Human body posture reconstruction and joint point positioning method and device and storage medium

Through the pressure smart clothing, obtain human posture data and calculate the joint position, and generate posture graphics, solving the limitations of human posture reconstruction in the existing technology, realizing accurate posture reconstruction at any time and place, improving the comfort and use Convenience.

CN118967929APending Publication Date: 2024-11-15UNIV OF SCI & TECH OF CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411030203.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The prior art cannot achieve long-term, anywhere, human posture reconstruction, and due to the limitations of equipment location and comfort, it cannot meet the application needs in daily life.

Method used

Use pressure smart clothing to obtain pressure data of human body posture, determine the position information of the joint node through computing devices and generate posture graphics, reducing interference to the human body, improving comfort and application convenience.

Benefits of technology

Accurate human posture reconstruction at any time and place is achieved, improving the comfort and acceptance of use, and reducing the limitations of posture reconstruction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118967929A_ABST
    Figure CN118967929A_ABST
Patent Text Reader

Abstract

The invention discloses a body attitude reconstruction and joint point positioning method and device and a storage medium. The method comprises the following steps: acquiring pressure data generated by a pressure smart garment in response to a posture of a target human body; determining first position information corresponding to a plurality of joint points of the target human body according to the pressure data; and generating a graph corresponding to the posture according to the first position information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device and storage medium for human body posture reconstruction and joint point positioning. Background Art

[0002] Posture estimation obtains data through devices such as cameras or inertial sensors, and then calculates the human body's ball-and-stick model (or human skeleton, which uses joints and their connections to represent posture movements) or volume model (using body surface grids to represent human shape and posture movements). It has been widely used in film animation, virtual reality, human-computer interaction, and sports analysis. However, due to the underlying method, posture estimation is still often limited to fixed space or short time, resulting in the unmet needs of those supported by long-term data.

[0003] At present, most posture estimation applications are still confined to fixed spaces and short periods of time (such as live broadcast rooms or studios), which is due to the underlying limitations of posture estimation methods. The two existing mature methods: visual methods and inertial methods, cannot be directly applied in daily life for a long time.

[0004] Among them, the visual solution uses a monocular or multi-camera to obtain the coordinates of key points based on the principle of triangulated imaging, or reconstructs a high-precision human body volume model through a camera array; the inertial solution wears an inertial sensor at a specific position on the human body, obtains the sensor coordinates based on quaternion integration, and then reassembles the human skeleton. They perform well in well-equipped venues, but the location of the equipment and the timing of use limit their downstream applications in daily life. The visual method not only has problems such as lighting and occlusion, but also privacy issues in private and sensitive venues such as bedrooms and hospitals. If you want to cover all the scenes reached in daily life, there are also human and material costs for setting up and maintaining camera equipment. The inertial solution requires wearing a hard inertial sensor unit, which has low comfort during long-term use (for example, the back sensor will hurt when sitting or lying down, and frequent putting on and taking off will seriously affect the willingness to use. As early as 2002, scholars studied wearability based on whether it is suitable for daily activities, and there are not many parts of the human body that are suitable for wearing hard devices all day long). If sensors are only set on the wrists and head, the fitting accuracy will be seriously affected because only local data is available. In addition, there is the problem of requiring additional calibration and the problem of trajectory drift caused by accumulated errors during the integration process.

[0005] In short, there is no long-term posture acquisition method that can support "anytime, anywhere", and the limitations of posture reconstruction are also quite significant.

[0006] With respect to the technical problem of high limitations in reconstructing human posture in the above-mentioned prior art, no effective solution has been proposed yet. Summary of the invention

[0007] The embodiments of the present application provide a method, device and storage medium for human body posture reconstruction and joint point positioning, so as to at least solve the technical problem of high limitations in reconstructing human body posture in the prior art.

[0008] According to one aspect of an embodiment of the present application, a method for reconstructing a human body posture is provided, including: obtaining pressure data generated by a pressure smart garment in response to the posture of a target human body; determining first position information corresponding to multiple joints of the target human body based on the pressure data; and generating a graphic corresponding to the posture based on the first position information.

