Point cloud data acquisition method for sternum corrector and sternum corrector
By acquiring and analyzing the key parameters of users, using neural network models and point cloud data technology, the problems of high cost of customization and radiation risk of existing sternum correction devices are solved, and efficient and accurate customization of sternum correction devices are achieved.
Patent Information
- Application Number
- CN202411845599.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-13
- Publication Date
- 2025-04-01
AI Technical Summary
The customization of existing sternum correctors is expensive and difficult, and traditional scanning technology has the problem of high radiation and causing damage to the patient's body.
By obtaining the user's first and second key parameters, the user's chest width, depth and sternum prediction curve are used to fit the user's chest cavity width, depth and sternum prediction curve, combined with the anterior chest contour curve, the user's sternum growth situation is established, and then exclusive point cloud data is generated to create customized sternum correction devices.
The precise manufacturing of customized sternum correctors is realized, which reduces customization costs and time, avoids radiation risks in traditional scanning technology, and improves the convenience and reliability of user experience.
Smart Images
Figure CN119741220A_ABST
Abstract
Description
Technical Field
[0001] The present invention mainly relates to the technical field of sternum correction, and in particular to a point cloud data acquisition method for a sternum corrector and a sternum corrector. Background Art
[0002] Currently, patients with chest deformity often need to wear sternal braces during the treatment stage. However, conventional models of sternal braces cannot achieve the best treatment effect for every patient, and hospital-customized products still have problems such as high cost and low efficiency, which have a great impact on the patient's recovery. Summary of the invention
[0003] The present invention provides a point cloud data acquisition method for a sternum corrector and a sternum corrector, aiming to solve the technical problems of high customization cost and great difficulty of the existing sternum corrector.
[0004] In a first aspect, the present application provides a method for acquiring point cloud data for a sternum corrector, comprising:
[0005] Obtaining a first key parameter, and fitting a chest contour curve of the user according to the first key parameter, wherein the first key parameter includes: a horizontal distance between the widest parts of the left and right sides of the chest of the user, and a vertical distance between the deepest part of the skin surface of the user and the deepest part of the sternum;
[0006] Obtaining a second key parameter, inputting the second key parameter into the neural network model, and obtaining a user's chest width prediction value, chest depth prediction value, and sternum prediction curve, wherein the second key parameter includes: the user's height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance, and sternum length;
[0007] Determine the sternum growth of the user based on the anterior chest contour curve, the thoracic width prediction value, the thoracic depth prediction value and the sternum prediction curve; and,
[0008] Point cloud data of a sternum corrector for a user is established based on the sternum growth.
[0009] In some embodiments, fitting a chest contour curve of a user according to the first key parameter includes:
[0010] Obtaining the type of curve used to form a front chest contour curve, wherein the curve types include: an elliptical curve and a parabola;
[0011] A chest elliptical curve is fitted according to the first key parameter, wherein the length of the horizontal axis of the chest elliptical curve is less than the horizontal distance between the widest parts of the left and right sides of the user's chest, and the length of the vertical axis of the chest elliptical curve is proportional to the vertical distance from the user's skin surface to the deepest part of the sternum;
[0012] A chest parabola having an intersection point with the chest elliptical curve is selected, and the chest elliptical curve and the chest parabola are spliced to obtain a chest contour curve of the user.
[0013] In some embodiments, before obtaining the second key parameter, the neural network model is trained in the following manner:
[0014] Obtaining a training sample set, wherein the training sample set includes a plurality of training samples, each training sample includes input features and scan data of a sample object, the input features include: height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance, and sternum length of the sample object, and the scan data includes chest surface data of the sample object acquired by a 3D scanner;
[0015] The neural network model is trained for multiple rounds of iterations using the training sample set. In one round of training:
[0016] Input features of at least one sample object are input into the neural network model to obtain prediction results of a corresponding round of training, and the neural network model is optimized and adjusted according to the prediction results of the round of training and the thorax surface data of at least one sample object.
