Early pregnancy embryo ultrasonic maximum measurement section automatic grabbing system based on real-time dynamic image

Abstract

The invention provides an early pregnancy embryo ultrasonic maximum measurement section automatic grabbing system based on a real-time dynamic image, and the system comprises a multi-classification semantic segmentation module which is used for recognizing and segmenting a gestation sac, a yolk sac and a germ in an early pregnancy embryo ultrasonic dynamic image, and outputting a segmentation result; the measurement section capturing module is used for dynamically maintaining the longest diameter table of each category in the segmentation result based on the segmentation result, and determining the maximum measurement section of each category in the segmentation result according to a dynamic updating strategy; and the automatic measurement module is used for automatically measuring the maximum measurement tangent plane and generating a measurement result. The real-time performance and the stability of the system are improved, the measurement speed is remarkably improved, the consistency and the accuracy of diagnosis results are improved, and measurement errors caused by experience differences of doctors are reduced.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to an automatic capture system for the maximum measurement section of early pregnancy embryo ultrasound based on real-time dynamic images. Background Technology

[0002] Because fetal health directly impacts pregnancy outcomes, monitoring fetal health during this period is crucial. Ultrasound examination has become one of the most important non-invasive screening methods during pregnancy, especially in early pregnancy, used to confirm fetal viability, assess gestational age, locate the gestational sac, observe embryonic anatomy, and screen for chromosomal abnormalities. This provides an opportunity for early risk assessment and identification of potentially high-risk pregnancies.

[0003] However, traditional ultrasound examinations still have shortcomings. First, they are highly dependent on the operator's experience and skills, leading to subjective results and poor diagnostic consistency among different operators. This is especially true in early pregnancy, when the embryonic structure is tiny, making it difficult to accurately measure key biological parameters (such as fetal heart position and heart rate) and susceptible to operational errors. Second, the processing and analysis of ultrasound images mainly rely on manual methods, which are slow and error-prone, failing to meet the demands for efficient and accurate clinical procedures. Therefore, introducing automation and intelligent technologies to improve accuracy and efficiency is crucial.

[0004] Artificial intelligence (AI) holds immense potential in ultrasound applications. While some progress has been made in automated ultrasound diagnostic methods based on AI and deep learning, their application in early pregnancy remains limited. The ability to automatically identify and measure early embryos remains weak. Existing intelligent tools lack comprehensive solutions for the precise assessment of early pregnancy embryos, and effective clinical applications have not yet been developed.

[0005] In summary, the existing technology has the following drawbacks: (1) Traditional manual measurement methods rely on doctors' experience and skills, resulting in low consistency of diagnostic results; (2) Traditional methods require doctors to manually identify the cut surface, select the structure and measure the parameters, which is tedious and time-consuming, and the workload is heavy; (3) It cannot handle large-scale data and is not suitable for large-scale clinical screening scenarios; (4) The positioning of the standard section is done manually by the physician, with a low degree of automation; (5) The examination results are subjective and one-sided, and are prone to misdiagnosis and missed diagnosis. Summary of the Invention

[0006] In view of this, the present invention provides an automatic capture system for the maximum measurement section of early pregnancy embryo ultrasound based on real-time dynamic imaging, so as to at least solve the above-mentioned technical problems.

[0007] According to a first aspect of the present invention, an automatic capture system for the maximum measurement section of an early pregnancy embryo via ultrasound based on real-time dynamic images is provided, comprising: a multi-class semantic segmentation module for identifying and segmenting the gestational sac, yolk sac, and embryo in an early pregnancy ultrasound dynamic image, and outputting segmentation results; a measurement section capture module for dynamically maintaining a table of the longest diameters for each category in the segmentation results based on the segmentation results, and determining the maximum measurement section for each category in the segmentation results according to a dynamic update strategy; and an automatic measurement module for automatically measuring the maximum measurement section and generating measurement results.

[0008] Optionally, the multi-class semantic segmentation module is specifically used to: identify and segment early pregnancy ultrasound dynamic images using a trained three-class semantic segmentation model, and output the segmentation results of the three categories of gestational sac, yolk sac and embryo through three independent channels respectively.