[0009] According to another aspect of an embodiment of the present application, a method for reconstructing a human body posture is also provided, comprising: acquiring pressure data generated by the pressure smart clothing in response to the posture of a target human body; and generating a graphic corresponding to the posture according to the pressure data.

[0010] According to another aspect of an embodiment of the present application, a method for locating human joints is also provided, including: obtaining pressure data generated by the pressure smart clothing in response to the posture of the target human body; and determining position information corresponding to multiple joints of the target human body based on the pressure data.

[0011] According to another aspect of an embodiment of the present application, a human body posture reconstruction device is also provided, including: a data acquisition module, used to acquire pressure data generated by the pressure smart clothing in response to the posture of the target human body; an information determination module, used to determine first position information corresponding to multiple joints of the target human body based on the pressure data; and a graphics generation module, used to generate a graphics corresponding to the posture based on the first position information.

[0012] According to another aspect of an embodiment of the present application, a human body posture reconstruction device is also provided, including: a processor; and a memory connected to the processor, for providing the processor with instructions for processing the following processing steps: obtaining pressure data generated by the pressure smart clothing in response to the posture of the target human body; determining first position information corresponding to multiple joints of the target human body based on the pressure data; and generating a graphic corresponding to the posture based on the first position information.

[0013] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented.

[0014] According to another aspect of the embodiments of the present application, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0015] In the embodiment of the present application, the pressure smart clothing generates pressure data, and the computing device determines the position information corresponding to multiple joints of the target human body according to the pressure data generated by the pressure smart clothing, and then generates a graph corresponding to the posture according to the position information, thereby reconstructing the posture. Compared with the prior art, the present technical solution generates pressure data corresponding to the target human body through the pressure smart clothing, reduces interference to the target human body, improves the comfort and acceptance of use, and reduces the limitation of not being able to obtain pressure data anytime and anywhere to reconstruct the posture of the target human body. This solves the technical problem of high limitations in reconstructing human body posture in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 is a hardware structure block diagram of a computing device for implementing the method according to Embodiment 1 of the present application;

[0018] Figure 2 is a schematic diagram of a human body posture reconstruction system according to Example 1 of the present application;

[0019] Figure 3A is a schematic diagram of a pressure smart garment according to Example 1 of the present application;

[0020] Figure 3B is a schematic diagram of a target human body wearing a pressure smart garment according to Example 1 of the present application;

[0021] Figure 4 is a flow chart of a method for reconstructing a human body posture according to the first aspect of Embodiment 1 of the present application;

[0022] Figure 5A is a schematic diagram of a pressure image corresponding to a top of a pressure smart garment according to Embodiment 1 of the present application;

[0023] Figure 5B is a schematic diagram of a pressure image corresponding to the trousers of the pressure smart clothing according to Embodiment 1 of the present application;

[0024] Figure 6 is a schematic diagram of a graphic corresponding to a posture according to Embodiment 1 of the present application;

[0025] Figure 7 is a schematic diagram of a process for generating first location information according to Embodiment 1 of the present application;

[0026] Figure 8is a schematic diagram of a process for generating a feature vector according to Embodiment 1 of the present application;

[0027] Fig. 9 is a schematic diagram of a process for processing pressure information according to the feature extraction model described in Example 1 of the present application;

[0028] Fig.10 is a schematic diagram of the number of joint points according to Example 1 of the present application;

[0029] Fig.11A is a schematic diagram of the first pressure data according to Example 1 of the present application;

[0030] Fig. 11B is a schematic diagram of the second pressure data according to Example 1 of the present application;

[0031] Fig. 12A is a schematic diagram of a frontal pressure image after the threshold is adjusted according to Example 1 of the present application;

[0032] Fig. 12B is a schematic diagram of a reverse pressure image after the threshold is adjusted according to Example 1 of the present application;

[0033] Fig.13 is a schematic diagram of a sequential flow of human body posture reconstruction according to Embodiment 1 of the present application;

[0034] Fig.14 is a schematic diagram of a human body posture reconstruction device according to Embodiment 2 of the present application; and

[0035] Fig.15 It is a schematic diagram of the human body posture reconstruction device described in Example 3 of the present application. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only embodiments of a part of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present application.

