A multi-lead electrocardio electrode patch positioning method based on human body surface features

Abstract

The application provides a multi-lead electrocardiogram electrode patch positioning method based on human body surface features, and belongs to the technical field of medical devices, and comprises the following steps: inputting user-worn electrocardiogram electrode patch video data into a pre-trained target positioning model to obtain human upper body surface features and electrocardiogram electrode patch feature state information; calculating and obtaining feature operation correlation results of the human upper body surface features and the electrocardiogram electrode patch feature state information, generating electrocardiogram electrode patch wearing state information, and determining electrocardiogram electrode patch positioning guidance results from the electrocardiogram electrode patch wearing state information. The terminal with a video function is used to collect the body surface features and electrode patch position information of the user in real time when the user wears the electrocardiogram electrode patch, and information such as whether the wearing is correct or error correction methods is fed back to the user, so that the user can correctly wear the electrocardiogram electrode patch, and the error of electrocardiogram signal measurement caused by incorrect wearing of the electrocardiogram electrode patch is effectively reduced or avoided.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, and in particular to a method for positioning multi-lead electrocardiogram electrode patches based on human body surface features. Background Technology

[0002] In the diagnosis of cardiovascular diseases, electrocardiogram (ECG) is one of the important indicators. ECG data is usually obtained by patients wearing special ECG testing equipment. The most common method is to wear multi-lead ECG electrode patches to obtain their own ECG. By observing the ECG, we can understand whether the examinee has heart disease.

[0003] With the increasing popularity of services such as remote consultations and home health monitoring, more and more users are obtaining their electrocardiogram (ECG) data at any time through home self-testing. However, most users have not received specialized training on the proper wearing of electrode patches, which may lead to significant errors in the collected ECGs due to inaccurate electrode patch placement. This can affect the accuracy of doctors' or instruments' analysis of ECG signals and cardiac diagnosis.

[0004] Therefore, there is an urgent need to provide an effective method to help users position ECG electrode patches and assist them in wearing ECG electrode patches correctly. Summary of the Invention

[0005] This invention provides a method for positioning multi-lead ECG electrode patches based on human body surface features, in order to address the deficiency in existing technologies that lack methods to guide users on the correct wearing of ECG electrode patches.

[0006] The present invention provides a method for positioning multi-lead ECG electrode patches based on human body surface features, comprising:

[0007] Collect video data of the user wearing ECG electrode patches, and input the video data of the user wearing ECG electrode patches into a pre-trained target localization model to obtain the upper body surface features and ECG electrode patch feature status information.

[0008] Calculate the feature operation association results of the upper body surface features and the ECG electrode patch feature state information;

[0009] Based on the feature operation association results, ECG electrode patch wearing status information is generated, and ECG electrode patch positioning guidance results are determined based on the ECG electrode patch wearing status information.

[0010] According to the present invention, a method for locating multi-lead ECG electrode patches based on human body surface features is provided. The method involves acquiring video data of a user wearing ECG electrode patches, inputting this video data into a pre-trained target localization model, and obtaining upper body surface features and ECG electrode patch feature state information, including:

[0011] Initialize an N-dimensional vector containing several feature state information, where N is the total number of human upper body surface features plus ECG electrode patches. The several feature state information includes target type, target score, x-coordinate of target center point, y-coordinate of target center point, target detection box height, and target detection box width.

[0012] If the target type is determined to be a human nipple, the components of the several feature state information corresponding to the human nipple are sorted in non-ascending order, and the first two components are taken as the human nipple detection result.

[0013] If the target type is determined to be the upper body boundary line of the human body, then the components of the several feature state information corresponding to the upper body boundary line of the human body are sorted in non-ascending order, and the first two components are taken as the human body boundary line detection result.

[0014] If the target type is determined to be an electrode patch, the components corresponding to the electrode patch in the several feature state information are sorted in non-ascending order, and the first few components are taken as the electrode patch detection result.

[0015] A two-dimensional matrix is ​​constructed based on the detection results of the human nipple, the detection results of the human body boundary line, and the detection results of the electrode patches. The two-dimensional matrix includes the human right nipple component, the human left nipple component, the human right boundary line component, the human left boundary line component, the right upper limb electrode patch component, the left upper limb electrode patch component, the first chest electrode patch component, the second chest electrode patch component, the third chest electrode patch component, the fourth chest electrode patch component, the fifth chest electrode patch component, the sixth chest electrode patch component, the right lower limb electrode patch component, and the left lower limb electrode patch component.

[0016] According to the present invention, a method for locating multi-lead electrocardiogram (ECG) electrode patches based on human body surface features is provided. The step of calculating and obtaining the feature operation association result of the upper body surface features and the ECG electrode patch feature state information includes:

[0017] If the x-coordinates of the first, second, third, fourth, fifth, and sixth chest electrode patch components are determined to increase sequentially, then the subsequent calculation steps are executed; otherwise, the first feature operation association result is returned.

