Acupoint detection effect determination method, device and electronic equipment

Abstract

This invention provides a method, apparatus, and electronic device for determining the effectiveness of acupoint detection; it relates to the fields of artificial intelligence and big data technology; it can quantify the effectiveness of acupoint detection and provide a reference for model optimization. The method includes: determining predicted acupoints in a target image by using an acupoint detection model; acquiring annotation information of the target image, and determining a scale factor for the target image using the annotation information; calculating the prediction error of the predicted acupoints using the scale factor, and determining the prediction effectiveness of the acupoint detection model using the prediction error.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and big data processing technology, and in particular to a method, device and electronic equipment for determining acupoint detection effect. Background Art

[0002] The convolutional neural network model needs to be trained before application. Using a well-labeled image dataset, the coordinates of the acupuncture points in each image are recorded in advance, and the model is trained using the image dataset to obtain the model parameters. Since there are many parameters that need to be set during the model training process, different parameter settings will train different models. The pixel error between the image detection result and the labeled position is usually used to evaluate the model positioning effect. However, there are differences in human body shape and the shooting angle of the image will also affect the labeled position. It is not accurate to evaluate by the pixel error between the detection result and the labeled position. Summary of the invention

[0003] The present invention provides a method, device and electronic device for determining the effect of acupoint detection, which can at least solve the problem in the prior art that the prediction error calculation of the model is inaccurate, resulting in difficulty in optimizing the model effect.

[0004] In a first aspect, the present invention provides a method for determining an acupoint detection effect, comprising:

[0005] Determine the predicted acupoints of the part to be detected in the target image through the acupoint detection model;

[0006] Acquire annotation information of the target image, and determine a scale factor according to the annotation information;

[0007] The prediction error of the predicted acupoint is calculated using the scale factor, and the prediction effect of the acupoint detection model is determined by the prediction error.

[0008] According to an exemplary embodiment provided by the present invention, the part to be detected includes a hand, and determining a scale factor for the target image by using the annotation information includes:

[0009] The length of two knuckles of the middle finger in the target image is determined by the annotation information as the scale factor.

[0010] According to an exemplary embodiment provided by the present invention, the annotation information includes a plurality of annotated acupoints, and determining the length of two knuckles of the middle finger in the target image by using the annotation information as the scale factor includes:

[0011] Calculating the distance between the first acupoint and the second acupoint in the marked acupoints to obtain the length of two knuckles of the middle finger in the target image;

[0012] Among them, the first acupoint is the Zhongchong point of the middle finger, and the second acupoint is the Sifeng point of the middle finger.

[0013] According to an exemplary embodiment provided by the present invention, the method further includes:

[0014] When the part to be detected in the target image is the back of the hand, adding a first annotation point and a second annotation point to the annotation information of the target image;

[0015] The first marking point is the coordinate of the tip of the middle finger, and the second marking point is the coordinate of the second knuckle of the middle finger.

[0016] According to an exemplary embodiment provided by the present invention, the calculating the prediction error of the predicted acupoint by using the scale factor includes:

[0017] Acquire the annotated acupoint corresponding to the predicted acupoint from the annotated information;

[0018] Calculating a first error between the predicted acupoint and the corresponding marked acupoint;

[0019] The first error is normalized by the scale factor to obtain a normalized prediction error.

[0020] According to an exemplary embodiment provided by the present invention, determining the prediction effect of the acupoint detection model by using the prediction error includes:

[0021] If the prediction error is within the error threshold, it is determined that the acupoint detection model correctly predicts the target image.

[0022] In a second aspect, the present invention further provides an acupoint detection effect determination device, comprising:

[0023] An acupoint prediction module is used to determine the predicted acupoints of the part to be detected in the target image through an acupoint detection model;

[0024] An error scale determination module, used to obtain the annotation information of the target image, and determine the scale factor for the target image according to the annotation information;

[0025] The error calculation module is used to calculate the prediction error of the predicted acupoint using the scale factor, and determine the prediction effect of the acupoint detection model through the prediction error.