[0037] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0038] Example 1

[0039] According to this embodiment, a method embodiment of a human body posture reconstruction method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0040] The method embodiment provided in this embodiment can be executed in a mobile terminal, a computer terminal, a server or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computing device for implementing a human posture reconstruction method. Figure 1 As shown, the computing device may include one or more processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory for storing data, and a transmission device for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0041] It should be noted that the one or more processors and / or other data processing circuits described above may generally be referred to herein as "data processing circuitry". The data processing circuitry may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuitry may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computing device. As described in the embodiments of the present application, the data processing circuitry serves as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0042] The memory can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the human posture reconstruction method in the embodiment of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, the human posture reconstruction method of the above-mentioned application is realized. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computing device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0043] The transmission device is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computing device. In one example, the transmission device includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device can be a radio frequency (Radio Frequency, RF) module, which is used to communicate with the Internet wirelessly.

[0044] The display may be, for example, a touch screen liquid crystal display (LCD) that may enable a user to interact with a user interface of the computing device.

[0045] It should be noted that, in some optional embodiments, the above Figure 1 The computing device shown may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware elements and software elements. Figure 1 This is merely one example of a particular embodiment and is intended to illustrate the types of components that may be present in the computing devices described above.

[0046] Figure 2is a schematic diagram of a human body posture reconstruction system according to this embodiment. Figure 2 As shown, the system includes: a pressure smart garment 10 and a computing device 20.

[0047] The pressure smart clothing 10 is used to collect pressure data of the target human body wearing the pressure smart clothing and send the pressure data to the computing device 20. The computing device 20 is used to process the pressure data and generate a graph corresponding to the posture of the target human body according to the pressure data, that is, to achieve reconstruction of the human body posture.

[0048] Figure 3A shows a schematic diagram of a pressure smart garment, Figure 3B A schematic diagram of a target human body wearing a pressure smart garment is shown. Figure 3A As shown, the pressure smart garment 10 may include, for example, a top and pants. The sensing area of ​​the pressure smart garment 10 is composed of transverse sensing stripes ( Figure 3A The gray stripes shown in FIG. 1 ) and the longitudinal sensing stripes (located on the inner side of the pressure smart garment 10, Figure 3A The horizontal sensing stripes and the vertical sensing stripes are composed of conductive materials with pressure-sensitive properties, and their intersection is a pressure sampling point. The pressure smart clothing 10 covers more than 80% of the human body surface and contains a total of 1952 pressure sampling points. The driving circuit of the pressure smart clothing 10 periodically applies voltage to the pressure sampling point, and calculates the pressure on the sampling point by measuring the partial pressure of the pressure sampling point. Thus, the pressure smart clothing 10 can transmit the pressure data of each pressure sampling point to the computing device 20.

[0049] Under the above operating environment, according to the first aspect of this embodiment, a method for reconstructing human posture is provided. Figure 2 The computing device 20 shown in FIG. Figure 4 A schematic diagram showing the process of the method is shown in FIG. Figure 4 As shown, the method includes:

[0050] S402: Acquire pressure data generated by the pressure smart clothing in response to the posture of the target human body;

[0051] S404: Determine first position information corresponding to multiple joints of the target human body according to the pressure data; and

[0052] S406: Generate a graphic corresponding to the posture according to the first position information.

[0053] Specifically, when the target human body wears the pressure smart clothing 10 and poses a certain posture, the multiple pressure sampling points of the pressure smart clothing 10 respond to the posture and generate pressure data corresponding to the posture. The pressure data generated by the pressure smart clothing 10 can be, for example, a pressure image generated according to the distribution of the pressure sampling points (see Figure 5A as well as Figure 5B As shown). The posture may be, for example, sitting with legs hugged, turning, and lunging. Then, the pressure smart garment 10 transmits the generated pressure image to the computing device 20, so that the computing device 20 obtains the pressure image from the pressure smart garment 10 (S402).