[0017] In some embodiments, the neural network model is optimized and adjusted according to the prediction results of a round of training and the chest surface data of at least one sample subject, including:
[0018] Calculating a loss value between a prediction result of a round of training and chest surface data of at least one sample object, and optimizing and adjusting the neural network model according to the loss value;
[0019] The expression of the loss value is:
[0020]
[0021] N is the number of training samples, TW pred,i is the predicted value of chest width obtained for the i-th training sample, TD pred,i is the predicted value of chest depth for the i-th training sample, TW true,i is the true value of the chest width in the chest surface data in the i-th training sample, TD true,i is the true value of the chest cavity depth in the chest surface data in the i-th training sample.
[0022] In some embodiments, after calculating the loss value between the prediction result of a round of training and the chest surface data of at least one sample object, the method further includes:
[0023] Use the back-propagation algorithm to update the weight matrix of the neural network model;
[0024] The expression of the weight matrix is:
[0025]
[0026] is the weight matrix of the lth layer of the updated neural network model, is the weight matrix of the lth layer of the neural network model before updating, W (l) is the current weight of the lth layer of the neural network model, MSE is the loss value, and α is the learning rate, which is used to control the step size of the update.
[0027] In some embodiments, after calculating the loss value between the prediction result of a round of training and the chest surface data of at least one sample object, the method further includes:
[0028] The bias vector of the neural network model is updated using the back-propagation algorithm;
[0029] The expression of the bias vector is:
[0030]
[0031] is the bias vector of the lth layer of the updated neural network model, is the bias vector of the lth layer of the neural network model before updating, b (l) is the current bias of the lth layer of the neural network model, MSE is the loss value, and α is the learning rate, which is used to control the step size of the update.
[0032] In some embodiments, after establishing point cloud data of a sternum corrector for a user according to the sternum growth, the method further includes:
[0033] Converting point cloud data into a mesh model; and,
[0034] A solid model with a specified thickness is established according to the mesh model, wherein the solid model is used to print a sternum corrector for the user.
[0035] In a second aspect, the present application provides a point cloud data acquisition device for a sternum corrector, comprising:
[0036] A first acquisition module is used to acquire a first key parameter, and fit a chest contour curve of the user according to the first key parameter, wherein the first key parameter includes: a horizontal distance between the widest parts of the left and right sides of the chest of the user, and a vertical distance between the deepest part of the skin surface of the user and the deepest part of the sternum;
[0037] A second acquisition module is used to acquire a second key parameter, input the second key parameter into the neural network model, and obtain a user's chest width prediction value, chest depth prediction value and sternum prediction curve, wherein the second key parameter includes: the user's height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance and sternum length;
[0038] A first processing module is used to determine the sternum growth of the user according to the front chest contour curve, the thoracic cavity width prediction value, the thoracic cavity depth prediction value and the sternum prediction curve; and
[0039] The second processing module is used to establish point cloud data of a sternum corrector for the user according to the growth of the sternum.
[0040] In a third aspect, the present application provides a sternum corrector, which is manufactured using point cloud data obtained by any method of the first aspect described above.
[0041] In a fourth aspect, the present application provides an electronic device, comprising a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is called by the processor individually or collectively, the processor executes any method described in the first aspect.
