Breast nodule ultrasound memory and re-identification method, device and system
Through neural network model and image feature extraction method, combined with rectangular box marking and database matching, the problems of low efficiency and missed detection of phacodile breast screening technology are solved, and efficient and accurate breast nodule screening and memory are achieved, which improves users confidence and reliability of screening results.
Patent Information
- Application Number
- CN202311345579.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-18
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2043-10-18
AI Technical Summary
The existing phacopad screening technology is inefficient and prone to missed detection, making it difficult for users to accurately identify and remember the location and number of breast nodules.
The neural network model and image feature extraction method are used to mark and color coding in rectangular box, combined with database preservation and feature matching, and memory and re-identification of breast nodules are achieved to ensure efficient screening and accurate recording.
It improves the efficiency of breast nodule screening, avoids missed examinations, enhances users' clear understanding and confidence in breast screening, and ensures the accuracy of the results.
Smart Images

Figure CN117392509B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of ultrasonic scanning technology, and in particular to a method, device and system for ultrasonic memory and re-identification of breast nodules. Background Art
[0002] Ultrasound is a safe, convenient and low-cost medical imaging technology. In its application, it mainly uses ultrasound probes to transmit and receive sound wave signals to the human body to understand the internal anatomical structure of the human body. For example, in breast screening, ultrasound is more suitable for screening dense breasts, which account for a large proportion in my country, than X-rays.
[0003] In the existing technology, breast screening requires traversing the entire breast tissue. Given that the field of view of ultrasound imaging is relatively limited, ultrasound users need to move the ultrasound probe over a large range, observe a large number of ultrasound images, and search for suspicious breast nodules. Ultrasound real-time detection technology based on artificial intelligence can locate nodules on the current ultrasound image, but because the breast tissue itself lacks obvious anatomical references, it is difficult for ultrasound users to determine the exact position of the current nodule relative to the overall breast tissue. In addition, inexperienced ultrasound users often need to perform repeated ultrasound scans of breast tissue to prevent omissions, and the same nodule may appear multiple times. It is difficult for ultrasound users to reliably identify whether the current nodule is a newly discovered nodule or a previously discovered nodule, as well as how many nodules have been discovered so far. These challenges have caused ultrasound users to lack a clear and complete understanding of breast screening, thereby reducing the confidence and reliability of ultrasound users in conducting breast screening.
[0004] Therefore, how to efficiently use ultrasound to screen breast nodules and avoid missed detection is an urgent problem to be solved. Summary of the invention
[0005] In view of the above-mentioned deficiencies in the prior art, the object of the present invention is to provide a method, device and system for ultrasonic memory and re-identification of breast nodules, aiming to solve the problem of low efficiency and easy missed detection when screening breast nodules with ultrasound.
[0006] The technical solution of the present invention is as follows: The present invention provides a method for ultrasonic memory and re-identification of breast nodules, including: based on a first ultrasonic image obtained by the first scanning of a user, applying a neural network model to determine a first ultrasonic image feature of a first breast nodule and a first rectangular box containing the first breast nodule; marking the first rectangular box as a first color; based on the first breast nodule, applying an image feature extraction method to determine a second ultrasonic image feature of the first breast nodule, and saving it to a database; and marking the first rectangular box as a second color; based on a second ultrasonic image obtained by continuing to scan the user, applying a neural network model to determine a third ultrasonic image feature of a second breast nodule, and a second rectangular box containing the second breast nodule; marking the second rectangular box as the first color; when the third ultrasonic image feature and the ultrasonic image feature in the database are not matched successfully, based on the second breast nodule, applying an image feature extraction method to determine a fourth ultrasonic image feature of the second breast nodule, and saving it to a database; and marking the second rectangular box as a second color.
[0007] In one embodiment of the present invention, the method for ultrasonic memory and re-identification of breast nodules further includes: when the third ultrasonic image feature matches the ultrasonic image feature in the database successfully, marking the second rectangular frame with a second color.
[0008] In one embodiment of the present invention, the dimension of the ultrasound image features in the database is greater than the dimension of the ultrasound image features determined by applying the neural network model.
[0009] In one embodiment of the present invention, the matching of the third ultrasonic image feature and the ultrasonic image feature in the database is unsuccessful, including: obtaining the fifth ultrasonic image feature in the database; the fifth ultrasonic image feature is any ultrasonic image feature in the database; based on the third ultrasonic image feature and the fifth ultrasonic image feature, applying a feature distance calculation method to determine the distance between the third ultrasonic image feature and the fifth ultrasonic image feature; when the distance is greater than a predetermined distance threshold, the matching is unsuccessful.