[0054] Then, the computing device 20 determines the position information (i.e., the first position information) of the multiple joints of the target human body corresponding to the posture according to the pressure image (S404). The multiple joints include, for example (but not limited to): 15 joints such as the center of the buttocks, the neck, the nose, the left shoulder, the right shoulder, the left elbow, the right elbow, the left wrist, the right wrist, the left hip, the right hip, the left knee, the right knee, the left ankle, and the right ankle. Thus, the computing device 20 can determine the position information (i.e., the first position information) of the multiple joints corresponding to the posture according to the pressure image. Specifically, the position information of the multiple joints can be represented by the three-dimensional coordinates of the multiple joints.

[0055] Further, refer to Figure 6 As shown, the computing device 20 draws a graphic corresponding to the gesture according to the determined position information of the multiple joint points (i.e., the first position information) (S406). Specifically, the computing device 20 can, for example, draw the multiple joint points on the display according to the determined position information of the multiple joint points, and then the computing device 20 draws the lines between the multiple joint points, thereby drawing a graphic corresponding to the gesture.

[0056] As described in the background technology, there is no long-term posture acquisition method that can support "anytime, anywhere", and the limitations of posture reconstruction are also quite significant. In response to the technical problems described above, through the technical solutions of the embodiments of the present application, the pressure smart clothing generates pressure data, and the computing device determines the position information corresponding to multiple joints of the target human body according to the pressure data generated by the pressure smart clothing, and then generates graphics corresponding to the posture according to the position information, thereby reconstructing the posture. Compared with the prior art, the technical solution generates pressure data corresponding to the target human body through the pressure smart clothing, reduces interference to the target human body, improves the comfort and acceptance of use, and reduces the limitation of not being able to obtain pressure data anytime and anywhere to reconstruct the posture of the target human body. This solves the technical problem of high limitations in reconstructing human body posture in the prior art.

[0057] Optionally, the operation of determining first position information corresponding to multiple joints of the target human body based on pressure data includes: determining rotation information of the multiple joints based on the pressure data; and determining the first position information of the multiple joints based on the rotation information using a forward dynamics algorithm.

[0058] Specifically, after receiving the pressure data sent by the pressure smart clothing 10, the computing device 20 determines the rotation information of multiple joints according to the pressure data. The rotation information is used to indicate the data information of multiple joints of the target human body after rotation relative to the previous joint in the current posture. For example, if the joint point "left elbow" is the previous joint of the joint point "left wrist", the rotation information of the joint point "left wrist" indicates the data information of the "left wrist" after rotation relative to the previous joint point "left elbow" in the current posture.

[0059] More specifically, refer to Figure 7 As shown, the computing device 20 determines the rotation matrix (i.e., rotation information) corresponding to the multiple joint points according to the pressure data corresponding to the multiple joint points. Each joint point corresponds to a rotation matrix, including rotation matrices 1 to n. In this embodiment, the computing device 20 pre-sets 15 joint points, including the center of the buttocks, the neck, the nose, the left shoulder, the right shoulder, the left elbow, the right elbow, the left wrist, the right wrist, the left hip, the right hip, the left knee, the right knee, the left ankle, and the right ankle, then n=15.

[0060] Furthermore, the computing device 20 uses a forward dynamics algorithm to determine first position information of each joint point according to rotation matrices 1 to n corresponding to the plurality of joint points, for example, position information 1 to n respectively.

[0061] Therefore, this technical solution can accurately estimate the human body posture by directly deriving the rotation information of the joint points based on the pressure data, and then using the forward dynamics algorithm to calculate the position information, thereby ensuring the accuracy of the calculated position information.