[0042] Compared with the prior art, the present invention has the following advantages:
[0043] The present application provides a point cloud data acquisition method and a sternum corrector for a sternum corrector, which can accurately grasp the growth of the user's sternum through parameter fitting and neural network technology, so as to generate point cloud data of the sternum corrector exclusive to the user, solving the problems of high customization difficulty and high cost. At the same time, the customization process does not require traditional CT scanning data, which solves the problem of high radiation in the scanning stage and damage to the patient's body. The customization cycle is greatly shortened, and the user experience is more convenient and reliable. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] The accompanying drawings are included to provide a further understanding of the present application. They are included and constitute a part of the present application. The accompanying drawings illustrate embodiments of the present application and together with the present specification serve to explain the principles of the present application. In the accompanying drawings:
[0045] Figure 1 A flow chart of a method for acquiring point cloud data for a sternum corrector is shown;
[0046] Figure 2 A schematic diagram showing an exemplary front chest contour curve;
[0047] Figure 3 A schematic diagram showing an exemplary sternum growth situation;
[0048] Figure 4A schematic diagram showing an exemplary point cloud data;
[0049] Figure 5 A schematic diagram of a process for obtaining a sternal corrector is shown;
[0050] Figure 6 A schematic diagram of a point cloud data acquisition device for a sternum corrector is shown;
[0051] Figure 7 A schematic diagram of an exemplary electronic device is shown. DETAILED DESCRIPTION
[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.
[0053] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.
[0054] In addition, it should be noted that the use of words such as "first" and "second" to define components is only for the convenience of distinguishing the corresponding components. If not otherwise stated, the above words have no special meaning and cannot be understood as limiting the scope of protection of this application. In addition, although the terms used in this application are selected from well-known and commonly used terms, some terms mentioned in the specification of this application may be selected by the applicant at his or her discretion, and their detailed meanings are explained in the relevant parts of the description of this article. In addition, it is required to understand this application not only by the actual terms used, but also by the meaning implied by each term.
[0055] refer to Figure 1 , showing a flow chart of a method for acquiring point cloud data for a sternum corrector, comprising:
[0056] S101, obtaining a first key parameter, and fitting a chest contour curve of a user according to the first key parameter.
[0057] The first key parameter includes: the horizontal distance between the widest parts of the left and right sides of the user's chest, and the vertical distance between the deepest part of the user's skin surface and the deepest part of the sternum.
[0058] It will be understood that the first key parameter can be measured by the user in a hospital, or can be obtained by the user himself using a measuring tool (such as a ruler, etc.), and there is no limitation on this.
[0059] In some embodiments, fitting a chest contour curve of a user according to the first key parameter includes:
[0060] Get the type of curve used to form the chest contour curve, where the curve types include: elliptical curve and parabola.
[0061] A chest elliptical curve is fitted according to the first key parameter, wherein the length of the horizontal axis of the chest elliptical curve is less than the horizontal distance between the widest points on the left and right sides of the user's chest, and the length of the vertical axis of the chest elliptical curve is proportional to the vertical distance from the user's skin surface to the deepest part of the sternum.
[0062] A chest parabola having an intersection point with the chest elliptical curve is selected, and the chest elliptical curve and the chest parabola are spliced to obtain a chest contour curve of the user.
[0063] For example, see Figure 2 , a schematic diagram of an exemplary chest contour curve is shown. The chest contour curve is composed of a chest elliptical curve and a chest parabola. Assume that the horizontal distance between the widest parts of the left and right sides of the user's chest is L, and the vertical distance from the user's skin surface to the deepest part of the sternum is D. The length of the horizontal axis of the chest elliptical curve is a, which is slightly smaller than L, and the length of the vertical axis of the chest elliptical curve is b, which is 2D. Taking into account that the user's sternum is concave inward, a chest parabola can be smoothly connected to the two ends of the chest elliptical curve to form a complete chest contour curve.
[0064] S102, obtaining a second key parameter, inputting the second key parameter into a neural network model, and obtaining a predicted value of the chest width, a predicted value of the chest depth, and a predicted sternum curve of the user.
[0065] The second key parameters include: user's height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance and sternum length.
[0066] It will be understood that using the above representative second key parameter as the input of the neural network model can ensure that the neural network model has a fast prediction capability and also has a high accuracy. In addition, other key parameters such as age can also be added on this basis through clinical data, and there is no restriction on this.