[0009] Optionally, the measurement section capture module includes: a longest diameter table maintenance unit, used to dynamically maintain the longest diameter table for each category in the segmentation result, initially set to zero; a dynamic update unit, used to update the longest diameter table according to a dynamic update strategy; and a maximum measurement section determination unit, used to determine the maximum measurement section for each category in the segmentation result.

[0010] Optionally, the step of dynamically maintaining the longest diameter table for each category in the segmentation results based on the segmentation results, and determining the maximum measurement section for each category in the segmentation results according to the dynamic update strategy, includes: extracting the segmentation contour of each category from the segmentation results; calculating the longest diameter of the segmentation contour of each category; comparing the longest diameter of the current frame with the longest diameter already stored in the longest diameter table; if the longest diameter of the current frame is larger than the longest diameter already stored in the longest diameter table, updating the longest diameter table and recording the corresponding segmentation contour and frame index; after all frames have been processed, determining the maximum measurement section for each category based on the longest diameter in the updated longest diameter table and its corresponding segmentation contour and frame index.

[0011] Optionally, the automatic measurement of the maximum measurement section and the generation of measurement results include: selecting the corresponding segmentation result according to the category of the maximum measurement section; analyzing the segmentation contour of the selected segmentation result and extracting the pixel points of the segmentation contour; calculating the circumscribed straight-line distance of the pixel points as the measurement basis for the longest diameter to determine the longest diameter of the target category; returning the specific measurement point coordinates of the longest diameter of the target category; and generating a measurement result containing the longest diameter data and the measurement point coordinates.

[0012] According to a second aspect of the present invention, an electronic device is provided, including a processor and a memory storing a program. The program includes instructions that, when executed by the processor, cause the processor to perform the steps performed by the method of the first aspect described above.

[0013] According to a third aspect of the present invention, a computer storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method of the first aspect described above.

[0014] Compared with the prior art, the present invention has the following beneficial effects: (1) Stability: The automatic capture strategy of the maximum measurement section of the present invention reduces the computational overhead caused by multi-model processing and improves the real-time performance and stability of the system.

[0015] (2) High efficiency: The automatic measurement process of the present invention does not require manual intervention, which significantly improves the measurement speed and is suitable for large-scale data processing.

[0016] (3) Accuracy: The measurement method based on segmented contours of the present invention can accurately capture the geometric features of the target structure and avoid errors caused by human subjective judgment.

[0017] (4) Versatility: Through a unified measurement process design, the present invention can adapt to different types of structural measurement needs, making multi-task measurement possible. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a structural block diagram of an automatic capture system for the maximum measurement section of early pregnancy embryo ultrasound based on real-time dynamic imaging, according to the present invention.

[0020] Figure 2 This is a network architecture diagram of a three-class semantic segmentation model.

[0021] Figure 3 This is a schematic diagram of the algorithm for determining the maximum measurement section based on a dynamic update strategy.

[0022] Figure 4 This is a schematic diagram showing the measurement of the gestational sac, yolk sac, and embryo based on the segmentation results.

[0023] Figure 5This is a diagram comparing the manual selection and measurement in Case 1 with the automatic capture and measurement by this system.

[0024] Figure 6 This is a diagram comparing the manual selection and measurement in Case 2 with the automatic capture and measurement by this system.

[0025] Figure 7 This is a schematic diagram comparing the manual selection and measurement in Case 3 with the automatic capture and measurement by this system. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] To facilitate understanding, prior art of the present invention will be described by way of example before a detailed description of specific embodiments thereof.

[0028] In recent years, the widespread adoption of image archiving and communication systems, which can systematically store ultrasound images in digital formats, has driven the rapid development of artificial intelligence. Deep learning and machine learning technologies can quantitatively analyze images and extract vast amounts of information, opening up new possibilities for imaging and analysis in early pregnancy. In early pregnancy imaging, AI algorithms can be used for tasks such as segmentation, detection, and classification. Segmentation tasks extract regions of interest from ultrasound images by detecting pixels and automatically identify standard sections or anatomical regions, which not only supports clinicians but also significantly accelerates the diagnostic process. Simultaneously, classification algorithms can analyze input and assign it to appropriate categories to support clinical decision-making.