[0062] Optionally, the operation of determining the rotation information of multiple joint points based on pressure data includes: performing feature extraction on the pressure data using a neural network-based feature extraction model to generate six-dimensional feature vectors corresponding to the multiple joint points respectively, wherein the feature vectors are associated with the rotation information of the corresponding joint points; and determining a corresponding rotation matrix as the rotation information based on the feature vectors through a preset mapping algorithm.

[0063] Specifically, refer to Figure 8 As shown, the computing device 20 inputs the pressure data into a pre-set neural network-based feature extraction model, extracts features from the pressure data through the feature extraction model, and generates six-dimensional feature vectors corresponding to multiple joint points. The feature vectors are feature vectors 1 to n (in this embodiment, n = 15). For example, feature vectors 1 to n are represented by [r6d1, r6d2, ..., r6d n ] T . The six-dimensional feature vector is, for example, a 6×1 r6d vector, for example:

[0064] r6d1=[r 1,1 ,r 1,2 ,r 1,3 ,r 1,4 ,r 1,5 ,r 1,6 ];

[0065] r6d2=[r 2,1 ,r 2,2 ,r 2,3 ,r 2,4 ,r 2,5 ,r 2,6 ];

[0066] …

[0067] r6d n =[r n,1 ,r n,2 ,r n,3 ,r n,4 ,r n,5 ,r n,6 ].

[0068] Furthermore, the computing device 20 converts the feature vectors corresponding to each joint point into a three-dimensional rotation matrix R (ie, rotation information) using a preset mapping algorithm. The mapping algorithm may be a Gram-Schmidt orthogonalization algorithm:

[0069]

[0070] Where r6d[0:3] represents the feature vector r6d i =[r i,1 ,r i,2 ,r i,3 ,r i,4 ,r i,5 ,r i,6 ] i,1 、r i,2 and r i,3 , r6d[3:6] represents the feature vector r6d i =[r i,1 ,r i,2 ,r i,3 ,r i,4 ,r i,5 ,r i,6 ] i,4 、r i,5 and r i,6 .

[0071] Thus, the computing device 20 converts the feature vectors 1-n corresponding to each joint point into three-dimensional rotation matrices 1-n (ie, rotation information) according to the Gram-Schmidt orthogonalization algorithm. The rotation matrices 1-n are represented by R1-Rn .

[0072] in

[0073] Where row0 = [c i,0 ,c i,1 ,c i,2 ],row1=[c i,3 ,c i,4 ,c i,5 ],row2=[c i,6 ,c i,7 ,c i,8 ].

[0074] And where c i,0 ~c i,5 Corresponding to the eigenvector r6d i =[r i,1 ,r i,2 ,r i,3 ,r i,4 ,r i,5 ,r i,6 ].

[0075] For example, the rotation matrix

[0076] where c 1,0 ~c 1,5 Corresponding to the eigenvector r6d1=[r 1,1 ,r 1,2 ,r 1,3 ,r 1,4 ,r 1,5 ,r 1,6 ].

[0077] Rotation Matrix

[0078] where c 2,0 ~c 2,5 Corresponding to the eigenvector r6d2=[r 2,1 ,r 2,2 ,r 2,3 ,r 2,4 ,r 2,5 ,r 2,6 ].

[0079] …

[0080] Rotation Matrix

[0081] where c n,0 ~c n,5 Corresponding to the eigenvector r6d n =[r n,1 ,rn,2 ,r n,3 ,r n,4 ,r n,5 ,r n,6 ].

[0082] Since the six-dimensional feature vector in this technical solution has better continuity than other rotation expression methods (such as Euler angles, quaternions, etc.), it is conducive to the learning of neural networks. Therefore, this technical solution generates rotation information based on the six-dimensional feature vector, ensuring the stability and accuracy of generating position information corresponding to the joint point based on the rotation information.

[0083] Optionally, the pressure data is subjected to feature extraction using a neural network-based feature extraction model to generate a six-dimensional feature vector corresponding to a plurality of joint points, including: performing a three-dimensional convolution operation on the pressure data through a convolution layer to generate a first feature map corresponding to the pressure data; encoding the first feature map through a residual network to generate a second feature map; and extracting features from the second feature map through a fully connected layer to generate a feature vector.