[0067] In some embodiments, before obtaining the second key parameter, the neural network model is trained in the following manner:
[0068] A training sample set is obtained, wherein the training sample set includes a plurality of training samples, each training sample includes input features and scan data of a sample object, the input features include: height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance, and sternum length of a sample object, and the scan data includes chest surface data of a sample object collected by a 3D scanner.
[0069] The neural network model is trained for multiple rounds of iterations using the training sample set. In one round of training:
[0070] Input features of at least one sample object are input into the neural network model to obtain prediction results of a corresponding round of training, and the neural network model is optimized and adjusted according to the prediction results of the round of training and the thorax surface data of at least one sample object.
[0071] For example, referring to Table 1, some input features in the training samples corresponding to the sample objects AE are shown, which are:
[0072] Table 1
[0073]
[0074] During the training process, the input features of the sample object AE are input into the neural network model in batches or all at once to obtain the prediction results of the corresponding round of training. The prediction results include the chest depth prediction value and chest width prediction value of the neural network model for the sample object AE. See Table 2, which shows the chest depth prediction value and chest width prediction value of the sample comparison AE, which are:
[0075] Table 2
[0076]
[0077]
[0078] It will be understood that the neural network model can adopt a full neural network. The full neural network includes an input layer, multiple hidden layers and an output layer, and each hidden layer includes multiple neurons. The layers (hidden layers or output layers) are connected to each other through weight matrices and bias vectors, and the expression of the output of the lth layer can be shown in the following equations (1) and (2):
[0079] z (l) =W (l) a (l-1) +b (l) (1)
[0080] a(l) =δ(z (l) ) (2)
[0081] a (l) is the output of the lth layer, W (l) is the weight matrix of the lth layer, b (l) is the bias vector of the lth layer, a (l-1) is the output of the l-1th layer, and δ(.) is the activation function.
[0082] In some embodiments, the neural network model is optimized and adjusted according to the prediction results of a round of training and the chest surface data of at least one sample subject, including:
[0083] The loss value between the prediction result of a round of training and the chest surface data of at least one sample object is calculated, and the neural network model is optimized and adjusted according to the loss value.
[0084] The expression of the loss value is shown in formula (3):
[0085]
[0086] N is the number of training samples, TW pred,i is the predicted value of chest width obtained for the i-th training sample, TD pred,i is the predicted value of chest depth for the i-th training sample, TW true,i is the true value of the chest width in the chest surface data in the i-th training sample, TD true,i is the true value of the chest cavity depth in the chest surface data in the i-th training sample.
[0087] It will be understood that, in addition to the mean square error loss value represented by equation (3), the neural network model may also use other loss functions to calculate the corresponding damage value, and there is no limitation to this.
[0088] In some embodiments, after calculating the loss value between the prediction result of a round of training and the chest surface data of at least one sample object, the method further includes:
[0089] The back-propagation algorithm is used to update the weight matrix of the neural network model.
[0090] The expression of the weight matrix is shown in formula (4):
[0091]
[0092] is the weight matrix of the lth layer of the updated neural network model, is the weight matrix of the lth layer of the neural network model before updating, W (l)is the current weight of the lth layer of the neural network model, MSE is the loss value, and α is the learning rate, which is used to control the step size of the update;
[0093] In some embodiments, after calculating the loss value between the prediction result of a round of training and the chest surface data of at least one sample object, the method further includes:
[0094] The back-propagation algorithm is used to update the bias vector of the neural network model.
[0095] The expression of the bias vector is shown in formula (5):
[0096]
[0097] is the bias vector of the lth layer of the updated neural network model, is the bias vector of the lth layer of the neural network model before updating, b (l) is the current bias of the lth layer of the neural network model.
[0098] In some embodiments, the back propagation algorithm may use a gradient descent method or other algorithms without limitation.
[0099] S103, determining the sternum growth of the user according to the anterior chest contour curve, the thoracic cavity width prediction value, the thoracic cavity depth prediction value and the sternum prediction curve.