[0018] According to the present invention, a method for locating multi-lead electrocardiogram (ECG) electrode patches based on human body surface features, wherein the step of calculating and obtaining the feature operation association result of the upper body surface features and the ECG electrode patch feature state information further includes:

[0019] Obtain the first ratio of the absolute value of the difference between the ordinate of the first chest electrode patch component and the ordinate of the second chest electrode patch component to the absolute value of the difference between the abscissa of the first chest electrode patch component and the abscissa of the second chest electrode patch component, and the second ratio of the absolute value of the difference between the ordinate of the right nipple component and the ordinate of the left nipple component to the absolute value of the difference between the abscissa of the right nipple component and the abscissa of the left nipple component.

[0020] If the absolute value of the difference between the first ratio and the second ratio is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the second feature operation association result is returned.

[0021] Obtain the first product obtained by multiplying the difference between the x-coordinate of the first chest electrode patch component and the x-coordinate of the left nipple component by the difference between the y-coordinate of the left nipple component and the y-coordinate of the right nipple component, and the second product obtained by multiplying the difference between the y-coordinate of the first chest electrode patch component and the y-coordinate of the left nipple component by the difference between the x-coordinate of the left nipple component and the y-coordinate of the right nipple component.

[0022] If it is determined that the difference between the first product and the second product is less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the third feature operation association is returned.

[0023] If the Euclidean distance between the first chest electrode patch component and the right nipple component is less than or equal to the second preset error value, and the Euclidean distance between the second chest electrode patch component and the left nipple component is less than or equal to the second preset error value, then proceed with the subsequent calculation steps; otherwise, return to the fourth feature operation association result.

[0024] According to the present invention, a method for locating multi-lead electrocardiogram (ECG) electrode patches based on human body surface features, wherein the step of calculating and obtaining the feature operation association result of the upper body surface features and the ECG electrode patch feature state information further includes:

[0025] The third product is obtained by multiplying the difference between the abscissa of the fourth chest electrode patch component and the abscissa of the left nipple component by the difference between the abscissa of the left nipple component and the abscissa of the right nipple component, and the fourth product is obtained by multiplying the difference between the ordinate of the fourth chest electrode patch component and the ordinate of the left nipple component by the difference between the ordinate of the left nipple component and the ordinate of the right nipple component.

[0026] If the difference between the third product and the fourth product is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the fifth feature operation is returned.

[0027] If the Euclidean distance between the fourth chest electrode patch component and the left nipple component is determined to be less than or equal to the second preset error value, then proceed with the subsequent calculation steps; otherwise, return to the sixth feature operation association result.

[0028] According to the present invention, a method for locating multi-lead electrocardiogram (ECG) electrode patches based on human body surface features, wherein the step of calculating and obtaining the feature operation association result of the upper body surface features and the ECG electrode patch feature state information further includes:

[0029] Obtain the first straight-line distance from the center point of the third chest electrode patch to the line connecting the center points of the second and fourth chest electrode patches, the second straight-line distance from the center point of the third chest electrode patch to the center point of the second chest electrode patch, and the third straight-line distance from the center point of the third chest electrode patch to the center point of the fourth chest electrode patch.

[0030] If it is determined that the difference between the first straight-line distance and the second preset error value is less than or equal to 0, then the subsequent calculation steps are executed; otherwise, the result of the seventh feature operation is returned.

[0031] If the absolute value of the difference between the second straight-line distance and the third straight-line distance is less than or equal to the second preset error value, then the subsequent calculation steps are executed; otherwise, the result of the eighth feature operation is returned.

[0032] According to the present invention, a method for locating multi-lead electrocardiogram (ECG) electrode patches based on human body surface features, wherein the step of calculating and obtaining the feature operation association result of the upper body surface features and the ECG electrode patch feature state information further includes:

[0033] The fifth product is obtained by multiplying the difference between the x-coordinate of the sixth chest electrode patch component and the x-coordinate of the left boundary component of the human body by multiplying the difference between the x-coordinate of the left nipple component and the x-coordinate of the right nipple component of the human body, and the sixth product is obtained by multiplying the difference between the y-coordinate of the sixth chest electrode patch component and the y-coordinate of the left boundary component of the human body by multiplying the difference between the y-coordinate of the left nipple component and the y-coordinate of the right nipple component of the human body.

[0034] If the difference between the fifth product and the sixth product is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the ninth feature operation is returned.

[0035] Obtain the seventh product obtained by multiplying the difference between the x-coordinate of the sixth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the y-coordinate of the left nipple component and the y-coordinate of the right nipple component, and the eighth product obtained by multiplying the difference between the x-coordinate of the sixth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the x-coordinate of the left nipple component and the y-coordinate of the right nipple component.

[0036] If the difference between the seventh product and the eighth product is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the tenth feature operation is returned.