[0084] Specifically, refer to Fig. 9 As shown in the figure, the neural network-based feature extraction model includes a 3D convolutional layer, a residual network, and a fully connected layer.

[0085] First, the computing device 20 inputs the pressure data into the 3D convolution layer of the feature extraction model, performs three-dimensional convolution processing on the pressure data through the 3D convolution layer, and generates a first feature map. The size of the pressure data is, for example, 2×5×64×32, representing a 2-channel 64×32 pressure image of 5 consecutive frames. Then the computing device 20 inputs the first feature map into the residual network, encodes the first feature map through the residual network, and generates a second feature map. Then the computing device inputs the second feature map into the fully connected layer, performs feature extraction on the second feature map through the fully connected layer, and generates a feature vector. The feature vector can be represented as a 15×6 feature vector, for example, representing 15 six-dimensional feature vectors.

[0086] Therefore, this technical solution first captures the spatial characteristics of pressure data through a 3D convolutional layer, then uses a residual network to deeply explore the intrinsic connections of the spatial characteristics, and finally integrates the above information through a fully connected layer to form a highly generalized feature vector, thereby improving the expressive power of the model.

[0087] Optionally, the operation of determining the first position information of multiple joints based on rotation information using a forward dynamics algorithm includes: selecting a target joint from multiple joints; determining first position deviation information between the target joint and the previous-level joint based on second position information of the target joint and the previous-level joint of the target joint, wherein the second position information indicates a reference position of the target joint and the previous-level joint; determining second position deviation information of the target joint after rotation relative to the previous-level joint based on the rotation matrix and the first position deviation information; and determining the first position information of the target joint based on the second position deviation information and the first position information of the previous-level joint.

[0088] Specifically, the computing device 20 pre-specifies a root joint point among the plurality of joint points, and then performs a classification process on each joint point according to the connection distance with the root joint point to determine the level of each joint point. The closer the joint point is to the root joint point, the higher the level. And the plurality of joint points may include, for example, the center of the buttocks, the neck, the nose, the left shoulder, the right shoulder, the left elbow, the right elbow, the left wrist, the right wrist, the left hip, the right hip, the left knee, the right knee, the left ankle, and the right ankle.

[0089] refer to Fig.10 As shown, for example, the root joint point designated by the computing device 20 is "hip center", so the computing device 20 performs hierarchical processing on other joint points according to the distance and connection relationship with the "hip center". For example, "neck", "left hip" and "right hip" are closest to the "hip center" and have a connection relationship with the "hip center", so the computing device 20 determines "neck", "left hip" and "right hip" as the next level joint points of "hip center", that is, the first level joint points.

[0090] Afterwards, the computing device 20 uses the joint points associated with the first-level joint points as second-level joint points. For example, if the computing device 20 determines that "nose", "left shoulder" and "right shoulder" are associated with the first-level joint point "neck", then the computing device 20 determines that "nose", "left shoulder" and "right shoulder" are associated with the first-level joint point "neck", and then determines that "neck" is the next-level joint point, that is, the second-level joint point.

[0091] Thus, the computing device 20 determines the levels of other joint points according to the connection relationship between the joint points. For example, the "left knee" is determined as the joint point of the next level (i.e., the second joint point) of the "left hip", and the "right knee" is determined as the joint point of the next level (i.e., the second joint point) of the "right hip". And the computing device 20 generates a joint point tree (e.g., the second joint point) according to the connection relationship between the joint points. Fig.10 shown).

[0092] Furthermore, the computing device 20 first selects the root joint point from the multiple joint points of the joint point tree as the target joint point, and then determines each joint point as the target joint point in turn according to the level of the joint point, and then uses the forward dynamics algorithm to calculate the first position information of each target joint point in turn.

[0093] For example, first, the computing device 20 takes the root joint point, that is, the "center of the buttocks", as the target joint point. The computing device 20 takes the position of the root joint point as the origin in advance, thereby determining the first position information of the root joint point as the target joint point.