[0100] For example, see Figure 3 , showing a schematic diagram of an exemplary sternum growth. The sternum prediction curve output by the neural network model and the anterior chest contour curve fitted in step S101 can eventually form a plane curve with a certain inclination, and the thoracic cavity width prediction value and the thoracic cavity depth prediction value can be used to reflect their changes in three-dimensional space, so as to accurately obtain the sternum growth of the user.
[0101] S104, establishing point cloud data of a sternum corrector for the user according to the growth of the sternum.
[0102] For example, see Figure 4 , shows a schematic diagram of an exemplary point cloud data. According to the growth of the user's sternum, the anterior chest contour curve, the predicted value of the thoracic cavity width, the predicted value of the thoracic cavity depth and the predicted sternum curve are converted into three-dimensional point cloud data. The point cloud data can reflect the changes of the user's sternum in three-dimensional space.
[0103] In some embodiments, after establishing point cloud data of a sternum corrector for a user according to the sternum growth, the method further includes:
[0104] Convert point cloud data into mesh models.
[0105] For example, the point cloud data is packaged to obtain a corresponding mesh model, which is composed of a plurality of triangular or quadrilateral facets and can better represent the shape and surface features of the object.
[0106] A solid model with a specified thickness is established according to the mesh model, wherein the solid model is used to print a sternum corrector for the user.
[0107] For example, the user can specify the direction and thickness of the thickening, so that the facets in the mesh model can be trimmed to increase the reliability of the model.
[0108] It will be understood that after obtaining the physical model, the user can also optimize and adjust the physical model according to actual needs, for example, removing redundant face pieces, smoothing the surface, repairing holes, etc., so as to further improve the quality of the physical model and increase the comfort and accuracy of the sternum corrector manufactured subsequently.
[0109] See also Figure 5 , showing a flow chart for obtaining a sternum corrector. First, the user can obtain the first key parameter and the second key parameter by self-measurement, and then input the second key parameter into the neural network model, and obtain the user's chest width prediction value, chest depth prediction value and sternum prediction curve through the neural network model to determine the user's sternum growth. Subsequently, the sternum growth is used to establish point cloud data for the user's sternum corrector. Further, the point cloud data is encapsulated, thickened, optimized, and other processes are performed to obtain a physical model that meets the user's needs. The user's exclusive sternum corrector can be obtained based on the physical model through 3D printing technology. Based on the above method, the present embodiment can generate point cloud data of a sternum corrector that is exclusive to the user, solving the problems of high customization difficulty and high cost. The customization process does not require traditional CT scan data, which solves the problem of large radiation in the scanning stage and damage to the patient's body. The customization cycle is greatly shortened, and the user experience is more convenient and reliable.
[0110] For further information, see Figure 6 , showing a schematic diagram of a point cloud data acquisition device for a sternum corrector, comprising:
[0111] The first acquisition module 601 is used to acquire the first key parameter and fit the user's chest contour curve according to the first key parameter, wherein the first key parameter includes: the horizontal distance between the widest parts of the left and right sides of the user's chest, and the vertical distance between the skin surface of the user to the deepest part of the sternum.
[0112] The second acquisition module 602 is used to acquire the second key parameter, input the second key parameter into the neural network model, and obtain the user's chest width prediction value, chest depth prediction value and sternum prediction curve, wherein the second key parameter includes: the user's height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance and sternum length.
[0113] The first processing module 603 is used to determine the sternum growth of the user according to the anterior chest contour curve, the thoracic cavity width prediction value, the thoracic cavity depth prediction value and the sternum prediction curve.
[0114] The second processing module 604 is used to establish point cloud data of a sternum corrector for the user according to the sternum growth condition.