[0029] In 2012, Ling Zhang et al., based on two-dimensional ultrasound videos of 31 patients, used a coarse-to-fine detection scheme to locate the gestational sac after training two cascaded AdaBoost classifiers. Then, they automatically selected the standard section of the gestational sac by eliminating false positives, combining the relative positional relationships of anatomical structures in the image sequence with local contextual information.

[0030] In 2022, Yu Wang et al., based on 1098 patients, trained a CNN model to segment the gestational sac in early pregnancy two-dimensional ultrasound static images after the embryo showed primitive heart tube pulsation at 6–8 weeks of gestation, and then constructed a predictive model for early pregnancy miscarriage.

[0031] In 2023, Yuanyuan Pei et al. used a specific procedure to automatically segment two-dimensional ultrasound videos of 214 pregnant patients from three centers into 38,941 frames. First, an optimal deep learning classifier was selected to extract standard planes with key anatomical structures from the ultrasound frames. Second, an optimal segmentation model was selected to delineate the gestational sac, and a novel biometric method was used for measurement.

[0032] In 2023, Lijue Liu et al. constructed a modified Unet model (AFG-net) based on 914 transvaginal ultrasound two-dimensional ultrasound images, which can automatically identify and locate the gestational sac, yolk sac and embryo in two-dimensional ultrasound images, and then completely separate and measure them through image segmentation.

[0033] In summary, current intelligent identification of early pregnancy embryos mainly focuses on two-dimensional static ultrasound images and primarily concentrates on the analysis of individual structures (mostly the gestational sac). Research based on dynamic video is relatively recent (2012, Ling Zhang) and either only analyzes the gestational sac structure. Furthermore, early pregnancy ultrasound images often suffer from image noise, low contrast, and dark areas, increasing the difficulty of measurement. While automated ultrasound diagnostic methods based on artificial intelligence and deep learning have made some progress, their application in early pregnancy remains limited. The ability to automatically identify and measure early embryos is still relatively weak. Existing intelligent tools lack comprehensive solutions for the detailed assessment of early pregnancy embryos and have not yet been developed into effective clinical applications.

[0034] To address these shortcomings, this invention aims to develop a deep learning-based intelligent image capture system for early pregnancy ultrasound examinations. The focus is on automatically identifying standard sections, which can then be used for further automated measurement of embryonic biological parameters. Specific objectives include: constructing an intelligent identification system for standard sections; utilizing deep learning algorithms to automatically identify standard sections in early pregnancy ultrasound images, thereby reducing operator reliance on image recognition and improving the standardization of the examination. This research is expected to effectively shorten the early pregnancy ultrasound assessment time, significantly improve physician efficiency, and effectively enhance the accuracy and consistency of ultrasound examinations. This will provide strong support for early screening, diagnosis, and treatment, thereby improving the quality of the newborn population and ensuring maternal and infant health.

[0035] Specifically, see Figure 1 The present invention provides an automatic capture system for the maximum measurement section of early pregnancy embryo ultrasound based on real-time dynamic imaging, comprising: The multi-class semantic segmentation module 10 is used to identify and segment the gestational sac, yolk sac and embryo in early pregnancy ultrasound dynamic images and output the segmentation results. The measurement section capture module 20 is used to dynamically maintain the longest diameter table of each category in the segmentation results based on the segmentation results, and determine the maximum measurement section of each category in the segmentation results according to the dynamic update strategy. The automatic measurement module 30 is used to automatically measure the maximum measurement section and generate measurement results.

[0036] Optionally, the multi-class semantic segmentation module is specifically used to: identify and segment early pregnancy ultrasound dynamic images using a trained three-class semantic segmentation model, and output the segmentation results of the three categories of gestational sac, yolk sac and embryo through three independent channels respectively.

[0037] This invention employs multi-class semantic segmentation technology, utilizing a single three-class semantic segmentation model to simultaneously identify the gestational sac, yolk sac, and embryo, thereby simplifying the system structure and improving efficiency.