[0094] Further, the computing device 20 uses the first-level joint point "neck" in the joint point tree as the target joint point, and then determines the first-level joint point of the first-level joint point "neck". Thus, the computing device 20 determines that the first-level joint point of the first-level joint point "neck" is "hip center" according to the number of joint points.

[0095] Further, the computing device 20 obtains the position information of the target joint point "neck" in the reference posture (i.e., the second position information) and the position information of the upper-level joint point "hip center" in the reference posture (i.e., the second position information). The position information in the reference posture is used to indicate the coordinates of the joint points of the target human body in the reference posture. For example, the reference posture can be the posture of the target human body standing vertically.

[0096] The computing device 20 then calculates the difference between the second position information of the target joint point “neck” and the second position information of the upper-level joint point “center of the hips” as the position deviation information between the target joint point “neck” and the upper-level joint point “center of the hips” (i.e., the first position deviation information).

[0097] Further, the computing device 20 obtains the rotation matrix corresponding to the target joint point "neck", calculates the product between the rotation matrix and the first position deviation information, and determines the position deviation information (second position deviation information) of the target joint point "neck" after rotation relative to the upper-level joint point "hip center". Afterwards, the computing device 20 calculates the sum of the second position deviation information and the first position information of the upper-level joint point to obtain the first position information of the target joint point "neck".

[0098] Furthermore, the computing device 20 sequentially takes other joint points as target joint points, and sequentially calculates the first position information of other joint points according to the above-mentioned method for calculating the first position information of the "neck", which will not be repeated here.

[0099] Preferably, the computing device 20 can calculate each joint point a in sequence by the following formula for calculating the first position information: i The first position information q i :

[0100] q i =R i (p i -p pa(i) )+q pa(i) .

[0101] Among them, R i is the joint point a i The rotation matrix, p i is the joint point a i The second position information in the reference posture, q i is the joint point a i The first position information, p pa(i) is the joint point a i The second position information of the previous joint point in the reference posture, q pa(i) is the joint point a i The first position information of the previous level joint point.

[0102] In actual applications, the inventors found that the pressure data generated by the pressure smart clothing 10 cannot represent the absolute position of the target body in space, so the pressure data cannot be directly used to determine the position of the target body's joints in space. However, the pressure data is related to the rotation of the target body's joints (for example, when the elbow is bent, the pressure smart clothing 10 will be tightened, and the elbow squeezes the pressure smart clothing 10 from the inside). Based on this, the rotation of the joints can be inferred from the pressure data. Therefore, in the present technical solution, the computing device 20 determines the rotation information of each joint point relative to the previous joint point based on the pressure data. And the computing device 20 takes the root joint point (i.e., the center of the buttocks) as the origin, and determines the position information of each joint point in space (i.e., the first position information) based on the rotation information through the forward dynamics algorithm.

[0103] Therefore, the technical solution calculates the first position information through the rotation information and the position deviation information of the joint point, thereby ensuring the accuracy of the position information of each joint point. In addition, the technical solution uses the forward dynamics algorithm to calculate the first position information of the joint points of each level downward from the root joint point, so that the first position information of the joint point that needs to be calculated can be calculated based on the first position information of the previous level joint point that has been calculated, thereby improving the calculation efficiency.

[0104] Optionally, the pressure data is data in matrix form, and before determining the position information corresponding to multiple joints of the target body based on the pressure data, it also includes at least one of the following operations: organizing the format of the pressure data into pressure data that reflects the pressure distribution of the target body, wherein the pressure data is data in matrix form; adjusting the value of the pressure data greater than a first threshold and less than a second threshold to the first threshold, and adjusting the value of the pressure data greater than the second threshold to 0; and repairing bad sectors on the pressure data.

[0105] Specifically, after the computing device 20 obtains the pressure data sent by the pressure smart garment 10, it pre-processes the pressure data.

[0106] For example, the computing device 20 arranges the format of the pressure data to obtain pressure data reflecting the pressure distribution of the target human body, wherein the pressure data is in the form of a matrix.