[0115] In some embodiments, the first acquisition module 601 is used to:
[0116] Obtaining the type of curve used to form a front chest contour curve, wherein the curve types include: an elliptical curve and a parabola;
[0117] A chest elliptical curve is fitted according to the first key parameter, wherein the length of the horizontal axis of the chest elliptical curve is less than the horizontal distance between the widest points of the left and right sides of the user's chest, and the length of the vertical axis of the chest elliptical curve is proportional to the vertical distance from the user's skin surface to the deepest part of the sternum;
[0118] A chest parabola having an intersection point with the chest elliptical curve is selected, and the chest elliptical curve and the chest parabola are spliced to obtain a chest contour curve of the user.
[0119] In some embodiments, the second acquisition module 602 is further used to:
[0120] Obtaining a training sample set, wherein the training sample set includes a plurality of training samples, each training sample includes input features and scan data of a sample object, the input features include: height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance, and sternum length of the sample object, and the scan data includes chest surface data of the sample object acquired by a 3D scanner;
[0121] The neural network model is trained for multiple rounds of iterations using the training sample set. In one round of training:
[0122] Input features of at least one sample object are input into the neural network model to obtain prediction results of a corresponding round of training, and the neural network model is optimized and adjusted according to the prediction results of the round of training and the thorax surface data of at least one sample object.
[0123] In some embodiments, the second acquisition module 602 is further used to:
[0124] The loss value between the prediction result of a round of training and the chest surface data of at least one sample object is calculated, and the neural network model is optimized and adjusted according to the loss value.
[0125] The expression of the loss value is:
[0126]
[0127] N is the number of training samples, TW pred,i is the predicted value of chest width obtained for the i-th training sample, TD pred,i is the predicted value of chest depth for the i-th training sample, TW true,i is the true value of the chest width in the chest surface data in the i-th training sample, TD true,i is the true value of the chest cavity depth in the chest surface data in the i-th training sample.
[0128] In some embodiments, the second acquisition module 602 is further used to:
[0129] Use the back-propagation algorithm to update the weight matrix of the neural network model;
[0130] The expression of the weight matrix is:
[0131]
[0132] is the weight matrix of the lth layer of the updated neural network model, is the weight matrix of the lth layer of the neural network model before updating, W (l) is the current weight of the lth layer of the neural network model, MSE is the loss value, and α is the learning rate, which is used to control the step size of the update.
[0133] In some embodiments, the second acquisition module 602 is further used to:
[0134] The bias vector of the neural network model is updated using the back-propagation algorithm;
[0135] The expression of the bias vector is:
[0136]
[0137] is the bias vector of the lth layer of the updated neural network model, is the bias vector of the lth layer of the neural network model before updating, b (l) is the current bias of the lth layer of the neural network model.
[0138] In some embodiments, the second processing module 604 is further configured to:
[0139] Converting point cloud data into a mesh model; and,
[0140] A solid model with a specified thickness is established according to the mesh model, wherein the solid model is used to print a sternum corrector for the user.
[0141] Furthermore, an exemplary embodiment of the present application also provides a sternum corrector, which is manufactured using the point cloud data in the point cloud data acquisition method for the sternum corrector mentioned above.
[0142] Further, if Figure 7 An exemplary embodiment of the present application further provides an electronic device, including a memory 701 and a processor 702, wherein the memory 701 stores instructions, and when the instructions are executed by the processor 702, the processor 702 executes the method as described above.
[0143] It should be understood that the processor mentioned in the embodiments of the present application may be a CPU, or other general-purpose processors, DSPs, ASICs, FPGAs or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
[0144] It should also be understood that the memory mentioned in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory, dynamic random access memory, synchronous dynamic random access memory, double data rate synchronous dynamic random access memory, enhanced synchronous dynamic random access memory, synchronously connected dynamic random access memory, and direct memory bus random access memory.
[0145] The present application also provides a storage medium storing computer program instructions, which, when executed by a processor of a computer, enables the computer to execute the steps of any of the methods mentioned above.
