation.

[0028] Preferably, the screenshot categories include BPD screenshots, HC screenshots, AC screenshots and FL screenshots.

[0029] Based on the same inventive principle, in order to achieve the above-mentioned third invention purpose, the present invention provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, it implements the method for automatically generating ultrasonic prenatal examination reports based on large models and image recognition as described in the first invention.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention first performs incremental training on a number of prenatal examination report text data based on natural language processing to obtain a large medical model, in which a basic template for a prenatal examination report is provided. When performing an ultrasonic prenatal examination, the doctor only needs to input the prenatal examination time and the fetal month to obtain a number of labels to be measured. For each label, the ultrasonic image is frozen, the measurement value is displayed, a screenshot is taken and the screenshot is sent to a workstation, which automatically classifies the screenshot and performs OCR text recognition to obtain the measurement value and replace the corresponding label. The doctor only needs to input the prenatal examination time and the fetal month, and the rest of the parts are automatically completed with high efficiency and accuracy, and the probability of error in the ultrasonic prenatal examination report is greatly reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The present invention is a flow chart of a method for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition;

[0033] Figure 2 The block diagram of the system for automatically generating ultrasonic prenatal examination reports based on large models and image recognition of the present invention;

[0034] Figure 3 It is a schematic diagram of the computer medium module of the present invention. DETAILED DESCRIPTION

[0035] The present invention is described in detail below in conjunction with the various embodiments shown in the accompanying drawings, but it should be noted that these embodiments are not limitations of the present invention, and any equivalent transformations or substitutions in functions, methods, or structures made by ordinary technicians in the field based on these embodiments are all within the scope of protection of the present invention.

[0036] The specific implementation process of the present invention is described below through multiple embodiments.

[0037] Embodiment 1:

[0038] Ginseng Figure 1As shown, this embodiment discloses a specific implementation method of automatically generating an ultrasound prenatal examination report based on a large model and image recognition (hereinafter referred to as the "method").

[0039] Ginseng Figure 1 As shown, in this embodiment, the method for automatically generating an ultrasound prenatal examination report based on a large model and image recognition includes the following steps:

[0040] Step S1: obtaining a number of prenatal examination report text data; specifically, the number of prenatal examination report text data is used as the basic standard data of the medical large model, the prenatal examination report text data in step S1 is cleaned, and the BPD data is replaced with a BPD label, the HC data is replaced with an HC label, the AC data is replaced with an AC label, and the FL data is replaced with an FL label, where BPD represents the biparietal diameter of the fetal head, HC represents the fetal head circumference, AC represents the fetal abdominal circumference, and FL represents the fetal femur length;

[0041] Step S2: using a large language model to perform incremental pre-training based on natural language processing on the prenatal examination report text data of step S1 to obtain a large medical model; the large medical model generates a basic form of an ultrasound prenatal examination report;

[0042] Step S3: Input the prenatal examination time and fetal month parameters into the medical big model obtained in step S2 to obtain a basic template for a prenatal examination report, wherein the basic template for a prenatal examination report has a number of labels to be measured; specifically, input the prenatal examination time and fetal month into the basic form of the ultrasound prenatal examination report in step S2, and the medical big model can generate report findings and diagnoses based on the input prenatal examination time and fetal month to obtain a basic template for a prenatal examination report, wherein the basic template for a prenatal examination report has a number of labels to be measured, such as a BPD label, an HC label, an AC label, and a FL label.

[0043] Step S4: Freeze the ultrasound image and measure the position corresponding to each label in step S3. The measured value will be displayed on the ultrasound image, and a screenshot will be taken of the area displaying the measured value. Specifically, the doctor operates the ultrasound device to measure the position corresponding to each label. The ultrasound image needs to be frozen during measurement. The measured value will be displayed on the ultrasound image, and a screenshot will be taken of the portion containing the measured value.

[0044] Step S5: The screenshot obtained in step S4 is sent to the workstation, and the screenshot obtained in step S4 is classified by the image classification model so that the screenshot corresponds to the label in step S3; specifically, the screenshot is divided into BPD screenshots, HC screenshots, AC screenshots and FL screenshots through the classification model, the BPD screenshot corresponds to the BPD label, the HC screenshot corresponds to the HC label, the AC screenshot corresponds to the AC label, and the FL screenshot corresponds to the FL label.