[0038] Optionally, the measurement section capture module includes: a longest diameter table maintenance unit, used to dynamically maintain the longest diameter table for each category in the segmentation result, initially set to zero; a dynamic update unit, used to update the longest diameter table according to a dynamic update strategy; and a maximum measurement section determination unit, used to determine the maximum measurement section for each category in the segmentation result.

[0039] Optionally, the step of dynamically maintaining the longest diameter table for each category in the segmentation results based on the segmentation results, and determining the maximum measurement section for each category in the segmentation results according to a dynamic update strategy, includes: extracting the segmentation contour of each category from the segmentation results; calculating the longest diameter of the segmentation contour of each category; comparing the longest diameter of the current frame with the longest diameter already stored in the longest diameter table; if the longest diameter of the current frame is larger than the longest diameter already stored in the longest diameter table, updating the longest diameter table and recording the corresponding segmentation contour and frame index; and after all frames have been processed, determining the maximum measurement section for each category based on the longest diameter in the updated longest diameter table and its corresponding segmentation contour and frame index.

[0040] Optionally, the automatic measurement of the maximum measurement section and the generation of measurement results include: selecting the corresponding segmentation result according to the category of the maximum measurement section; analyzing the segmentation contour of the selected segmentation result and extracting the pixel points of the segmentation contour; calculating the circumscribed straight-line distance of the pixel points as the measurement basis for the longest diameter to determine the longest diameter of the target category; returning the specific measurement point coordinates of the longest diameter of the target category; and generating a measurement result containing the longest diameter data and the measurement point coordinates.

[0041] This invention significantly improves the consistency and accuracy of diagnostic results, provides a reliable auxiliary diagnostic tool, and reduces measurement errors caused by differences in doctors' experience.

[0042] Specific implementation steps 1) such as Figure 2 As shown, the multi-class semantic segmentation module of this invention outputs the segmentation results of the gestational sac, yolk sac and embryo through independent channels.

[0043] 2) After obtaining the segmentation results, this invention designs a measurement-driven strategy to send the three-channel segmentation results into the measurement process. The measurement process is responsible for maintaining the longest path table for each category and selecting the optimal measurement cross section through dynamic updates. Specifically, when the length of a cross section for a certain category in a video frame exceeds the currently recorded longest path, the system automatically replaces the previous record with a new cross section, thereby achieving the capture of the optimal cross section.

[0044] 3) After automatically capturing the largest measurement section, the system will select the corresponding segmentation result based on the category of the section (gestational sac, yolk sac, or embryo) and send it to the measurement process for subsequent automatic measurement processing. In simple terms, the measurement process will analyze the segmented contour and calculate the longest diameter of the target structure.

[0045] Specifically, the system of the present invention is further described below: 1. Automatic capture strategy for maximum measurement cross-section In early pregnancy, the anatomical features of the embryo are not yet clearly defined, making traditional scoring rules based on structural visualization inapplicable at this stage. Therefore, this invention proposes a novel automatic maximum measurement section capture strategy to address this challenge. After acquiring the segmentation results, this invention designs a measurement-driven strategy to feed the three-channel segmentation results into the measurement process. The measurement process is responsible for maintaining the longest diameter table for each category and selecting the optimal measurement section through dynamic updates.

[0046] In early pregnancy, the gestational sac, yolk sac, and embryo can coexist in a single image frame. Using separate deep learning models for each of these three sections could significantly reduce the system's real-time performance and reliability. To address this issue, this invention employs a multi-class semantic segmentation technique, attempting to simultaneously identify the gestational sac, yolk sac, and embryo using a single three-class semantic segmentation model, thereby simplifying the system structure and improving efficiency. Multi-class semantic segmentation techniques have been extensively explored in related fields, and this invention builds upon this foundation with model design and optimization.

[0047] Specifically, this three-classification model outputs segmentation results for the gestational sac, yolk sac, and embryo through independent channels. The model network architecture used in this invention is as follows: Figure 2 As shown.

