A method and apparatus for locating a pressure point for hemostasis
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-08-31
- Publication Date
- 2026-03-20
AI Technical Summary
Most users lack basic knowledge of acupressure hemostasis and find it difficult to accurately identify the location of acupressure hemostasis points. Existing textual descriptions of location methods are inaccurate.
A digital pressure hemostasis point detection model was established using the YOLO algorithm. Training images were annotated using annotation software, the model was trained, and adjusted according to the loss function to ensure that the positioning accuracy of different human body parts reached the qualified threshold. A digital pressure hemostasis point positioning device was implemented using a processor and memory.
It enables direct marking of finger pressure hemostasis points on user-provided test samples, resulting in more accurate positioning, improved recognition speed and work efficiency, and reduced errors.
Smart Images
Figure CN117159369B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of data processing, in particular to a finger pressure hemostasis point positioning method and device. BACKGROUND
[0002] In daily life with frequent natural disasters and accidental injuries, trauma to the human body often causes bleeding. In various emergency situations, there is not enough medical resources and medical facilities, and severe bleeding can be life-threatening. Among various types of blood vessel bleeding, arterial bleeding is more dangerous. When an artery bleeds, the blood is bright red and sprays out, and with the beating of the heart, it sprays out in bursts. At this time, the finger pressure hemostasis method (the part most easily hemostatic by pressing the bleeding artery with fingers) is used. As an emergency measure, it can temporarily stop bleeding.
[0003] However, most users do not have the basic knowledge of finger pressure hemostasis, and it is not easy to confirm the accurate position of the finger pressure hemostasis point. The finger pressure hemostasis point positioning method on the market only stays in the description of the text, and the positioning is not accurate. SUMMARY
[0004] Therefore, the purpose of the present application is to provide a finger pressure hemostasis point positioning method and device to overcome the problem of inaccurate positioning of the finger pressure hemostasis point by text description.
[0005] To achieve the above purpose, the present application adopts the following technical solutions:
[0006] On the one hand, the present application provides a finger pressure hemostasis point positioning method, comprising:
[0007] obtaining a detection sample input by a user;
[0008] inputting the detection sample into a pre-created finger pressure hemostasis compression point detection model;
[0009] marking a finger pressure hemostasis point in the detection sample through the finger pressure hemostasis compression point detection model.
[0010] Further, the method described above, before obtaining the detection sample input by the user, further comprises:
[0011] establishing the finger pressure hemostasis compression point detection model based on a YOLO algorithm;
[0012] obtaining a training image, processing the training image to obtain a verification image; wherein the training image comprises a human body part image;
[0013] training the finger pressure hemostasis compression point detection model according to the training image, and verifying the training result according to the verification image;
[0014] determine whether the verification is passed, and if not, retrain the finger pressure hemostasis compression point detection model.
[0015] Further, the method described above, the finger pressure hemostasis compression point detection model is established based on YOLO algorithm, including:
[0016] According to different human body parts, a plurality of finger pressure hemostasis compression point detection models are established based on YOLO algorithm.
[0017] Further, the method described above, the training image is obtained, and the training image is processed to obtain a verification image, including:
[0018] According to the human body parts, the training images are classified;
[0019] The classified training images are processed to obtain verification images.
[0020] Further, the method described above, the classified training images are processed to obtain verification images, including:
[0021] The finger pressure hemostasis compression points are marked in the classified training images by labelme marking software to obtain verification images.
[0022] Further, the method described above, the finger pressure hemostasis compression point detection model is trained according to the training image, and the training result is verified according to the verification image, including:
[0023] The verification images are classified according to human body parts and converted into data conforming to the YOLO algorithm,
[0024] The data and the training images of the same human body part are input into a single finger pressure hemostasis compression point detection model for a preset number of times of training, so as to train the finger pressure hemostasis compression point detection model of different human body parts;
[0025] The training results output by the finger pressure hemostasis compression point detection model of different human body parts are obtained, and the annotation accuracy of the training results is calculated according to the verification image and the loss function.
[0026] Further, the method described above, the determination whether the verification is passed, and if not, retrain the finger pressure hemostasis compression point detection model, including:
[0027] According to a preset qualified accuracy threshold, it is determined whether the finger pressure hemostasis compression point detection model of different human body parts is qualified;
[0028] If the labeling accuracy of the finger pressure hemostasis compression point detection model of the current human body part reaches the qualified accuracy threshold, the finger pressure hemostasis compression point detection model training of the current human body part is qualified.
[0029] If the labeling accuracy of the finger pressure hemostasis compression point detection model of the current human body part does not reach the qualified accuracy threshold, the finger pressure hemostasis compression point detection model of the current human body part is trained again.