[0045] Step S6: Perform OCR recognition on the screenshot to obtain a measurement value, and replace the corresponding label with the measurement value; specifically, perform OCR recognition on the BPD screenshot, HC screenshot, AC screenshot and FL screenshot respectively, and replace the corresponding labels with the measurement values, replace the BPD label with the OCR recognition measurement value of the BPD screenshot, replace the HC label with the OCR recognition measurement value of the HC screenshot, replace the AC label with the OCR recognition measurement value of the AC screenshot, and replace the FL label with the OCR recognition measurement value of the FL screenshot.

[0046] Step S7: Repeat steps S4 to S6 until all tags are replaced by corresponding measurement values, and the ultrasound prenatal examination report is completed. Specifically, the doctor checks the ultrasound prenatal examination report obtained in step S7, and sends it after confirmation.

[0047] This embodiment first performs incremental training on a number of prenatal examination report text data based on natural language processing to obtain a large medical model. The large medical model has a basic template for prenatal examination reports. When performing an ultrasonic prenatal examination, the doctor only needs to enter the prenatal examination time and fetal month to obtain a number of labels to be measured. For each label, the ultrasonic image is frozen, the measurement value is displayed, a screenshot is taken, and the screenshot is sent to the workstation. The workstation automatically classifies the screenshot and performs OCR text recognition to obtain the measurement value and replace the corresponding label. The only thing that needs to be entered during the entire process is the prenatal examination time and fetal month. The rest of the parts are automatically completed with high efficiency and accuracy, and the probability of error in the ultrasonic prenatal examination report is greatly reduced.

[0048] Embodiment 2:

[0049] Ginseng Figure 2 As shown, this embodiment discloses a specific implementation of an automatic generation system for ultrasound prenatal examination reports based on large models and image recognition (hereinafter referred to as the "system").

[0050] Ginseng Figure 2As shown, in this embodiment, the automatic generation system of ultrasound prenatal examination report based on large model and image recognition includes an ultrasound examination unit and a PACS workstation, a video signal line is set between the ultrasound examination unit and the PACS workstation, and the video signal line is used to transmit images such as ultrasound screenshots in real time; the ultrasound examination unit is used to obtain fetal ultrasound images, and to generate screenshots containing measurement values ​​by screenshots, and the screenshots are sent to the PACS workstation, and the screenshot categories include BPD screenshots, HC screenshots, AC screenshots and FL screenshots; the PACS workstation includes a medical large model unit, an OCR recognition unit and an image classification unit; the medical large model unit generates a basic template for a prenatal examination report, and the basic template for the prenatal examination report has a number of labels to be measured, a number of labels to be measured The determined labels include BPD labels, HC labels, AC labels and FL labels; the image classification unit classifies the screenshots so that the screenshots correspond to the labels to be measured, and the screenshot categories include BPD screenshots, HC screenshots, AC screenshots and FL screenshots; the OCR recognition unit performs OCR recognition on the screenshots to obtain measurement values, and the measurement values ​​replace the corresponding labels, replacing the BPD label with the OCR recognition measurement value of the BPD screenshot, replacing the HC label with the OCR recognition measurement value of the HC screenshot, replacing the AC label with the OCR recognition measurement value of the AC screenshot, and replacing the FL label with the OCR recognition measurement value of the FL screenshot; the system executes the method for automatically generating ultrasonic prenatal examination reports based on large models and image recognition as described in Example 1.

[0051] The automatic generation system of ultrasound prenatal examination reports based on large models and image recognition disclosed in this embodiment has the same technical solutions as those in Embodiment 1. Please refer to Embodiment 1 and will not be repeated here.

[0052] Embodiment three:

[0053] Ginseng Figure 3 As shown, this embodiment discloses a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in the first embodiment is implemented.

[0054] The computer-readable storage medium disclosed in this embodiment has the same technical solutions as those in Embodiment 1. Please refer to Embodiment 1 and will not be described in detail here.