[0037] According to the present invention, a method for locating multi-lead electrocardiogram (ECG) electrode patches based on human body surface features, wherein the step of calculating and obtaining the feature operation association result of the upper body surface features and the ECG electrode patch feature state information further includes:

[0038] The ninth product is obtained by multiplying the difference between the x-coordinate of the fifth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the y-coordinate of the left nipple component and the x-coordinate of the right nipple component, and the tenth product is obtained by multiplying the difference between the y-coordinate of the fifth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the y-coordinate of the left nipple component and the x-coordinate of the right nipple component.

[0039] If the difference between the ninth product and the tenth product is determined to be less than the first preset error value, then the subsequent calculation steps are executed; otherwise, the eleventh feature operation association result is returned.

[0040] According to the present invention, a method for locating multi-lead ECG electrode patches based on human body surface features is provided. The second preset error value is obtained by multiplying the average border length of the sum of the target frame length and the target frame width of 10 electrode patches in a two-dimensional matrix by a preset ratio.

[0041] According to the present invention, a method for positioning multi-lead ECG electrode patches based on human body surface features includes generating ECG electrode patch wearing status information based on the feature calculation association results, and determining ECG electrode patch positioning guidance results based on the ECG electrode patch wearing status information, comprising:

[0042] If the result of the feature operation is determined to be a wearing error, the user is given feedback via voice or text regarding the incorrect electrode patch number and the electrode patch adjustment method.

[0043] If the result of the feature operation is determined to be correct wearing information, a voice or text message will be used to provide the user with a successful wearing notification.

[0044] The present invention provides a multi-lead ECG electrode patch positioning method based on human body surface features. It uses a terminal with video function to collect the user's body surface features and electrode patch position information in real time when wearing ECG electrode patches, and provides feedback to the user on whether the wearing is correct or how to correct the error, so as to assist the user in wearing ECG electrode patches correctly and effectively reduce or avoid errors in ECG signal measurement caused by incorrect wearing of ECG electrode patches. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0046] Figure 1 This is one of the flowcharts of the multi-lead ECG electrode patch positioning method based on human body surface features provided by the present invention;

[0047] Figure 2 This is the second flowchart of the multi-lead ECG electrode patch positioning method based on human body surface features provided by the present invention.

[0048] Figure 3 This is a schematic diagram showing the results of the electrocardiogram electrode patch and human body surface features provided by the present invention;

[0049] Figure 4 This is a schematic diagram of the target numbering results provided by the present invention;

[0050] Figure 5 This is a schematic diagram showing the result of wearing the electrode patch assisted by the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0052] With the popularization of online remote consultations and home health monitoring, users' demand for medical devices is also increasing. In terms of cardiovascular disease monitoring, users generally wear multi-lead ECG electrode patches to obtain ECG data and provide it to professional doctors for cardiovascular health monitoring and diagnosis. This invention takes the most commonly used 12-lead ECG electrode patch wearing method as the starting point. The overall idea includes using an automated target localization model to detect the user's upper body, identify the user's body surface features and electrode patch positions, then determine the target number, and after determining the number, calculate the relative positional relationship between the body surface features and the electrode patch, and determine whether the relative positional relationship meets the wearing specifications, thereby helping users to wear ECG electrode patches correctly.

[0053] Figure 1 This is one of the flowcharts illustrating the multi-lead ECG electrode patch positioning method based on human body surface features provided in this embodiment of the invention, such as... Figure 1 As shown, it includes:

[0054] Step 100: Collect video data of the user wearing ECG electrode patches, and input the video data of the user wearing ECG electrode patches into a pre-trained target localization model to obtain the upper body surface features and ECG electrode patch feature status information.

[0055] Step 200: Calculate the feature operation association result of the obtained upper body surface features and ECG electrode patch feature state information;

[0056] Step 300: Generate ECG electrode patch wearing status information based on the feature operation association results, and determine the ECG electrode patch positioning guidance result based on the ECG electrode patch wearing status information.

[0057] This invention first collects video data of a user wearing ECG electrode patches. A pre-trained automated target localization model processes the images or video streams in the video data for detection and localization. The target localization model outputs the features of the upper body surface and the ECG electrode patch's feature state information. Then, it calculates the feature operation association result of the upper body surface features and the ECG electrode patch's feature state information. Based on this feature operation association result, it determines whether the ECG electrode patch was successfully worn. If successful, the user is informed of the successful wearing result via voice or text; otherwise, the user is informed via voice or text how to adjust the incorrectly worn ECG electrode patch.

[0058] Specifically, the overall implementation process of the embodiments of the present invention is as follows: Figure 2As shown, after the user puts on the ECG electrode patch, a pre-trained target localization model is used to train the input video stream or image based on target detection of the video stream or image to obtain the body surface feature reference points and the position of the ECG electrode patch. Then, the wearing result of the ECG electrode patch is judged according to the auxiliary wearing algorithm. If the wearing is successful, the user is directly informed of the successful wearing result. If there is an error, the error information is fed back to the user in real time through voice or text, and the user is guided to adjust the position of the ECG electrode patch.