[0010] In one embodiment of the present invention, the second ultrasound image feature of the first breast nodule is determined and saved in a database, including: when the number N of second ultrasound image features is greater than a preset value, fusing the N second ultrasound image features to obtain a fused ultrasound image feature; and saving the fused ultrasound image feature in the database.
[0011] In one embodiment of the present invention, after marking the second rectangular frame with the second color, the method further includes: adding 1 to the number of breast nodules; and marking the clock position and number of the second breast nodule.
[0012] The present invention provides a breast nodule ultrasonic memory and re-identification device, comprising: a first determination module, used to apply a neural network model to determine a first ultrasonic image feature of a first breast nodule and a first rectangular frame containing the first breast nodule based on a first ultrasonic image obtained by an initial scan of a user; marking the first rectangular frame as a first color; a first processing module, used to apply an image feature extraction method to determine a second ultrasonic image feature of the first breast nodule based on the first breast nodule, and save it to a database; and marking the first rectangular frame as a second color; a second determination module, used to apply a neural network model to determine a third ultrasonic image feature of a second breast nodule and a second rectangular frame containing the second breast nodule based on a second ultrasonic image obtained by a continued scan of the user; marking the second rectangular frame as the first color; a second processing module, used to apply an image feature extraction method to determine a fourth ultrasonic image feature of the second breast nodule based on the second breast nodule, and save it to the database when the third ultrasonic image feature matches the ultrasonic image feature in the database; and marking the second rectangular frame as a second color.
[0013] In one embodiment of the present invention, the second processing module is further configured to: when the third ultrasonic image feature matches the ultrasonic image feature in the database successfully, mark the second rectangular frame with a second color.
[0014] The present invention provides a storage medium storing program instructions, wherein when the program instructions are executed, the steps of a method for ultrasonic memory and re-identification of breast nodules are implemented.
[0015] The present invention provides an ultrasonic processing system, comprising: a nodule positioning module, a feature extraction module, a feature matching module, a nodule memory module, a nodule re-identification module, a visualization module, and a human-computer interaction module; the nodule positioning module is respectively connected to the feature extraction module, the nodule memory module, and the nodule re-identification module, and is used to apply a neural network model to determine a first ultrasonic image feature of a first breast nodule and a rectangular frame containing the first breast nodule based on an ultrasonic image obtained by scanning a user, and mark the rectangular frame as a first color; the feature extraction module is respectively connected to the nodule memory module, the nodule re-identification module, and the feature matching module, and is used to apply an image feature extraction method to the first breast nodule to determine a second ultrasonic image feature of the first breast nodule and save it to a database; the number of dimensions of the second ultrasonic image feature is greater than that of the first ultrasonic image feature; the feature matching module is respectively connected to the nodule memory module and the nodule re-identification module, and is used to apply a feature distance calculation method to paired ultrasonic image features to determine paired The distance between the ultrasonic image features; when the distance is greater than a predetermined distance threshold, the match is unsuccessful; a nodule re-identification module is used to receive the third ultrasonic image feature of the second breast nodule determined by the nodule positioning module when continuing the ultrasonic scanning of the user; the third ultrasonic image feature is transmitted to the feature matching module; the nodule re-identification module is also used to receive the matching result of the third ultrasonic image feature determined by the feature matching module and the ultrasonic image feature in the database; when the matching result is that there is a successful match, the number of the breast nodule corresponding to the ultrasonic image feature in the database is returned; when the matching result is that none of them are successful, the nodule memory module is activated; the nodule re-identification module is also connected to the visualization module for real-time display of the number, clock position and quantity of the breast nodules; the nodule memory module is respectively connected to the visualization module and the human-computer interaction module for receiving the fourth ultrasonic image feature of the second breast nodule determined by the feature extraction module; and according to the operation of the human-computer interaction module, it is transmitted to the visualization module in real time for display.
[0016] As described above, the breast nodule ultrasound memory and re-identification method, device and system of the present invention have the following beneficial effects:
[0017] (1) It can efficiently complete the ultrasound screening of breast nodules.
[0018] (2) It can avoid missed detection and ensure the accuracy of breast screening results.