[0146] A computer-readable medium may include a propagated data signal containing computer program code, such as in baseband or as part of a carrier wave. The propagated signal may have a variety of manifestations, including electromagnetic, optical, etc., or a suitable combination. A computer-readable medium may be any computer-readable medium other than a computer-readable storage medium, which may be connected to an instruction execution system, device or apparatus to communicate, propagate or transmit a program for use. The program code on the computer-readable medium may be propagated via any suitable medium, including radio, cable, fiber optic cable, radio frequency signal, or similar medium, or any combination of the above mediums.
[0147] The basic concepts have been described above. Obviously, for those skilled in the art, the above invention disclosure is only used as an example and does not constitute a limitation of the present application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements and amendments to the present application. Such modifications, improvements and amendments are suggested in the present application, so such modifications, improvements and amendments still belong to the spirit and scope of the exemplary embodiments of the present application.
[0148] At the same time, the present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present application can be appropriately combined.
[0149] Some aspects of the present application may be performed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may be referred to as "data blocks", "modules", "engines", "units", "components" or "systems". The processor may be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DAPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, or combinations thereof. In addition, various aspects of the present application may be expressed as computer products located in one or more computer-readable media, which include computer-readable program codes. For example, computer-readable media may include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, tapes ...), optical disks (e.g., compact disks CDs, digital versatile disks DVDs ...), smart cards, and flash memory devices (e.g., cards, sticks, key drives ...).
[0150] A computer-readable medium may include a propagated data signal containing computer program code, such as in baseband or as part of a carrier wave. The propagated signal may have a variety of manifestations, including electromagnetic, optical, etc., or a suitable combination. A computer-readable medium may be any computer-readable medium other than a computer-readable storage medium, which may be connected to an instruction execution system, device or apparatus to communicate, propagate or transmit a program for use. The program code on the computer-readable medium may be propagated via any suitable medium, including radio, cable, fiber optic cable, radio frequency signal, or similar medium, or any combination of the above mediums.
[0151] Similarly, it should be noted that in order to simplify the description of the disclosure of this application and thus help understand one or more embodiments of the invention, in the above description of the embodiments of this application, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the object of this application are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.
[0152] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of the present application are approximate values, in specific embodiments, the setting of such numerical values is as accurate as possible within the feasible range.
[0153] Although the present application has been described with reference to the current specific embodiments, ordinary technicians in this technical field should recognize that the above embodiments are only used to illustrate the present application, and various equivalent changes or substitutions may be made without departing from the spirit of the present application. Therefore, as long as the changes and modifications to the above embodiments are within the essential spirit of the present application, they will fall within the scope of the claims of the present application.
Claims
1. A point cloud data acquisition method for a sternum corrector, characterized in that: include: Acquire a first key parameter, and fit a chest contour curve of the user according to the first key parameter, wherein the first key parameter includes: a horizontal distance between the widest parts of the left and right sides of the chest of the user, and a vertical distance between the deepest part of the skin surface of the user and the deepest part of the sternum; Acquire a second key parameter, input the second key parameter into a neural network model, and obtain a predicted value of the chest width, a predicted value of the chest depth, and a predicted sternum curve of the user, wherein the second key parameter includes: the height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance, and sternum length of the user; Determining the sternum growth of the user according to the anterior chest contour curve, the thoracic cavity width prediction value, the thoracic cavity depth prediction value and the sternum prediction curve; and, Point cloud data of a sternum corrector for the user is established according to the sternum growth condition.
2. The method according to claim 1, characterized in that The step of fitting the chest contour curve of the user according to the first key parameter includes: Acquire the type of curve used to form the front chest contour curve, wherein the curve type includes: an elliptical curve and a parabola; Fitting a chest elliptical curve according to the first key parameter, wherein the length of the horizontal axis of the chest elliptical curve is less than the horizontal distance between the widest parts of the left and right sides of the chest of the user, and the length of the vertical axis of the chest elliptical curve is proportional to the vertical distance from the skin surface of the user to the deepest part of the sternum; A chest parabola having an intersection point with the chest elliptical curve is selected, and the chest elliptical curve and the chest parabola are spliced to obtain a chest contour curve of the user.