[0026] In a third aspect, the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for determining the effect of acupoint detection as described above is implemented.

[0027] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for determining the effect of acupoint detection.

[0028] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-mentioned methods for determining the effect of acupoint detection.

[0029] The method, device and electronic device for determining the effect of acupoint detection provided by the present invention determine a scale factor for the image through the annotation information of the image, calculate the prediction error through the scale factor, and then determine whether the model is correct in prediction. The error of each image can be unified to the same scale through the scale factor, which can improve the accuracy of error measurement. In addition, the scale factor of each image only depends on the annotation of the image itself, which can adapt to the differences between different individuals and improve the robustness of the model. Furthermore, the prediction effect of the model can be more accurately evaluated through the error, which is conducive to model optimization and improves the prediction effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0031] Figure 1 It is one of the flow charts of the method for determining the acupoint detection effect provided by the present invention;

[0032] Figure 2 It is one of the schematic diagrams of the labeled information in the method for determining the acupoint detection effect provided by the present invention;

[0033] Figure 3 This is the second schematic diagram of the labeled information in the method for determining the acupoint detection effect provided by the present invention;

[0034] Figure 4 This is the second flow chart of the method for determining the acupoint detection effect provided by the present invention;

[0035] Figure 5 This is an application scenario diagram of the acupoint detection effect determination method provided by the present invention;

[0036] Figure 6 It is a structural schematic diagram of the acupoint detection effect determination device provided by the present invention;

[0037] Figure 7It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] The present invention provides a method for determining the effect of acupoint detection. The method can be executed by electronic devices such as personal computers, tablet computers, mobile phones, or servers, and the present invention does not specifically limit the form of expression of the electronic device.

[0040] Combine the following Figures 1 to 5 The technical solution of the method for determining the acupoint detection effect of the present invention is described.

[0041] like Figure 1 As shown, the acupoint detection effect determination method may include the following steps:

[0042] Step S10: Determine the predicted acupoints of the part to be detected in the target image through the acupoint detection model.

[0043] The acupoint detection model refers to a model constructed by a convolutional neural network algorithm, which is used to predict the location points of acupoints on a part to be detected, such as a hand. The predicted acupoint refers to the location of the acupoint in the image output by the acupoint detection model. For example, (1, 1) is a predicted acupoint, and the predicted acupoint can be the predicted location point of acupoint A. The part to be detected may include the hand, and may also include the head, limbs, etc. If there are multiple acupoints on the part to be detected, the predicted acupoints may include multiple acupoints. The target image may be an image of the part to be detected. For example, if it is necessary to detect the acupoints on the user's hand, the user's hand can be photographed to obtain the target image. The target image is input into the acupoint detection model, and the model can predict the acupoints on the hand in the target image, obtain the position of the acupoint in the target image, and then mark it in the target image.

[0044] Step S20: Acquire the annotation information of the target image, and determine the scale factor according to the annotation information.

[0045] The annotation information is information that is manually marked in advance, and may specifically include the location of each acupuncture point in the target image, and may also include other information that needs to be marked, such as the type of the part to be detected. The type of the part to be detected refers to the shooting angle of the part to be detected in the target image, and may include the front and back of the part to be detected. For example, when the part to be detected is a hand, the palm and back of the hand may be marked. The scale factor for the target image may be determined by the multiple acupuncture points marked in the annotation information.