[0059] This invention uses a terminal with video functionality to collect real-time information on the user's body surface features and electrode position when wearing ECG electrode patches. It provides feedback to the user on whether the wear is correct or provides methods for correcting errors, thus assisting the user in correctly wearing the ECG electrode patches and effectively reducing or avoiding errors in ECG signal measurement caused by incorrect wear of the ECG electrode patches.

[0060] Based on the above embodiments, step 100 includes:

[0061] Initialize an N-dimensional vector containing several feature state information, where N is the total number of human upper body surface features plus ECG electrode patches. The several feature state information includes target type, target score, x-coordinate of target center point, y-coordinate of target center point, target detection box height, and target detection box width.

[0062] If the target type is determined to be a human nipple, the components of the several feature state information corresponding to the human nipple are sorted in non-ascending order, and the first two components are taken as the human nipple detection result.

[0063] If the target type is determined to be the upper body boundary line of the human body, then the components of the several feature state information corresponding to the upper body boundary line of the human body are sorted in non-ascending order, and the first two components are taken as the human body boundary line detection result.

[0064] If the target type is determined to be an electrode patch, the components corresponding to the electrode patch in the several feature state information are sorted in non-ascending order, and the first few components are taken as the electrode patch detection result.

[0065] A two-dimensional matrix is ​​constructed based on the detection results of the human nipple, the detection results of the human body boundary line, and the detection results of the electrode patches. The two-dimensional matrix includes the human right nipple component, the human left nipple component, the human right boundary line component, the human left boundary line component, the right upper limb electrode patch component, the left upper limb electrode patch component, the first chest electrode patch component, the second chest electrode patch component, the third chest electrode patch component, the fourth chest electrode patch component, the fifth chest electrode patch component, the sixth chest electrode patch component, the right lower limb electrode patch component, and the left lower limb electrode patch component.

[0066] This invention uses a pre-trained automatic target localization model to automatically locate targets in images or video streams input from the device, and obtains human body surface features and feature state information of multiple electrode patches in the images or video streams.

[0067] Specifically, images of naked human bodies and human bodies with electrode patches worn on the upper body were taken under various background environments. Targets appearing in all images (human nipples, left and right boundaries of the upper body, and electrode patches) were labeled to obtain a dataset for the automatic target localization model. An automatic target localization model for detecting and locating electrode patches and human body surface features was constructed and trained using the obtained dataset. During training, the number of target categories in the automatic target localization model was set to 3. The trained automatic target localization model was obtained, and the results for ECG electrode patches and human body surface features are as follows: Figure 3 As shown.

[0068] Initialize an N-dimensional vector object of data type feature state information, and store all the feature state information of the detected targets in this N-dimensional vector object; the rank of the object vector is N, where N is the number of human body surface features and electrode patches detected by the automatic target localization model in the image or video stream; the structure of the feature state information includes the target type, target score, x-coordinate of the target center point, y-coordinate of the target center point, height of the target detection box, and width of the target detection box; any component of the N-dimensional vector can be divided into three target types, here assuming target type 1 is human nipple, target type 2 is the human upper body boundary line, and target type 3 is electrode patch; sort the components of the N-dimensional vector belonging to the same target type in non-ascending order according to the target score value of the component:

[0069] (1) For target type 1, take the first two components after sorting the N-dimensional vector as the final detection result of human nipple;

[0070] (2) For target type 2, take the first two components after sorting the N-dimensional vector as the final detection result of the human body boundary line;

[0071] (3) For target type 3, take the first few components after sorting the N-dimensional vector as the final detection result of the electrode patch. Note that the number here should be consistent with the number of electrode patches worn.

[0072] Then, based on the positional relationships of components belonging to the same target type in the N-dimensional vector `object`, each detected target is assigned a number. Simultaneously, a two-dimensional matrix `matrix` is initialized, storing the target's number and corresponding feature state information in each column of each row of the matrix. The number of rows in the two-dimensional matrix `matrix` is the number of targets, meaning the index of each row corresponds to the target's number. The number of columns in the two-dimensional matrix should be the same as the number of feature states of the targets; that is, each column, from smallest to largest, corresponds to the target type, target score, target center point x-coordinate, target center point y-coordinate, target detection box height, and target detection box length.

[0073] First, the target elements of the human nipple are numbered, and the x-coordinates of the center points of the target types "human nipple" in the N-dimensional vector are compared. The component with the smaller x-coordinate is denoted as "RN" (right nipple) and stored in a 2D matrix, while the component with the larger x-coordinate is denoted as "LN" (left nipple) and stored in a 2D matrix. Next, the target elements of the human boundary line are numbered, and the x-coordinates of the center points of the target types "human boundary line" in the N-dimensional vector are compared. The component with the smaller x-coordinate is denoted as "RB" (right boundary line) and stored in a 2D matrix, while the component with the larger x-coordinate is denoted as "LB" (left boundary line) and stored in a 2D matrix. Finally, the target elements of the electrode patches are numbered, and the numbering here is related to the number of electrode patches. The components in the N-dimensional vector are sorted in non-descending order according to the x-coordinate of the center point of the target type, which is the human body boundary line. Taking a 12-lead system as an example, the sequential component numbers are "RA", "LA", "V1", "V2", "V3", "V4", "V5", "V6", "RL", and "LL", and stored in a two-dimensional matrix. The determined target results are as follows: Figure 4 As shown.