[0048] After training the segmentation model, this invention proposes a dynamic algorithm to capture the maximum measurement plane of the gestational sac, yolk sac, and embryo in early pregnancy ultrasound videos. First, each frame of the video is processed by the segmentation model to generate multi-class segmentation results for the three anatomical structures. These segmentation results are then passed to a measurement process responsible for recording the maximum diameter for each class. For each frame, the algorithm calculates the longest diameter for each segmentation contour. If the calculated diameter for a particular class exceeds the currently recorded value, the algorithm updates the record and stores the corresponding frame and segmentation contour. This iterative process ensures continuous optimization of the recorded values, enabling the algorithm to dynamically capture the optimal measurement plane as the video progresses. By integrating the segmentation and measurement processes, this method provides a robust solution for accurate and efficient embryo parameter analysis in clinical applications. Figure 3 As shown, the process of capturing the maximum measurement cross-section is illustrated.

[0049] 2. Automatic measurement strategy for the maximum measurement cross-section After achieving automatic grasping of the maximum measurement cross section, this invention further investigates how to perform efficient and accurate automatic measurement of the grasped optimal cross section.

[0050] Specifically, when the capture strategy determines that a certain frame of image is the optimal cross-section, the system will select the corresponding segmentation result according to the category of the cross-section (gestational sac, yolk sac or embryo) and send it into the measurement process for subsequent processing.

[0051] The measurement process begins with analyzing the segmented contour and calculating the longest diameter of the target structure, such as... Figure 4 As shown. To ensure measurement accuracy, this invention employs a geometry-based analysis method, extracting the pixels of the segmented contour and calculating the distance between their circumscribed lines as the measurement basis for the longest diameter. After completing the length calculation, the system returns the specific coordinates of the measurement point of the longest diameter, providing support for subsequent evaluation and visualization.

[0052] See Figure 5-7 Characterization data and effect data of the products used in the examples and comparative examples. Figure 5 The images from top to bottom are the maximum measurement cross-sections of the embryo, yolk sac, and gestational sac. The left column shows manually selected and measured images, while the right column shows images automatically captured and measured using this system. Figure 6 The images shown from top to bottom are the maximum measurement cross-sections of the gestational sac, embryo, and yolk sac. The left column shows manually selected and measured images, while the right column shows images automatically captured and measured using this system. Figure 7 The images shown from top to bottom are the maximum measurement cross-sections of the gestational sac, embryo, and yolk sac. The left column shows manually selected and measured images, while the right column shows images automatically captured and measured using this system.

[0053] The system of the present invention has the following significant advantages: 1. Regarding the automatic capture strategy for the maximum measurement section: It reduces the computational overhead caused by multi-model processing, improving the real-time performance and stability of the system; it dynamically adjusts the capture strategy through a unified measurement process, making the selection of measurement sections more accurate; it effectively solves the difficulty of section capture caused by the lack of obvious structural features in early pregnancy, laying the foundation for automatic measurement; it can simultaneously predict the segmentation results of decoded features and hand over the decoded feature vectors to a multilayer perceptron to predict the probability of having this structure in the frame.

[0054] 2. Regarding the automated measurement strategy: The entire process, from section capture to parameter measurement, has been successfully automated. Compared to traditional manual measurement, the advantages include: 1) High efficiency: The automatic measurement process requires no manual intervention, which significantly improves the measurement speed and is suitable for large-scale data processing; 2) Accuracy: The measurement method based on segmented contours can accurately capture the geometric features of the target structure, avoiding errors caused by subjective human judgment; 3) Versatility: Through a unified measurement process design, this scheme can adapt to different types of structural measurement needs, making multi-task measurement possible.

[0055] As another example, the present invention also provides an electronic device, which will now be described as an example of a hardware device that can be applied to various aspects of the present invention, serving as a server or client of the invention. The term "electronic device" is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0056] The electronic device may include a processor, a communications interface, memory, and a communication bus.

[0057] The processor, communication interface, and memory communicate with each other via a communication bus. The communication interface is used to communicate with other electronic devices or servers.

[0058] The processor is used to execute programs, specifically the relevant steps in the above method embodiments.

[0059] Specifically, the program may include program code, which includes computer operation instructions.

[0060] A processor may be a CPU, an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits. A smart device may include one or more processors, which can be of the same type, such as one or more CPUs, or of different types, such as one or more CPUs and one or more ASICs.