[0046] The annotation information may include multiple manually annotated acupoints. Taking the hand as an example, the annotated acupoints may include Hegu, Erming, Zhongchong, etc. By calculating the distance between two specific acupoints in the annotation information, such as the first acupoint and the second acupoint, the calculated distance may be used as a scale factor. For example, the length of the two knuckles of the middle finger of the hand in the target image is used as the scale factor. Figure 2 As shown in the figure, the length between the two knuckles of the middle finger of the hand can be calculated by the Zhongchong point and the Sifeng point on the palm. Specifically, when annotating an image, each acupoint can be numbered first, the coordinates corresponding to each acupoint can be determined, and then the corresponding number can be marked at the corresponding coordinates on the image. According to the numbers of Zhongchong point and Sifeng point, the two annotated coordinates of Zhongchong point and Sifeng point are queried from the annotated information of the target image, and then the Euclidean distance between Zhongchong point and Sifeng point is calculated to obtain the scale factor of the target image.

[0047] In an exemplary embodiment, the scale factor may also be determined by other means, such as calculating the length of the middle finger of the hand as the scale factor through annotation information, calculating the length of the two knuckles of the index finger as the scale factor, etc., which also fall within the scope of protection of the present invention.

[0048] The scale factor of the target image only depends on the annotation information of the target image, which can eliminate the influence of different shooting angles, annotation errors, etc. between different target images, ensure the reliability of the scale factor, and thus help to unify the errors of all images at the same scale, thereby improving the rationality and accuracy of error measurement.

[0049] Some target images may be images of the back of the user's hand. In order to avoid the inability to obtain the Zhongchong acupoint and the Sifeng acupoint due to the image being an image of the back of the hand, when the part to be detected in the target image is the back of the hand, two additional annotation points are added to the annotation information of the target image, namely the first annotation point and the second annotation point, so as to determine the length of the two knuckles of the middle finger. The first annotation point and the second annotation point may be the coordinates of the tip of the middle finger and the second knuckle of the back of the hand, respectively. Figure 3 As shown, when determining the scale factors of these target images, the distance between the tip of the middle finger and the second knuckle can be calculated as the scale factor.

[0050] Step S30: Calculate the prediction error of the predicted acupoint using the scale factor, and determine the prediction effect of the acupoint detection model through the prediction error.

[0051] The predicted acupoint is the acupoint of the target image marked by the acupoint detection model, and there can be multiple predicted acupoints. Exemplarily, for a predicted acupoint A, the marked acupoint corresponding to the predicted acupoint A can be obtained from the marking information. The predicted acupoint A and the marked acupoint corresponding to it are the coordinate information of the same acupoint. For example, for a predicted acupoint "1", the marked acupoint with the same number "1" can be obtained from the marking information as the marked acupoint corresponding to the predicted acupoint. Specifically, for the predicted acupoint A of the target image, the Euclidean distance between the predicted acupoint A and the marked acupoint corresponding to it is calculated, and the calculated result is used as the first error of the predicted acupoint A. Then, the first error is normalized by the scale factor of the target image, so that the prediction error of each target image is unified to the same scale. Exemplarily, the normalized prediction error of the i-th predicted acupoint in the j-th target image can be expressed as: Among them, f j is the scale factor of the jth target image; i and j are positive integers. In this embodiment, there is no special limitation on the number of predicted acupoints and the number of target images.

[0052] In an exemplary embodiment, by setting an error threshold, it is possible to determine whether the predicted acupoints output by the acupoint detection model are correct, thereby judging the quality of the acupoint detection model. Exemplarily, the prediction error of the predicted acupoints in the target image is calculated. If the prediction error is within the error threshold, it can be determined that the prediction of the target image by the acupoint detection model is correct. The number of correct predictions of the acupoint detection model in a certain number of target images can be counted to obtain the accuracy of the acupoint detection model. The accuracy rate can be used to measure the advantages and disadvantages of different acupoint detection models, thereby selecting the model with the highest accuracy rate for prediction, which is conducive to ensuring the accuracy of acupoint prediction.