[0019] (3) Ensure that users have a clear and complete understanding of breast screening and improve their confidence and reliability in conducting breast screening. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 Schematic diagram of the flow of the method for ultrasonic memory and re-identification of breast nodules in one embodiment of the present invention.
[0021] Figure 2Schematic diagram of the clock position of the method for ultrasonic memory and re-identification of breast nodules in one embodiment of the present invention.
[0022] Figure 3 It is a schematic diagram of the structure of the breast nodule ultrasonic memory and re-identification device in one embodiment of the present invention.
[0023] Figure 4 FIG. 4 is a system schematic diagram of a breast nodule ultrasound memory and re-identification device in one embodiment of the present invention.
[0024] Component number description
[0025] 31 First determination module
[0026] 32 First processing module
[0027] 33 Second determination module
[0028] 34 Second processing module
[0029] 41 Nodule Localization Module
[0030] 42 Feature Extraction Module
[0031] 43 Feature matching module
[0032] 44 Nodule Memory Module
[0033] 45 Nodule Re-identification Module
[0034] 46 Visualization Module
[0035] 47 Human-computer interaction module DETAILED DESCRIPTION
[0036] The present invention provides a method, device and system for ultrasonic memory and re-identification of breast nodules. To make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0037] The breast nodule ultrasonic memory and re-identification method, device and system provided by the present invention can efficiently complete the ultrasonic screening of breast nodules; avoid missed detection and ensure the accuracy of breast screening results; ensure that users have a clear and complete understanding of breast screening, and improve the confidence and reliability of conducting breast screening.
[0038] like Figure 1 As shown, in one embodiment, the breast nodule ultrasound memory and re-identification method of the present invention includes the following steps:
[0039] Step S1: Based on a first ultrasound image obtained by first scanning the user, a neural network model is applied to determine a first ultrasound image feature of a first breast nodule and a first rectangular frame containing the first breast nodule; and the first rectangular frame is marked with a first color.
[0040] Specifically, a breast ultrasound scan is started for the user, and ultrasound images are collected in real time. When a breast nodule, i.e., a first breast nodule, appears on the ultrasound image for the first time, a first ultrasound image containing the first breast nodule is obtained, and a neural network model is applied to determine the first ultrasound image features of the first breast nodule, and a first rectangular box containing the first breast nodule is marked, and the first rectangular box is set to a first color, for example, the first color is red, indicating that the ultrasound system has not memorized the first breast nodule.
[0041] Step S2: Based on the first breast nodule, applying an image feature extraction method, determining a second ultrasound image feature of the first breast nodule, and saving the feature to a database; and marking the first rectangular frame with a second color.
[0042] Specifically, the nodule memory program is started for the first breast nodule, that is, the image feature extraction method is applied to the first breast nodule to determine the extraction of high-dimensional ultrasound image features, namely, the second ultrasound image features. The second ultrasound image features here are multiple features obtained by the ultrasound system through a full range of ultrasound scanning of the first breast nodule. Through different anatomical sections, the recognition of the first breast nodule becomes more complete and accurate.
[0043] Further specifically, the number of dimensions of the ultrasound image features in the database is greater than the number of dimensions of the ultrasound image features determined by applying the neural network model.
[0044] More specifically, when the number N of the extracted second ultrasonic image features is greater than a preset value, the N second ultrasonic image features are fused to obtain fused ultrasonic image features; and the fused ultrasonic image features are saved in a database.
[0045] More specifically, marking the first rectangular frame with a second color, for example, green, indicates that the ultrasound system has completed a comprehensive scan and memory of the first breast nodule, and can also mark the clock position and number of the first breast nodule.
[0046] Step S3: Based on the second ultrasound image acquired by continuing to scan the user, the neural network model is applied to determine the third ultrasound image features of the second breast nodule and the second rectangular box containing the second breast nodule; and the second rectangular box is marked with the first color.
[0047] Specifically, the user continues to perform breast ultrasound scanning and collects ultrasound images in real time. When a breast nodule appears again on the ultrasound image, such as a second breast nodule, the neural network model is applied to determine the third ultrasound image features of the second breast nodule, mark the second rectangular box containing the second breast nodule, and set the second rectangular box to a first color, for example, red, indicating that the ultrasound system has not memorized the second breast nodule.
[0048] Step S4: When the third ultrasonic image feature and the ultrasonic image feature in the database are not matched successfully, based on the second breast nodule, an image feature extraction method is applied to determine the fourth ultrasonic image feature of the second breast nodule and save it to the database; and the second rectangular box is marked with a second color.