[0074] This invention uses the user's own body surface characteristics as a reference point for auxiliary examination, so it is applicable to a variety of complex background environments, and has a high accuracy rate and environmental tolerance.

[0075] Based on the above embodiments, step 200 includes:

[0076] Based on the human body surface features and the positional information of electrode patches in the image or video stream within the two-dimensional matrix, the relative positional relationships between the electrode patches and their positional relationships relative to the human body surface features are analyzed to determine whether they conform to medical electrode patch wearing standards. A result response code, resultCode, is returned, with specific classifications shown in Table 1.

[0077] Table 1

[0078]

[0079]

[0080] The first scenario involves analyzing whether the relative positions of electrode patches V1 to V6 satisfy the condition of sequentially increasing horizontal coordinates. If the condition is satisfied, the result response code resultCode = "correct" is returned; otherwise, the subsequent steps are executed. If the condition is not satisfied, the result response code resultCode = "error_1~6" is returned. The process for determining whether the relative positional relationship of electrode patches V1 to V6 meets the requirements is as follows:

[0081]

[0082] Where matrix[n][2] represents the x-coordinate of the center point of the electrode patch numbered Vn in the image or video stream.

[0083] The second scenario involves analyzing and checking whether the line connecting the center points of electrode patches V1 and V2 satisfies the positional relationship of being parallel to the line connecting the center points of RN and LN within the error range; if the relationship is satisfied, proceed with the subsequent steps; otherwise, return the result response code resultCode = "error_1~2_1";

[0084] The process for determining whether the line connecting the center points of electrode patches V1 and V2 meets the requirements with the line connecting the center points of RN and LN is as follows:

[0085]

[0086] Where matrix[n][3] represents the ordinate of the center point of the electrode patch numbered Vn in the image or video stream. matrix[6][3], matrix[7][3], matrix[0][3], and matrix[1][3] represent the ordinates of the center points of the electrode patches numbered V1 and V2 and the center points of the human nipples numbered RN and LN in the image or video, respectively. matrix[6][2], matrix[7][2], matrix[0][2], and matrix[1][2] represent the abscissas of the center points of the electrode patches numbered V1 and V2 and the center points of the human nipples numbered RN and LN in the image or video, respectively. α is the maximum permissible error value, i.e., the first preset error value.

[0087] Further analysis and inspection are conducted to check whether the line connecting the center points of electrode patches V1 and V2 satisfies the positional relationship of being on the same straight line as the line connecting the center points of RN and LN, and whether the distances between V1 and RN and V2 and LN are respectively within the allowable error range; if the relationship is satisfied, proceed to the next step; otherwise, return the result response code resultCode = "error_1~2_2";

[0088] The process for determining whether the line connecting the center points of electrode patches V1 and V2 meets the requirements of the line connecting the center points of RN and LN is as follows:

[0089] (matrix[6][2]-matrix[1][2])×(matrix[1][3]-matrix[0][3])-(matrix[6][3]-matrix[1][3])×(matrix[1][2]-matrix[0][2])≤α;

[0090] The process for determining whether the straight-line distances between electrode patches V1 and V2 and RN and LN respectively meet the requirements is shown in the Euclidean distance formula below:

[0091]

[0092] and

[0093]

[0094] Where matrix[6][2], matrix[0][2], matrix[7][2], and matrix[1][2] represent the horizontal coordinates of the center points of the electrode patches numbered V1 and V2 and the center points of the human nipples numbered RN and LN in the image or video stream, respectively; matrix[6][3], matrix[0][3], matrix[7][3], and matrix[1][3] represent the vertical coordinates of the center points of the electrode patches numbered V1 and V2 and the center points of the human nipples numbered RN and LN in the image or video stream, respectively; and β is the maximum permissible error value, i.e., the second preset error value.

[0095] The third scenario involves analyzing and checking whether the V4 electrode patch satisfies the positional relationship relative to the human body directly below the LN and whether the distance between the two is within the allowable error range. If both relationships are satisfied, the subsequent steps are executed; otherwise, if the former relationship is not satisfied, the result response code resultCode="error_4_1" is returned; if the former relationship is satisfied but the latter relationship is not satisfied, the result response code resultCode="error_4_2" is returned.