[0053] If there are multiple acupoints to be predicted in the part to be detected, the accuracy of the acupoint detection model for each acupoint can be determined separately. For example, a test data set is used to test the prediction effect of the target image. Assume that there are M target images in the test data set, and there are C images in which the i-th acupoint is correctly identified. i Zhang, then the average accuracy rate of the i-th acupoint is: This method can be used to determine the average accuracy of the acupoint detection model for each acupoint, so that a model with better prediction effect can be selected for different predicted acupoints. For example, when it is necessary to predict "Acupoint 1" and "Acupoint 2" in image a, acupoint detection model A has the highest average accuracy for "Acupoint 1" and "Acupoint 2", and the predicted acupoints of these two acupoints in image a can be obtained through the acupoint detection model A. Alternatively, after calculating the average accuracy of each predicted acupoint, an acupoint detection model with a relatively high average accuracy for all acupoints can be selected to predict the acupoints of the image.

[0054] If the average accuracy of a certain acupoint of the acupoint detection model is not high, multiple training images can be obtained specifically for the acupoint to optimize the training acupoint detection model, thereby ensuring the accuracy of each acupoint of the acupoint detection model. The accuracy can be used to measure the quality of the model, thereby quantitatively evaluating the model, which is conducive to optimizing the model, obtaining the best model, and improving the accuracy of acupoint detection.

[0055] Furthermore, the present invention also provides an acupoint positioning system, such as Figure 4 As shown. The acupoint positioning system 50 includes an electronic device 51, a binocular camera 52 and a robotic arm 53. Among them, the electronic device 51 can be a personal computer, a tablet computer, etc. An algorithm capable of acupoint detection is deployed in the electronic device 51, such as a convolutional neural network model, and the algorithm can determine the location of the acupoint. The binocular camera 52 can be used to collect depth images of the part to be detected, and the collected images can be sent to the electronic device 51. For example, the binocular camera 52 can collect images of human hands and send the collected hand images to the electronic device 51. After receiving the hand image, the electronic device 51 can input the hand image into the convolutional neural network model, and the convolutional neural network model can detect the acupoints of the hand in the image, and then mark the coordinates of the acupoints in the hand image, and output the image with the mark, and the electronic device 51 can display the image with the mark. Exemplarily, the electronic device 51 can determine whether the acupoints detected by it are correct during the training process of the convolutional neural network model through the above-mentioned acupoint detection effect determination method of the present invention. If the coordinates of the acupuncture points determined by the model are incorrect, the electronic device can optimize the model, or select a model that detects the correct acupuncture points. The electronic device 51 can send the coordinates of the acupuncture points determined by the convolutional neural network model to the robotic arm 53. The robotic arm 53 is a mechanism designed according to the human arm, which can imitate the posture and operation of the human hand, such as grasping, etc. Exemplarily, the robotic arm 53 can be controlled by instructions so that it can autonomously move to the coordinates of the human hand according to the received coordinates to perform massage, acupuncture and other operations on the human hand.

[0056] based on Figure 4The system shown in the figure, the present invention may include steps S41 to S47, such as Figure 5 shown. Figure 5 The steps shown can be executed by the above-mentioned electronic device 51, or by other electronic devices, and the present invention does not make special limitations on this. Specifically, in step S41, a hand image data set and annotation information are obtained. The hand image data set includes multiple hand images. Each hand image is manually annotated, and the position of each acupoint of the hand is annotated as annotation information. If the image is the back of the hand, the position of the middle finger tip and the second finger joint is added to the annotation information. In step S42, a convolutional neural network model is trained. The convolutional neural network model for acupoint detection is trained using the above-mentioned hand image data set. In step S43, the prediction effect of the convolutional neural network model is determined. The prediction error of each image in the hand image data set is calculated by the method of calculating the prediction error in the above-mentioned steps S10 to S30, and the accuracy of the convolutional neural network model for each acupoint is determined. If the accuracy of a certain acupoint does not meet the requirements, the convolutional neural network model continues to be optimized until the accuracy of the convolutional neural network model meets the requirements. The convolutional neural network model that meets the requirements after optimization can be deployed and applied, such as Figure 4 As shown, the convolutional neural network model can be deployed in the electronic device 51. Next, the acupuncture points of the human body can be automatically detected using the trained neural network model in the electronic device 51. In step S44, the hand image to be identified is collected. The user can place his arm on the workbench, and the binocular camera 52 can take a picture of the user's hand to obtain a hand image. The hand image obtained by the binocular camera 52 has depth information. In step S45, the convolutional neural network model obtained in the above steps is used to output the predicted acupuncture points of the hand image. In step S46, the three-dimensional coordinates of the predicted acupuncture points are measured. Through the principle of binocular vision, the depth information of the hand image can be combined to transform the predicted acupuncture points into the world coordinate system to obtain the three-dimensional coordinates of the predicted acupuncture points. In step S47, the robot is controlled to perform massage, acupuncture and other operations according to the three-dimensional coordinates. The robot may include Figure 4 The three-dimensional coordinates of the predicted acupoints are sent to the robotic arm 53, and the robotic arm 53 can perform massage, acupuncture and other operations on the acupoints according to the movement to the position of the three-dimensional coordinates. In this embodiment, the robot realizes automatic acupoint massage, acupuncture and other operations, and applies artificial intelligence to acupoint treatment, which can expand the scope of use of artificial intelligence and provide a new research direction for artificial intelligence.