[0030] Further, the method described above, the detection sample is input to the pre-created finger pressure hemostasis compression point detection model, comprising:
[0031] Obtain a user instruction, and determine the finger pressure hemostasis compression point detection model of the corresponding human body part according to the user instruction;
[0032] The detection sample is input to the finger pressure hemostasis compression point detection model.
[0033] Further, the method described above, the detection sample at least includes: picture, video and real-time video stream.
[0034] On the other hand, the application also provides a finger pressure hemostasis point positioning device, comprising a processor and a memory, the processor is connected with the memory:
[0035] The processor is used to call and execute the program stored in the memory.
[0036] The memory is used to store the program, and the program is used to execute at least the finger pressure hemostasis point positioning method described above.
[0037] The beneficial effects of the application are:
[0038] The application first obtains the detection sample input by the user, inputs the detection sample into the pre-created finger pressure hemostasis compression point detection model, and marks the finger pressure hemostasis point in the detection sample through the finger pressure hemostasis compression point detection model. In the application, the finger pressure hemostasis point is directly marked on the detection sample provided by the user through the pre-created finger pressure hemostasis compression point detection model, so as to solve the problem of inaccurate positioning of the finger pressure hemostasis point by text description at present. BRIEF DESCRIPTION OF DRAWINGS
[0039] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or the prior art description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.
[0040] Figure 1 is a flow chart provided by an embodiment of the application;
[0041] Figure 2 is a structural schematic diagram provided by an embodiment of the application. DETAILED DESCRIPTION
[0042] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described in detail below. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.
[0043] Most users generally do not have the basic knowledge of finger pressure hemostasis, and it is not easy to confirm the accurate position of the finger pressure hemostasis point. The finger pressure hemostasis point positioning method on the market only stays in the description of the text, and the positioning is not accurate.
[0044] Therefore, the present application aims to provide a finger pressure hemostasis point positioning method and device to overcome the problem of inaccurate positioning of the finger pressure hemostasis point by text description.
[0045] Figure 1 is a flow chart provided by an embodiment of the application. Please refer to Figure 1 , the embodiment can include the following steps:
[0046] S1, obtaining the detection sample input by the user.
[0047] S2, inputting the detection sample into a pre-created finger pressure hemostasis compression point detection model.
[0048] S3, marking the finger pressure hemostasis point in the detection sample by the finger pressure hemostasis compression point detection model.
[0049] It can be understood that the present application first obtains the detection sample input by the user, inputs the detection sample into a pre-created finger pressure hemostasis compression point detection model, and marks the finger pressure hemostasis point in the detection sample by the finger pressure hemostasis compression point detection model. In the present application, the finger pressure hemostasis point is directly marked on the detection sample provided by the user by the pre-created finger pressure hemostasis compression point detection model, which is more accurate than determining the finger pressure hemostasis point by the user's basic common sense and text description, thereby solving the problem of inaccurate positioning of the finger pressure hemostasis point by text description.
[0050] Preferably, before step S1, it further includes:
[0051] establish a finger pressure hemostasis compression point detection model based on a YOLO algorithm;
[0052] obtain training images, and process the training images to obtain verification images; wherein the training images include human body part images;
[0053] train the finger pressure hemostasis compression point detection model according to the training images, and verify the training result according to the verification images;
[0054] determine whether the verification is passed, and if not, train the finger pressure hemostasis compression point detection model again.
[0055] It can be understood that the YOLO algorithm is integrated by a neural network through four stages of generating target candidate regions, extracting target features by a basic network, fusing shallow and deep feature information by a strengthened feature extraction network, and verifying target candidate detection, realizing a new end-to-end target detection algorithm. The finger pressure hemostasis compression point detection model created by the YOLO algorithm positions the finger pressure hemostasis compression point in the image, and the YOLO neural network can recognize 45 pictures per second, has fast positioning speed and high positioning accuracy.
[0056] Preferably, the finger pressure hemostasis compression point detection model established based on the YOLO algorithm comprises:
[0057] According to different human body parts, multiple finger pressure hemostasis compression point detection models are established based on the YOLO algorithm.
[0058] It can be understood that in order to ensure more accurate positioning, multiple finger pressure hemostasis compression point detection models are established according to different parts of the human body.
[0059] Preferably, the training images are obtained, and the training images are processed to obtain verification images, comprising:
[0060] classify the training images according to human body parts;
[0061] process the classified training images to obtain verification images.
[0062] Preferably, the classified training images are processed to obtain verification images, comprising:
[0063] label the finger pressure hemostasis compression points in the classified training images by using a labelme labeling software to obtain the verification images.