[0096] The process for determining whether the V4 electrode patch and LN meet the requirements is as follows:

[0097] (matrix[9][2]-matrix[1][2])×(matrix[1][2]-matrix[0][2])+(matrix[9][3]-matrix[1][3])×(matrix[1][3]-matrix[0][3])≤α;

[0098] The process for determining whether the distance between the V4 electrode patch and the LN meets the requirements is as follows:

[0099]

[0100] Where matrix[9][2], matrix[1][2], and matrix[0][2] represent the abscissas of the center point of the electrode patch numbered V4 and the center point of the human nipple numbered LN and RN in the image or video stream, respectively; matrix[9][3], matrix[1][3], and matrix[0][3] represent the ordinates of the center point of the electrode patch numbered V4 and the center point of the human nipple numbered LN and RN in the image or video stream, respectively; and α and β are the maximum permissible error values.

[0101] The fourth case involves calculating the distance range from the center point of electrode patch V3 to the straight line connecting the center points of electrode patches V2 and V4. (v3,v2v4) And the distances from the center point of electrode patch V3 to the center points of electrode patches V2 and V4 are respectively called range. (v3,v2) and range (v3,v4) and determine the range (v3,v2v4) Does it satisfy the relationship that is 0 within the allowable error range and the range? (v3,v2) and range (v3,v4) Check if the relationship of equality within the error range is satisfied; if both satisfy the relationship, proceed to the next step; if the former does not satisfy the relationship, return the result response code resultCode="error_3_1"; if the former satisfies the relationship but the latter does not, return the result response code resultCode="error_3_2".

[0102] range (v3,v2v4) The process for determining whether the requirement of zero within the allowable error range is as follows:

[0103] range (v3,v2v4) -β≤0;

[0104] range (v3,v2) and range(v2,v4) The process for determining whether the requirement of equality within the allowable error range is met is as follows:

[0105] |range (v3,v2) -range (v2,v4) |≤β;

[0106] Where β is the maximum allowable error range.

[0107] The fifth scenario involves analyzing whether the center point of the V6 electrode patch satisfies the positional relationship on the left boundary line of the human body, and whether the line connecting the center points of the V6 and V4 electrode patches satisfies the positional relationship with the nipples of both individuals. If both relationships are satisfied, the subsequent steps are executed; otherwise, if the former relationship is not satisfied, the result response code resultCode="error_6_1" is returned; if the former relationship is satisfied but the latter relationship is not satisfied, the result response code resultCode="error_6_2" is returned.

[0108] The process for determining whether the center point of the V6 electrode patch meets the requirement of being on the left side boundary of the human body is as follows:

[0109] (matrix

[11] [2]-matrix[1][2])×(matrix[1][2]-matrix[0][2])+(matrix

[11] [3]-matrix[3][3])×(matrix[2][3]-matrix[0][3])≤α;

[0110] The process for determining whether the line connecting the center points of the V6 electrode patch and the V4 electrode patch meets the requirement of being at the same level as the nipples of the two individuals is as follows:

[0111] (matrix[9][2]-matrix[9][2])×(matrix[1][3]-matrix[0][3])-(matrix

[11] [3]-matrix[9][3])×(matrix[1][2]-matrix[0][2])≤α;

[0112] Where matrix[9][2], matrix

[11] [2], matrix[1][2], matrix[0][2], and matrix[3][2] represent the horizontal coordinates of the center points of electrode patches numbered V4 and V6, the center points of human nipples numbered LN and RN, and the center point of the left boundary line of the human body numbered LB in the image or video stream, respectively. matrix[9][3], matrix

[11] [3], matrix[1][3], matrix[0][3], and matrix[3][3] represent the vertical coordinates of the center points of electrode patches numbered V4 and V6, the center points of human nipples numbered LN and RN, and the center point of the left boundary line of the human body numbered LB in the image or video stream, respectively. α is the maximum permissible error value.

[0113] The sixth scenario involves analyzing and checking whether the V4, V5, and V6 electrode patches are in a horizontal position relative to the human body within the error range. If the relationship is satisfied, the result response code resultCode="correct" is returned; otherwise, the result response code resultCode="error_5" is returned.

[0114] The process for determining whether the V4, V5, and V6 electrode patches meet the requirements of being horizontal relative to the human body within the allowable error range is as follows:

[0115] (matrix

[10] [2]-matrix[9][2])×(matrix[1][3]-matrix[0][3])-(matrix

[10] [3]-matrix[9][3])×(matrix[1][2]-matrix[0][2])≤α;

[0116] Where matrix[9][2], matrix

[10] [2], matrix[1][2], and matrix[0][2] represent the horizontal coordinates of the center points of electrode patches numbered V4 and V5 and the center points of human nipples numbered LN and RN in the image or video stream, respectively; matrix[9][3], matrix

[10] [3], matrix[1][3], and matrix[0][3] represent the vertical coordinates of the center points of electrode patches numbered V4 and V5 and the center points of human nipples numbered LN and RN in the image or video stream, respectively; and α is the maximum permissible error value.

[0117] It should be noted that the maximum permissible error α in this embodiment of the invention is the permissible error set for determining the perpendicularity of the straight line; the maximum permissible error β is here set to two-thirds of the average border length of the sum of the lengths and widths of the 10 electrode patches in the obtained target image:

[0118]

[0119] Where matrix[i+4][4] and matrix[i+4][5] represent the width and height of the i-th electrode patch target box, respectively.