[0057] The present invention also provides an acupoint detection effect determination device, which can be used to execute the above-mentioned acupoint detection effect determination method. The acupoint detection effect determination device provided by the present invention is described below.

[0058] like Figure 6 As shown, the acupoint detection effect determination device 600 may include an acupoint prediction module 610 , an error scale determination module 620 , and an error calculation module 630 .

[0059] Specifically, the acupoint prediction module 610 is used to determine the predicted acupoints of the to-be-detected part in the target image through the acupoint detection model. The error scale determination module 620 is used to obtain the annotation information of the target image, and determine the scale factor for the target image through the annotation information. The error calculation module 630 is used to calculate the prediction error of the predicted acupoint using the scale factor, and determine the prediction effect of the acupoint detection model through the prediction error.

[0060] In an exemplary embodiment of the present invention, the part to be detected includes a hand, and the error scale determination module 620 is specifically configured to determine the length of two knuckles of the middle finger in the target image through the annotation information as the scale factor.

[0061] In an exemplary embodiment of the present invention, the annotation information includes multiple annotated acupoints, and the error scale determination module 620 can be configured to: calculate the distance between the first acupoint and the second acupoint in the annotated acupoints to obtain the length of two knuckles of the middle finger in the target image; wherein the first acupoint is the Zhongchong acupoint of the middle finger, and the second acupoint is the Sifeng acupoint of the middle finger.

[0062] In an exemplary embodiment of the present invention, the acupoint detection effect determination module 600 further includes a labeling module. The labeling module is used to add a first labeling point and a second labeling point to the labeling information of the target image when the part to be detected in the target image is the back of the hand; wherein the first labeling point is the coordinate of the tip of the middle finger, and the second labeling point is the coordinate of the second knuckle of the middle finger.

[0063] In an exemplary embodiment of the present invention, the error calculation module 630 includes a label acquisition module, a distance calculation module and an error normalization module. Specifically, the label acquisition module is used to obtain the labeled acupoint corresponding to the predicted acupoint from the label information. The label acquisition module is used to calculate the first error between the predicted acupoint and the corresponding labeled acupoint. The error normalization module is used to normalize the first error by the scale factor to obtain the normalized prediction error.

[0064] In an exemplary embodiment of the present invention, the error calculation module 630 is configured to determine that the acupoint detection model correctly predicts the target image if the prediction error is within an error threshold.

[0065] Since the various functional modules of the acupoint detection effect determination device in the exemplary embodiment of the present invention correspond to the steps of the exemplary implementation method of the acupoint detection effect determination device described above, for details not disclosed in the device embodiment of the present invention, please refer to the implementation method of the acupoint detection effect determination device described above.