[0055] The various illustrative logic blocks or units described in the embodiments of the present invention can be implemented or operated by a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of the above. The general-purpose processor can be a microprocessor, and optionally, the general-purpose processor can also be any conventional processor, controller, microcontroller or state machine. The processor can also be implemented by a combination of computing devices, such as a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a digital signal processor core, or any other similar configuration.

[0056] The series of detailed descriptions listed above are only specific descriptions of feasible implementation methods of the present invention. They are not intended to limit the scope of protection of the present invention. Any equivalent implementation methods or changes that do not deviate from the technical spirit of the present invention should be included in the scope of protection of the present invention.

[0057] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0058] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A method for automatically generating ultrasound prenatal examination reports based on large models and image recognition, characterized in that: The following steps are involved: Step S1: Obtaining some prenatal examination report text data; Step S2: using a large language model to perform incremental pre-training based on natural language processing on the prenatal examination report text data of step S1 to obtain a large medical model; Step S3: inputting the prenatal examination time and fetal month parameters into the medical large model obtained in step S2 to obtain a basic template for a prenatal examination report, wherein the basic template for a prenatal examination report has a plurality of tags to be measured; Step S4: Freeze the ultrasound image and measure the position corresponding to each label in step S3, the measured value will be displayed on the ultrasound image, and the area where the measured value is displayed will be captured; Step S5: The screenshot obtained in step S4 is sent to the workstation, and the image classification model classifies the screenshot obtained in step S4 so that the screenshot corresponds to the label in step S3; Step S6: performing OCR recognition on the screenshot to obtain a measurement value, and replacing the corresponding label with the measurement value; Step S7: Repeat steps S4 to S6 until all labels are replaced by corresponding measurement values, and the ultrasound prenatal examination report is completed.

2. The method for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition as claimed in claim 1, characterized in that: Several tags to be assayed include BPD tag, HC tag, AC tag and FL tag.

3. The method for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition as claimed in claim 2, characterized in that: The prenatal examination report text data in step S1 is cleaned, and the BPD data is replaced with a BPD label, the HC data is replaced with an HC label, the AC data is replaced with an AC label, and the FL data is replaced with an FL label.

4. The method for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition as claimed in claim 1, characterized in that: The medical large model generates a basic form for ultrasound prenatal examination reports.

5. The method for automatically generating ultrasonic prenatal examination reports based on large models and image recognition according to claim 1, characterized in that: The doctor checks the ultrasound prenatal examination report obtained in step S7 and sends it after confirmation.

6. The automatic generation system of ultrasound prenatal examination report based on large model and image recognition is characterized by: Includes ultrasound examination unit and PACS workstation; The ultrasound examination unit is used to obtain a fetal ultrasound image and to generate a screenshot containing a measurement value, and the screenshot is sent to a PACS workstation; The PACS workstation includes a large medical model unit, an OCR recognition unit and an image classification unit; The medical large model unit generates a basic template for a prenatal examination report, wherein the basic template for a prenatal examination report has a plurality of tags to be measured; The image classification unit classifies the screenshots so that the screenshots correspond to the labels to be measured; The OCR recognition unit performs OCR recognition on the screenshot to obtain a measurement value, and the measurement value replaces the corresponding label; The system executes the method for automatically generating ultrasonic prenatal examination reports based on large models and image recognition as described in any one of claims 1-5.

7. The automatic generation system of ultrasound prenatal examination report based on large model and image recognition as claimed in claim 6, characterized in that: Several tags to be assayed include BPD tag, HC tag, AC tag and FL tag.

8. The automatic generation system of ultrasound prenatal examination report based on large model and image recognition as claimed in claim 6, characterized in that: A video signal line is arranged between the ultrasonic inspection unit and the PACS workstation.

9. The automatic generation system of ultrasound prenatal examination reports based on large models and image recognition as described in claim 6, wherein the screenshot categories include BPD screenshots, HC screenshots, AC screenshots and FL screenshots.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, the method for automatically generating an ultrasonic prenatal examination report based on a large model and image recognition as described in any one of claims 1 to 5 is implemented.

Images