[0120] Understandably, if the calculation results are not among the above-mentioned cases, the result response code resultCode = "correct" will be returned.

[0121] Based on the above embodiments, step 300 includes:

[0122] If the result of the feature operation is determined to be a wearing error, the user is given feedback via voice or text regarding the incorrect electrode patch number and the electrode patch adjustment method.

[0123] If the result of the feature operation is determined to be correct wearing information, a successful wearing prompt message will be sent to the user via voice or text.

[0124] Specifically, according to the different result response codes obtained in the foregoing embodiments, the electrode patch number designed for the result response code and the corresponding prompt information are obtained from Table 1, and fed back to the customer in the form of voice or text, such as... Figure 5 The diagram shows an example of using voice or text to provide feedback on the results of an auxiliary examination involving electrode patch placement. Figure 5 The system will highlight the problematic electrode patch and provide corresponding voice and text prompts, such as "The V4 electrode patch is not located directly below the left nipple. Please adjust it."

[0125] The method proposed in this invention can be applied to edge devices such as mobile or PC applications or web pages. When users need to wear multi-lead ECG electrode patches themselves during home health monitoring or online consultations, they can use the method provided by this invention to complete the wearing by interacting with the device, thereby ensuring the successful wearing of the electrode patches and effectively improving the quality of the acquired ECG signals.

[0126] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0127] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for positioning multi-lead ECG electrode patches, characterized in that, include: Collect video data of the user wearing ECG electrode patches, and input the video data of the user wearing ECG electrode patches into a pre-trained target localization model to obtain the upper body surface features and ECG electrode patch feature status information. Calculate the feature operation association results of the upper body surface features and the ECG electrode patch feature state information; Based on the feature operation association results, ECG electrode patch wearing status information is generated, and ECG electrode patch positioning guidance results are determined based on the ECG electrode patch wearing status information; The process involves collecting video data of the user wearing ECG electrode patches, inputting this data into a pre-trained target localization model, and obtaining information on the upper body surface features and ECG electrode patch features, including: Initialize an N-dimensional vector containing several feature state information, where N is the total number of human upper body surface features plus ECG electrode patches. The several feature state information includes target type, target score, x-coordinate of target center point, y-coordinate of target center point, target detection box height, and target detection box width. If the target type is determined to be a human nipple, the components of the several feature state information corresponding to the human nipple are sorted in non-ascending order, and the first two components are taken as the human nipple detection result. If the target type is determined to be the upper body boundary line of the human body, then the components of the several feature state information corresponding to the upper body boundary line of the human body are sorted in non-ascending order, and the first two components are taken as the human body boundary line detection result. If the target type is determined to be an electrode patch, the components corresponding to the electrode patch in the several feature state information are sorted in non-ascending order, and the first few components are taken as the electrode patch detection result. A two-dimensional matrix is ​​constructed based on the detection results of the human nipple, the detection results of the human body boundary line, and the detection results of the electrode patches. The two-dimensional matrix includes the human right nipple component, the human left nipple component, the human right boundary line component, the human left boundary line component, the right upper limb electrode patch component, the left upper limb electrode patch component, the first chest electrode patch component, the second chest electrode patch component, the third chest electrode patch component, the fourth chest electrode patch component, the fifth chest electrode patch component, the sixth chest electrode patch component, the right lower limb electrode patch component, and the left lower limb electrode patch component.

2. The multi-lead ECG electrode patch positioning method according to claim 1, characterized in that, The calculation of the feature operation association results of the upper body surface features and the ECG electrode patch feature state information includes: If the x-coordinates of the first, second, third, fourth, fifth, and sixth chest electrode patch components are determined to increase sequentially, then the subsequent calculation steps are executed; otherwise, the first feature operation association result is returned.

3. The multi-lead ECG electrode patch positioning method according to claim 2, characterized in that, The calculation of the feature operation association result of obtaining the upper body surface features and the ECG electrode patch feature state information also includes: Obtain the first ratio of the absolute value of the difference between the ordinate of the first chest electrode patch component and the ordinate of the second chest electrode patch component to the absolute value of the difference between the abscissa of the first chest electrode patch component and the abscissa of the second chest electrode patch component, and the second ratio of the absolute value of the difference between the ordinate of the right nipple component and the ordinate of the left nipple component to the absolute value of the difference between the abscissa of the right nipple component and the abscissa of the left nipple component. If the absolute value of the difference between the first ratio and the second ratio is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the second feature operation association result is returned. Obtain the first product obtained by multiplying the difference between the x-coordinate of the first chest electrode patch component and the x-coordinate of the left nipple component by the difference between the y-coordinate of the left nipple component and the y-coordinate of the right nipple component, and the second product obtained by multiplying the difference between the y-coordinate of the first chest electrode patch component and the y-coordinate of the left nipple component by the difference between the x-coordinate of the left nipple component and the y-coordinate of the right nipple component. If it is determined that the difference between the first product and the second product is less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the third feature operation association is returned. If the Euclidean distance between the first chest electrode patch component and the right nipple component is less than or equal to the second preset error value, and the Euclidean distance between the second chest electrode patch component and the left nipple component is less than or equal to the second preset error value, then proceed with the subsequent calculation steps; otherwise, return to the fourth feature operation association result.