[0049] Specifically, the fifth ultrasound image feature in the database is obtained; the fifth ultrasound image feature is any ultrasound image feature in the database; based on the third ultrasound image feature and the fifth ultrasound image feature, the characteristic distance calculation method is applied to determine the distance between the third ultrasound image feature and the fifth ultrasound image feature; when the distance is greater than the predetermined distance threshold, the matching is unsuccessful. That is, the ultrasound image features in the database and the third ultrasound image features are sequentially calculated by the characteristic distance calculation method, and the distance is calculated one by one. When the results are all greater than the predetermined distance threshold, the matching is unsuccessful. The unsuccessful matching indicates that the ultrasound system has not memorized the second breast nodule, and the nodule memory program needs to be started for the second breast nodule.
[0050] More specifically, after marking the second rectangular box as the second color, step S4 also includes: adding 1 to the number of breast nodules; and marking the clock position and number of the second breast nodule. That is, each time the second rectangular box is marked as the second color, such as green, it means that the ultrasound system has completed a comprehensive scan and memory of the corresponding breast nodule. The clock position and number of the corresponding breast nodule can be marked at the same time; and the number of breast nodules that have been identified and memorized is added by 1, which is convenient for later statistics.
[0051] In one embodiment, the breast nodule ultrasound memory and re-identification method of the present invention may further include: step S5, when the third ultrasound image feature and the ultrasound image feature in the database match successfully, marking the second rectangular box with a second color. That is, the ultrasound system has memorized the second breast nodule, and there is no need to start the nodule memory program for the second breast nodule, and the second rectangular box is directly marked with a second color, such as green.
[0052] like Figure 2 As shown, in one embodiment, a schematic diagram of the clock orientation of the present invention, for example, the ultrasound system remembers the clock orientation of the first breast nodule as 2 o'clock, numbered 1; the number of breast nodules is 1; the ultrasound system remembers the clock orientation of the second breast nodule as 5 o'clock, numbered 2; the number of breast nodules is 2.
[0053] By using the method including the above steps S1-S4, ultrasonic screening of breast nodules can be completed efficiently; missed detection can be avoided and the accuracy of breast screening results can be ensured; it can ensure that users have a clear and complete understanding of breast screening and improve the confidence and reliability of breast screening.
[0054] like Figure 3 As shown, in one embodiment, the breast nodule ultrasound memory and re-identification device of the present invention includes:
[0055] A first determination module 31 is used to determine, based on a first ultrasound image obtained by first scanning the user, a first ultrasound image feature of a first breast nodule and a first rectangular frame containing the first breast nodule by applying a neural network model; and mark the first rectangular frame with a first color;
[0056] The first processing module 32 is used to determine the second ultrasound image feature of the first breast nodule based on the first breast nodule by applying an image feature extraction method, and save the feature to a database; and mark the first rectangular frame as a second color;
[0057] A second determination module 33 is used to determine, based on the second ultrasound image acquired by continuing scanning of the user, a third ultrasound image feature of the second breast nodule and a second rectangular frame containing the second breast nodule by applying a neural network model; and marking the second rectangular frame with a first color;
[0058] The second processing module 34 is used to determine the fourth ultrasound image feature of the second breast nodule based on the second breast nodule by applying the image feature extraction method and save it to the database when the third ultrasound image feature and the ultrasound image feature in the database are not matched successfully; and mark the second rectangular box with a second color.
[0059] The second processing module 34 is further configured to: when the third ultrasonic image feature matches the ultrasonic image feature in the database successfully, mark the second rectangular frame with a second color.
[0060] The technical features specifically implemented in the breast nodule ultrasonic memory and re-identification device of this embodiment are basically the same as the principles of each step in the breast nodule ultrasonic memory and re-identification method in Example 1, and the technical contents that are common between the method and the device will not be repeated.
[0061] The storage medium of the present invention stores a computer program, which, when executed by a processor, implements the above-mentioned method for ultrasonic memory and re-identification of breast nodules.