[0061] Memory is used to store programs. Memory may include high-speed RAM, and may also include non-volatile memory, such as at least one disk drive.

[0062] When executed by a processor, the program is used to enable electronic devices to perform an automatic capture system for the maximum measurement section of early pregnancy embryos based on real-time dynamic images, according to the present invention.

[0063] Furthermore, the specific implementation of each step in the program can be found in the corresponding descriptions of the steps and units in the above method embodiments, and will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices and modules described above can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0064] The present invention also provides a computer storage medium storing a computer program, wherein when the computer program is executed by a processor, the methods for implementing the various embodiments of the present invention can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0065] The methods described above according to embodiments of the present invention can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code originally stored in a remote recording medium or a non-transitory machine-readable medium and subsequently stored in a local recording medium, downloaded via a network. Thus, the methods described herein can be processed by software stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components (e.g., RAM, ROM, flash memory, etc.) capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods described herein. Furthermore, when a general-purpose computer accesses code used to implement the methods shown herein, the execution of the code transforms the general-purpose computer into a dedicated computer for executing the methods shown herein.

[0066] Specific embodiments of the invention have now been described. Other embodiments are within the scope of the appended claims. In some cases, the actions described in the claims can be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing can be advantageous.

[0067] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the embodiments of the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the embodiments of the present invention. Therefore, all equivalent technical solutions also fall within the scope of the embodiments of the present invention, and the patent protection scope of the embodiments of the present invention should be defined by the claims.

Claims

1. An automatic capture system for the maximum measurement section of early pregnancy embryo ultrasound based on real-time dynamic imaging, characterized in that, include: The multi-class semantic segmentation module is used to identify and segment the gestational sac, yolk sac, and embryo in early pregnancy ultrasound dynamic images and output the segmentation results. The measurement section capture module is used to dynamically maintain the longest diameter table of each category in the segmentation results based on the segmentation results, and determine the maximum measurement section of each category in the segmentation results according to the dynamic update strategy. An automatic measurement module is used to automatically measure the maximum measurement section and generate measurement results.

2. The system according to claim 1, characterized in that, The multi-class semantic segmentation module is specifically used for: The trained three-class semantic segmentation model is used to identify and segment early pregnancy ultrasound dynamic images, and the segmentation results of the gestational sac, yolk sac and embryo are output through three independent channels.

3. The system according to claim 1, characterized in that, The measurement section grasping module includes: The longest path table maintenance unit is used to dynamically maintain the longest path table for each category in the segmentation results, and is initially set to zero. The dynamic update unit is used to update the longest path table according to the dynamic update strategy; The maximum measurement section determination unit is used to determine the maximum measurement section for each category in the segmentation result.

4. The system according to claim 3, characterized in that, Based on the segmentation results, the longest path table for each category in the segmentation results is dynamically maintained, and the maximum measurement section for each category in the segmentation results is determined according to a dynamic update strategy, including: Extract the segmentation contour of each category from the segmentation results; Calculate the longest diameter of the segmentation contour for each category; Compare the longest diameter of the current frame with the longest diameter already stored in the longest diameter table. If the longest diameter of the current frame is larger than the longest diameter already stored in the longest diameter table, update the longest diameter table and record the corresponding segmentation contour and frame index. After all frames have been processed, the maximum measurement section for each category is determined based on the longest diameter in the updated longest diameter table and its corresponding segmentation profile and frame index.

5. The system according to claim 1, characterized in that, The automatic measurement of the maximum measurement section and the generation of measurement results include: Select the corresponding segmentation result based on the category of the largest measured cross section; Analyze the segmentation contour of the selected segmentation result and extract the pixels of the segmentation contour; The circumscribed straight-line distance of the pixel is calculated as the measurement basis for the longest diameter, in order to determine the longest diameter of the target category; Return the coordinates of the specific measurement point for the longest diameter of the target category; Generate measurement results containing the longest diameter data and the coordinates of the measurement points.

6. An electronic device, characterized in that, include: processor; Memory for stored programs; The program includes instructions that, when executed by the processor, cause the processor to perform the steps performed by the system as described in any one of claims 1-5.

7. A computer storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the system as described in any one of claims 1-5.

Images