3. The method according to claim 1 or 2, characterized in that Before obtaining the second key parameter, the neural network model is trained in the following manner: Acquire a training sample set, wherein the training sample set includes a plurality of training samples, each training sample includes input features and scan data of a sample object, the input features include: height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance, and sternum length of the sample object, and the scan data includes chest surface data of the sample object acquired by a 3D scanner; The neural network model is trained for multiple rounds of iterations using the training sample set, wherein in one round of training: Input features of at least one of the sample objects are input into the neural network model to obtain prediction results corresponding to the one round of training, and the neural network model is optimized and adjusted based on the prediction results of the one round of training and the chest surface data of at least one of the sample objects.
4. The method according to claim 3, characterized in that The optimizing and adjusting the neural network model according to the prediction result of the training round and the chest surface data of at least one of the sample objects includes: Calculating a loss value between the prediction result of the training round and the chest surface data of at least one of the sample objects, and optimizing and adjusting the neural network model according to the loss value; Among them, the expression of the loss value MSE is: N is the number of training samples, TW pred,i is the predicted value of chest width obtained for the i-th training sample, TD pred,i is the predicted value of chest depth obtained for the i-th training sample, TW true,i is the true value of the chest width in the chest surface data in the i-th training sample, TD true,i is the true value of the chest cavity depth in the chest surface data in the i-th training sample.
5. The method according to claim 4, characterized in that After calculating the loss value between the prediction result of the training round and the chest surface data of at least one of the sample objects, the method further includes: Using a back propagation algorithm to update the weight matrix of the neural network model; The expression of the weight matrix is: is the weight matrix of the lth layer of the neural network model after update, is the weight matrix of the lth layer of the neural network model before updating, W (l) is the current weight of the lth layer of the neural network model, MSE is the loss value, and α is the learning rate, which is used to control the step size of the update.
6. The method according to claim 4, characterized in that After calculating the loss value between the prediction result of the training round and the chest surface data of at least one of the sample objects, the method further includes: Updating the bias vector of the neural network model using a back propagation algorithm; The expression of the bias vector is: is the bias vector of the updated lth layer of the neural network model, is the bias vector of the lth layer of the neural network model before updating, b (l) is the current bias of the lth layer of the neural network model, MSE is the loss value, and α is the learning rate, which is used to control the step size of the update.
7. The method according to claim 1, characterized in that After the point cloud data of the sternum corrector for the user is established according to the sternum growth, the method further includes: converting the point cloud data into a mesh model; and, A solid model of a specified thickness is established according to the grid model, wherein the solid model is used to print a sternum corrector for the user.
8. A point cloud data acquisition device for a sternum corrector, characterized in that: include: A first acquisition module is used to acquire a first key parameter, and fit a chest contour curve of the user according to the first key parameter, wherein the first key parameter includes: a horizontal distance between the widest parts of the left and right sides of the chest of the user, and a vertical distance between the deepest part of the skin surface of the user and the deepest part of the sternum; A second acquisition module is used to acquire second key parameters, input the second key parameters into the neural network model, and obtain the user's chest width prediction value, chest depth prediction value and sternum prediction curve, wherein the second key parameters include: the user's height, weight, chest circumference, breast distance, clavicle distance, shoulder blade distance and sternum length; A first processing module is used to determine the sternum growth of the user according to the anterior chest contour curve, the thoracic cavity width prediction value, the thoracic cavity depth prediction value and the sternum prediction curve; and The second processing module is used to establish point cloud data of a sternum corrector for the user according to the sternum growth condition.
9. A sternum corrector, characterized in that: The sternum corrector is manufactured using the point cloud data obtained by the method according to any one of claims 1-7.
10. An electronic device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is called by the processor individually or collectively, the processor executes the method according to any one of claims 1 to 7.