[0062] like Figure 4As shown, in one embodiment, the breast nodule ultrasound memory and re-identification system of the present invention includes: a nodule positioning module 41, a feature extraction module 42, a feature matching module 43, a nodule memory module 44, a nodule re-identification module 45, a visualization module 46, and a human-computer interaction module 47; the nodule positioning module 41 is connected to the feature extraction module 42, the nodule memory module 44, and the nodule re-identification module 45, respectively, and is used to determine the first ultrasound image feature of the first breast nodule based on the ultrasound image obtained by scanning the user, and applying the neural network model, and A rectangular frame containing a first breast nodule is marked with a first color; a feature extraction module 42 is connected to a nodule memory module 44, a nodule re-identification module 45, and a feature matching module 43, respectively, and is used to apply an image feature extraction method to the first breast nodule, determine a second ultrasound image feature of the first breast nodule, and save it to a database; the dimension of the second ultrasound image feature is greater than that of the first ultrasound image feature; the feature matching module 43 is connected to the nodule memory module 44, and the nodule re-identification module 45, respectively, and is used to apply a feature extraction method to paired ultrasound image features. The feature distance calculation method is used to determine the distance between paired ultrasound image features; when the distance is greater than a predetermined distance threshold, the matching is unsuccessful; the nodule re-identification module 45 is used to receive the third ultrasound image feature of the second breast nodule determined by the nodule positioning module 41 when continuing the ultrasound scanning of the user; the third ultrasound image feature is transmitted to the feature matching module 43; the nodule re-identification module 45 is also used to receive the matching result between the third ultrasound image feature determined by the feature matching module 43 and the ultrasound image feature in the database; when the matching result is that there is a successful match, the number of the breast nodule corresponding to the ultrasound image feature in the database is returned; when the matching result is that none of them are successful, the nodule memory module 44 is activated; the nodule re-identification module 45 is also connected to the visualization module 46 for real-time display of the number, clock position and quantity of the breast nodules; the nodule memory module 44 is respectively connected to the visualization module 46 and the human-computer interaction module 47 for receiving the fourth ultrasound image feature of the second breast nodule determined by the feature extraction module 42; and according to the operation of the human-computer interaction module 47, it is transmitted to the visualization module 46 in real time for display.
[0063] In summary, the breast nodule ultrasonic memory and re-identification method, device and system provided by the present invention can efficiently complete the ultrasonic screening of breast nodules; avoid missed detection and ensure the accuracy of breast screening results; ensure that users have a clear and complete understanding of breast screening, and improve the confidence and reliability of breast screening.
[0064] It should be understood that the application of the present invention is not limited to the above examples. For those of ordinary skill in the art, it can be improved or transformed according to the above description, and all these improvements and transformations should belong to the protection scope of the claims attached to the present invention. That is, the above are only embodiments of the present invention, and the patent scope of the present invention is not limited thereto, and the terms "include", "comprise" or any other variants thereof in this article are intended to cover non-exclusive inclusion, so that a process, function or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, function or method. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are similarly included in the patent protection scope of the present invention.
Claims
1. A method for ultrasonic memory and re-identification of breast nodules, characterized in that: include: Based on a first ultrasound image obtained by first scanning the user, applying a neural network model, determining a first ultrasound image feature of a first breast nodule and a first rectangular frame containing the first breast nodule; marking the first rectangular frame with a first color; Based on the first breast nodule, applying an image feature extraction method to determine a second ultrasound image feature of the first breast nodule, and saving the feature to a database; and marking the first rectangular frame as a second color; and determining the second ultrasound image feature of the first breast nodule and saving it to a database, including: When the number N of the second ultrasonic image features is greater than a preset value, fusing the N second ultrasonic image features to obtain a fused ultrasonic image feature; Saving the fused ultrasound image features to the database; Based on a second ultrasound image acquired by continuing to scan the user, applying the neural network model, determining a third ultrasound image feature of a second breast nodule and a second rectangular frame containing the second breast nodule; marking the second rectangular frame with the first color; When the third ultrasound image feature and the ultrasound image feature in the database are not matched successfully, based on the second breast nodule, the image feature extraction method is applied to determine the fourth ultrasound image feature of the second breast nodule, and save it to the database; and the second rectangular frame is marked with the second color; The number of dimensions of the ultrasound image features in the database is greater than the number of dimensions of the ultrasound image features determined by applying the neural network model.
2. The method according to claim 1, characterized in that Also includes: When the third ultrasound image feature matches the ultrasound image feature in the database successfully, the second rectangular frame is marked with the second color.