4. The multi-lead ECG electrode patch positioning method according to claim 3, characterized in that, The calculation of the feature operation association result of obtaining the upper body surface features and the ECG electrode patch feature state information also includes: The third product is obtained by multiplying the difference between the abscissa of the fourth chest electrode patch component and the abscissa of the left nipple component by the difference between the abscissa of the left nipple component and the abscissa of the right nipple component, and the fourth product is obtained by multiplying the difference between the ordinate of the fourth chest electrode patch component and the ordinate of the left nipple component by the difference between the ordinate of the left nipple component and the ordinate of the right nipple component. If the difference between the third product and the fourth product is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the fifth feature operation is returned. If the Euclidean distance between the fourth chest electrode patch component and the left nipple component is determined to be less than or equal to the second preset error value, then proceed with the subsequent calculation steps; otherwise, return to the sixth feature operation association result.

5. The multi-lead ECG electrode patch positioning method according to claim 4, characterized in that, The calculation of the feature operation association result of obtaining the upper body surface features and the ECG electrode patch feature state information also includes: Obtain the first straight-line distance from the center point of the third chest electrode patch to the line connecting the center points of the second and fourth chest electrode patches, the second straight-line distance from the center point of the third chest electrode patch to the center point of the second chest electrode patch, and the third straight-line distance from the center point of the third chest electrode patch to the center point of the fourth chest electrode patch. If it is determined that the difference between the first straight-line distance and the second preset error value is less than or equal to 0, then the subsequent calculation steps are executed; otherwise, the result of the seventh feature operation is returned. If the absolute value of the difference between the second straight-line distance and the third straight-line distance is less than or equal to the second preset error value, then the subsequent calculation steps are executed; otherwise, the result of the eighth feature operation is returned.

6. The multi-lead ECG electrode patch positioning method according to claim 5, characterized in that, The calculation of the feature operation association result of obtaining the upper body surface features and the ECG electrode patch feature state information also includes: The fifth product is obtained by multiplying the difference between the x-coordinate of the sixth chest electrode patch component and the x-coordinate of the left boundary component of the human body by multiplying the difference between the x-coordinate of the left nipple component and the x-coordinate of the right nipple component of the human body, and the sixth product is obtained by multiplying the difference between the y-coordinate of the sixth chest electrode patch component and the y-coordinate of the left boundary component of the human body by multiplying the difference between the y-coordinate of the left nipple component and the y-coordinate of the right nipple component of the human body. If the difference between the fifth product and the sixth product is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the ninth feature operation is returned. Obtain the seventh product obtained by multiplying the difference between the x-coordinate of the sixth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the y-coordinate of the left nipple component and the y-coordinate of the right nipple component, and the eighth product obtained by multiplying the difference between the x-coordinate of the sixth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the x-coordinate of the left nipple component and the y-coordinate of the right nipple component. If the difference between the seventh product and the eighth product is determined to be less than or equal to the first preset error value, then the subsequent calculation steps are executed; otherwise, the result of the tenth feature operation is returned.

7. The multi-lead ECG electrode patch positioning method according to claim 6, characterized in that, The calculation of the feature operation association result of obtaining the upper body surface features and the ECG electrode patch feature state information also includes: The ninth product is obtained by multiplying the difference between the x-coordinate of the fifth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the y-coordinate of the left nipple component and the x-coordinate of the right nipple component, and the tenth product is obtained by multiplying the difference between the y-coordinate of the fifth chest electrode patch component and the x-coordinate of the fourth chest electrode patch component by the difference between the y-coordinate of the left nipple component and the x-coordinate of the right nipple component. If the difference between the ninth product and the tenth product is determined to be less than the first preset error value, then the subsequent calculation steps are executed; otherwise, the eleventh feature operation association result is returned.

8. The method for positioning multi-lead ECG electrode patches according to any one of claims 3 to 5, characterized in that, The second preset error value is obtained by multiplying the average border length of the sum of the target frame length and the target frame width of the 10 electrode patches in the two-dimensional matrix by a preset ratio.

9. The multi-lead ECG electrode patch positioning method according to claim 1, characterized in that, The step of generating ECG electrode patch wearing status information based on the feature calculation association result, and determining the ECG electrode patch positioning guidance result based on the ECG electrode patch wearing status information, includes: If the result of the feature operation is determined to be a wearing error, the user is given feedback via voice or text regarding the incorrect electrode patch number and the electrode patch adjustment method. If the result of the feature operation is determined to be correct wearing information, a successful wearing prompt message will be sent to the user via voice or text.

Images