[0064] Preferably, the finger pressure hemostasis compression point detection model is trained according to the training images, and the training result is verified according to the verification images, comprising:
[0065] classify the verification images according to human body parts, and convert them into data conforming to the YOLO algorithm,
[0066] inputting the data and the training images of the same human body part into a single finger pressure hemostasis compression point detection model for a preset number of times of training, thereby training the finger pressure hemostasis compression point detection models for different human body parts;
[0067] obtaining a training result output by the finger pressure hemostasis compression point detection model for different human body parts, and calculating a labeling accuracy of the training result according to a verification image and a loss function.
[0068] It can be understood that when training all the created finger pressure hemostasis compression point detection models, in order to ensure accurate training of the finger pressure hemostasis compression point detection models for different human body parts, the data and the training images of the same human body part are inputted into a single finger pressure hemostasis compression point detection model for a preset number of times of training.
[0069] The specific training process is as follows:
[0070] First, a plurality of human body part images are obtained, and the plurality of human body part images are subjected to grayscale processing to obtain grayscale processed images, and a human body part sample database is established according to the grayscale processed images; then, using a labelme labeling software, the human body part images after grayscale processing are subjected to rectangular frame labeling and finger pressure hemostasis compression point labeling to obtain a plurality of marker frame center point coordinates and a plurality of finger pressure hemostasis compression point coordinates, the human body part image data after labeling of the finger pressure hemostasis compression points by the labelme labeling software is converted into data conforming to the YOLO algorithm format, the training images in the sample database and the converted YOLO algorithm data are taken as training inputs and subjected to a preset number of times of training, and the accuracy of recognition and detection is obtained according to the human body part images after labeling of the finger pressure hemostasis compression points and a loss function.
[0071] Preferably, it is judged whether the verification is passed, and if not, the finger pressure hemostasis compression point detection model is trained again, including:
[0072] judging whether the finger pressure hemostasis compression point detection model for different human body parts is qualified according to a preset qualified accuracy threshold;
[0073] if the labeling accuracy of the finger pressure hemostasis compression point detection model for the current human body part reaches the qualified accuracy threshold, the finger pressure hemostasis compression point detection model for the current human body part is qualified;
[0074] if the labeling accuracy of the finger pressure hemostasis compression point detection model for the current human body part does not reach the qualified accuracy threshold, the finger pressure hemostasis compression point detection model for the current human body part is trained again.
[0075] It can be understood that, in order to ensure the positioning accuracy of the finger pressure hemostasis compression point detection model, after the model training is completed, whether the model training is qualified is judged according to a preset qualified accuracy threshold, if the qualified accuracy threshold is not reached, the model is trained again until the qualified accuracy threshold is reached.
[0076] Preferably, step S2 comprises:
[0077] Obtaining a user instruction, and determining a finger pressure hemostasis compression point detection model corresponding to a human body part according to the user instruction;
[0078] Inputting the detection sample into the finger pressure hemostasis compression point detection model.
[0079] Preferably, the detection sample at least comprises a picture, a video and a real-time video stream.
[0080] In specific practice, the detection sample input by the user in the user terminal is obtained, including an image, a video file and an online video. The detection sample is sent to the server terminal; the server terminal analyzes the human body key points in the image by judging the image and the video image frame, locates the finger pressure hemostasis point position in the image, and marks the position by a color point. The image with the marked point or the video synthesized by the image frame is returned to the user terminal; the recognition detection model is generated based on the YOLO algorithm key point detection, and the recognition detection model is used for locating and marking the finger pressure hemostasis point of the detection sample uploaded by the user terminal, which can effectively improve the recognition speed and work efficiency, and achieve the effect of accurate finger pressure hemostasis point positioning and low error.
[0081] The application further provides a finger pressure hemostasis point positioning device for implementing the above method embodiments. Figure 2 is a structural schematic diagram provided by an embodiment of a finger pressure hemostasis point positioning device of the application. As shown in the figure, Figure 2 The finger pressure hemostasis point positioning device of the embodiment comprises a processor 21 and a memory 22, and the processor 21 is connected with the memory 22. The processor 21 is used for calling and executing a program stored in the memory 22; the memory 22 is used for storing the program, and the program is used for at least executing the finger pressure hemostasis point positioning method in the above embodiments.
[0082] The specific implementation of the finger pressure hemostasis point positioning device provided by the embodiment of the application can refer to the implementation mode of the finger pressure hemostasis point positioning method of any of the above embodiments, which will not be described here.
[0083] It can be understood that the same or similar parts in the above embodiments can be mutually referred to, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.
[0084] It should be noted that, in the description of the present application, the terms "first", "second" and the like are used only for descriptive purposes, and cannot be construed as indicating or implying relative importance. In addition, in the description of the present application, the meaning of "a plurality of" is at least two, unless otherwise specified.