3. The method according to claim 1, characterized in that: The third ultrasonic image feature and the ultrasonic image feature in the database are not matched successfully, including: Acquire a fifth ultrasonic image feature in the database; the fifth ultrasonic image feature is any ultrasonic image feature in the database; Based on the third ultrasonic image feature and the fifth ultrasonic image feature, applying a feature distance calculation method to determine the distance between the third ultrasonic image feature and the fifth ultrasonic image feature; When the distance is greater than a predetermined distance threshold, the matching is unsuccessful.
4. The method according to claim 1, characterized in that: After marking the second rectangular frame with a second color, the method further includes: Add 1 to the number of breast nodules; and mark the clock position and number of the second breast nodule.
5. A breast nodule ultrasound memory and re-identification device, characterized in that: include: A first determination module is configured to determine, based on a first ultrasound image obtained by first scanning a user, a first ultrasound image feature of a first breast nodule and a first rectangular frame containing the first breast nodule by applying a neural network model; and mark the first rectangular frame as a first color; A first processing module, used for applying an image feature extraction method based on the first breast nodule to determine a second ultrasound image feature of the first breast nodule, and saving the feature to a database; and marking the first rectangular frame as a second color; and determining the second ultrasound image feature of the first breast nodule and saving it to a database, including: When the number N of the second ultrasonic image features is greater than a preset value, fusing the N second ultrasonic image features to obtain a fused ultrasonic image feature; Saving the fused ultrasound image features to the database; A second determination module is configured to apply the neural network model to determine, based on a second ultrasound image acquired by continuing to scan the user, a third ultrasound image feature of a second breast nodule and a second rectangular frame containing the second breast nodule; and mark the second rectangular frame with the first color; a second processing module, for, when the third ultrasonic image feature and the ultrasonic image feature in the database are not matched successfully, applying the image feature extraction method based on the second breast nodule, determining the fourth ultrasonic image feature of the second breast nodule, and saving it to the database; and marking the second rectangular frame with the second color; The number of dimensions of the ultrasound image features in the database is greater than the number of dimensions of the ultrasound image features determined by applying the neural network model.
6. The device according to claim 5, characterized in that The second processing module is also used for: When the third ultrasound image feature matches the ultrasound image feature in the database successfully, the second rectangular frame is marked with the second color.
7. A storage medium storing program instructions, wherein: When the program instructions are executed, the steps of the method for ultrasonic memory and re-identification of breast nodules as described in any one of claims 1-4 are implemented.
8. An ultrasonic processing system, characterized in that: include: Nodule localization module, feature extraction module, feature matching module, nodule memory module, nodule re-identification module, visualization module, and human-computer interaction module; The nodule positioning module is connected to the feature extraction module, the nodule memory module, and the nodule re-identification module, respectively, and is used to determine a first ultrasound image feature of a first breast nodule and a rectangular frame containing the first breast nodule based on an ultrasound image obtained by scanning a user, and to mark the rectangular frame with a first color by applying a neural network model; The feature extraction module is connected to the nodule memory module, the nodule re-identification module, and the feature matching module respectively, and is used to apply an image feature extraction method to the first breast nodule to determine a second ultrasound image feature of the first breast nodule, and save it to a database; the dimension number of the second ultrasound image feature is greater than that of the first ultrasound image feature; The feature matching module is connected to the nodule memory module and the nodule re-identification module respectively, and is used to apply a feature distance calculation method to the paired ultrasound image features to determine the distance between the paired ultrasound image features; when the distance is greater than a predetermined distance threshold, the matching is unsuccessful; The nodule re-identification module is used to receive the third ultrasound image feature of the second breast nodule determined by the nodule positioning module when continuing the ultrasound scanning of the user; and transmit the third ultrasound image feature to the feature matching module; The nodule re-identification module is further used to receive the matching result of the third ultrasound image feature determined by the feature matching module and the ultrasound image feature in the database; when the matching result is that there is a successful match, return the number of the breast nodule corresponding to the ultrasound image feature in the database; when the matching result is that none of them are successful, activate the nodule memory module; The nodule re-identification module is also connected to the visualization module and is used to display the number, clock position and quantity of the breast nodules in real time; The nodule memory module is connected to the visualization module and the human-computer interaction module respectively, and is used to receive the fourth ultrasound image feature of the second breast nodule determined by the feature extraction module; and transmit it to the visualization module for display in real time according to the operation of the human-computer interaction module.
Citation Information
Patent Citations
Image matching method and device based on portrait recognition, storage medium and application
CN110728242A
Real-time lung cancer focus identification system under thoracoscope
CN114532965A