[0066] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the above-mentioned acupoint detection effect determination device method, which includes: step S10, determining the predicted acupoint of the to-be-detected part in the target image through the acupoint detection model; step S20, obtaining the annotation information of the target image, and determining the scale factor for the target image through the annotation information; step S30, calculating the prediction error of the predicted acupoint using the scale factor, and determining the prediction effect of the acupoint detection model through the prediction error.

[0067] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0068] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the acupoint detection effect determination device method provided by the above-mentioned methods, and the method includes: step S10, determining the predicted acupoints of the to-be-detected part in the target image through an acupoint detection model; step S20, obtaining the annotation information of the target image, and determining the scale factor for the target image through the annotation information; step S30, calculating the prediction error of the predicted acupoint using the scale factor, and determining the prediction effect of the acupoint detection model through the prediction error.

[0069] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the acupoint detection effect determination device method provided by the above-mentioned methods, the method comprising: step S10, determining the predicted acupoints of the to-be-detected part in the target image through an acupoint detection model; step S20, acquiring annotation information of the target image, and determining a scale factor for the target image through the annotation information; step S30, calculating the prediction error of the predicted acupoint using the scale factor, and determining the prediction effect of the acupoint detection model through the prediction error.

[0070] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0071] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining acupoint detection effect, characterized in that: include: Determine the predicted acupoints of the part to be detected in the target image through the acupoint detection model; Acquire annotation information of the target image, and determine a scale factor for the target image according to the annotation information, wherein the annotation information includes a shooting angle of a part to be detected in the target image; Calculating the prediction error of the predicted acupoint by using the scale factor, and determining the prediction effect of the acupoint detection model by using the prediction error; The part to be detected includes a hand, and determining a scale factor for the target image by using the annotation information includes: Determine the length of two knuckles of the middle finger in the target image according to the annotation information as the scale factor; The annotation information includes a plurality of annotated acupoints, and determining the length of two knuckles of the middle finger in the target image by using the annotation information as the scale factor includes: Calculating the distance between the first acupoint and the second acupoint in the marked acupoints to obtain the length of two knuckles of the middle finger in the target image; Among them, the first acupoint is the Zhongchong point of the middle finger, and the second acupoint is the Sifeng point of the middle finger.

2. The method for determining acupoint detection effect according to claim 1, characterized in that: The method further comprises: When the part to be detected in the target image is the back of the hand, adding a first annotation point and a second annotation point to the annotation information of the target image; The first marking point is the coordinate of the tip of the middle finger, and the second marking point is the coordinate of the second knuckle of the middle finger.

3. The method for determining acupoint detection effect according to claim 1, characterized in that: The step of calculating the prediction error of the predicted acupoint by using the scale factor comprises: Acquire the annotated acupoint corresponding to the predicted acupoint from the annotated information; Calculating a first error between the predicted acupoint and the corresponding marked acupoint; The first error is normalized by the scale factor to obtain a normalized prediction error.

4. The method for determining acupoint detection effect according to claim 1, characterized in that: Determining the prediction effect of the acupoint detection model by using the prediction error includes: If the prediction error is within the error threshold, it is determined that the acupoint detection model correctly predicts the target image.

5. An acupoint detection effect determination device for implementing the acupoint detection effect determination method according to claim 1, characterized in that: include: An acupoint prediction module is used to determine the predicted acupoints of the part to be detected in the target image through an acupoint detection model; An error scale determination module, used to obtain the annotation information of the target image, and determine the scale factor for the target image according to the annotation information; The error calculation module is used to calculate the prediction error of the predicted acupoint using the scale factor, and determine the prediction effect of the acupoint detection model through the prediction error.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for determining the acupoint detection effect as described in any one of claims 1 to 4 is implemented.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for determining the acupoint detection effect as claimed in any one of claims 1 to 4 is implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for determining the acupoint detection effect as claimed in any one of claims 1 to 4 is implemented.

Images