[0085] Any process or method descriptions or descriptions of the flow diagrams described herein or otherwise described in the present application can be understood as representing the steps of the codes implemented as one or more modules, segments or portions of codes including executable instructions for performing specific logical functions or steps, and the scope of the preferred embodiments of the present application includes additional implementation involving other processes or methods.
[0086] It should be understood that each part of the present application can be realized by hardware, software, firmware or their combination. In the above embodiments, a plurality of steps or methods can be realized by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if realized by hardware, and as in another embodiment, it can be realized by any one or their combination of the following technologies known in the art: discrete logic circuit with logic gate circuit for implementing logic function on data signal, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA) and the like.
[0087] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment method can be completed by program instructions to the relevant hardware, and the program can be stored in a computer readable storage medium, which includes one or a combination of steps of the method embodiments when executed.
[0088] In addition, each functional unit in each embodiment of the present application can be integrated in one processing module, or each unit can exist physically, or two or more units can be integrated in one module. The above integrated module can be realized in the form of hardware or in the form of software functional module. When the integrated module is realized in the form of software functional module and sold or used as an independent product, it can also be stored in a computer readable storage medium.
[0089] The storage medium mentioned above can be read-only memory, magnetic disk or optical disk, etc.
[0090] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the specification, the illustrative description of the above terms does not necessarily mean the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0091] Although the embodiments of the present application have been shown and described above, it is understood that the above-described embodiments are exemplary, and cannot be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-described embodiments within the scope of the present application.
Claims
1. A method for locating acupressure points for hemostasis, characterized in that, include: A model for detecting pressure points in acupressure hemostasis was established based on the YOLO algorithm. Acquire training images and process the training images to obtain verification images; wherein, the training images include: images of human body parts; The finger pressure hemostasis compression point detection model is trained based on the training images, and the training results are verified based on the verification images; Determine whether the verification is successful. If it is not successful, then train the finger pressure hemostasis compression point detection model again. Obtain the detection sample input by the user; The test sample is input into a pre-created finger pressure hemostasis compression point detection model; The finger pressure hemostasis point detection model is used to mark the finger pressure hemostasis points in the test sample.
2. The method according to claim 1, characterized in that, The method for establishing the finger pressure hemostasis compression point detection model based on the YOLO algorithm includes: Based on different body parts, multiple acupressure hemostasis pressure point detection models were established using the YOLO algorithm.
3. The method according to claim 2, characterized in that, Acquiring training images and processing the training images to obtain verification images includes: The training images are classified according to body parts; The classified training images are processed to obtain verification images.
4. The method according to claim 3, characterized in that, The step of processing the classified training images to obtain verification images includes: The labelme annotation software is used to annotate the pressure points for acupressure hemostasis in the classified training images to obtain verification images.
5. The method according to claim 4, characterized in that, The step of training the finger pressure hemostasis compression point detection model based on the training images and validating the training results based on the verification images includes: The verification images are categorized according to human body parts and converted into data conforming to the YOLO algorithm. The data and training images of the same human body parts are input into a single acupressure hemostasis pressure point detection model for a preset number of training cycles, thereby training the acupressure hemostasis pressure point detection model for different human body parts. The training results of the finger pressure hemostasis compression point detection model for different human body parts are obtained, and the annotation accuracy of the training results is calculated based on the verification image and loss function.
6. The method according to claim 5, characterized in that, If the verification fails, the acupressure hemostasis point detection model is retrained, including: The training accuracy threshold is used to determine whether the finger pressure hemostasis compression point detection model for different human body parts is qualified. If the labeling accuracy of the finger pressure hemostasis compression point detection model for the current human body part reaches the qualified accuracy threshold, then the finger pressure hemostasis compression point detection model for the current human body part is qualified for training. If the labeling accuracy of the finger pressure hemostasis compression point detection model for the current human body part does not reach the qualified accuracy threshold, then the finger pressure hemostasis compression point detection model for the current human body part will be trained again.
7. The method according to claim 6, characterized in that, The step of inputting the detected sample into a pre-created finger pressure hemostasis compression point detection model includes: Obtain user instructions and determine the corresponding acupressure hemostasis pressure point detection model for the human body part based on the user instructions; The test sample is input into the finger pressure hemostasis compression point detection model.
8. The method according to claim 7, characterized in that, The test samples include at least: images, videos, and real-time video streams.
9. A device for locating acupressure points for hemostasis, characterized in that, It includes a processor and a memory, wherein the processor is connected to the memory: The processor is used to call and execute the program stored in the memory; The memory is used to store the program, which is at least used to execute the finger pressure hemostasis point location method according to any one of claims 1-8.
Citation Information
Patent Citations
Cardiopulmonary resuscitation guide method, apparatus and computer program using augmented reality
KR102274034B1