[0107] More specifically, the pressure data received by the computing device 20 is a pressure image corresponding to the top of the pressure smart clothing (eg Figure 5A As shown), and the pressure image corresponding to the pants of the pressure smart clothing (as shown Figure 5B The pressure image corresponding to the top includes the pressure data of the front and back of the upper body of the target human body, and the pressure image corresponding to the pants of the pressure smart clothing includes the pressure data of the front and back of the lower body of the target human body.

[0108] Furthermore, the computing device 20 formats the pressure data and generates pressure data reflecting the pressure distribution of the target human body. The pressure data may be a pressure image (such as a pressure image) reflecting the pressure distribution of the front of the target human body. Fig.11A ) and a pressure image reflecting the pressure distribution on the back of the target human body (as shown Fig. 11B It can also be a pressure image reflecting the pressure distribution of the whole body of the target human body, or it can be 15 pressure images reflecting the pressure distribution of 15 joints of the target human body, including the center of the buttocks, neck, nose, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip, right hip, left knee, right knee, left ankle and right ankle. Among them, each pressure image can be partially overlapped. It should be noted that the pressure data obtained after formatting can be set according to current needs, and no specific limitation is made here.

[0109] Therefore, this technical solution ensures that the data is organized in a way that matches the natural form of the human body by organizing the format of the pressure data into pressure data that reflects the pressure distribution of the target human body, so that subsequent analysis can be more intuitively mapped to the pressure distribution of various aspects and angles of the human body, thereby enhancing the intuitiveness of data interpretation.

[0110] In addition, the preprocessing operation also includes that the computing device 20 adjusts the value of the pressure data that is greater than the first threshold and less than the second threshold to the first threshold, and adjusts the value of the pressure data that is greater than the second threshold to zero.

[0111] Since the pressure sample points on the pressure smart clothing may be affected by the short circuit, the generated pressure data will have noise points with too high values. Therefore, the computing device pre-sets the first threshold and the second threshold, for example, the first threshold is 512 and the second threshold is 1024. The first threshold and the second threshold can be specifically set according to empirical values.

[0112] Therefore, when the value of a data point in the pressure data is greater than the first threshold 512 and less than the second threshold 1024, the computing device 20 adjusts the value of the data point to the same value as the first threshold, that is, 512.

[0113] When the value of a data point in the pressure data is greater than the second threshold 1024, the computing device 20 adjusts the value of the data point to 0.

[0114] Thus, the computing device 20 adjusts the pressure data with excessively high values ​​to a normal range using a preset threshold value, and obtains an adjusted pressure image reflecting the pressure distribution on the front side of the target human body (eg, Fig. 12A ) and a pressure image reflecting the pressure distribution on the back of the target body (as shown Fig. 12B shown).

[0115] Therefore, the computing device 20 sets a threshold and adjusts the data that exceeds the threshold, which can effectively remove or reduce the impact of noise, achieve data standardization, ensure that all data is within a reasonable range, and is conducive to the stability of the algorithm and the consistency of the results.

[0116] In addition, the preprocessing operation also includes repairing bad sectors of the pressure data.

[0117] Specifically, when a pressure sample point in the pressure smart clothing 10 is disconnected, a row or a column of the pressure data generated in the form of a matrix will lack data values, generating a bad track. Therefore, the computing device 20 uses a Gaussian filter method with a 3×3 Gaussian kernel to perform differential fitting on the bad track based on the pressure data adjacent to the bad track, thereby filling and repairing the bad track.

[0118] In addition, the preprocessing operation also includes that the computing device 20 divides the pressure data obtained after adjusting the data points according to the preset threshold and the pressure data obtained after the bad track is repaired by 512, thereby normalizing the pressure data.

[0119] Therefore, in the event of data loss or damage, the computing device 20 can compensate for the loss by repairing bad sectors, thereby ensuring the integrity of the data.

[0120] Therefore, this technical solution significantly improves the quality and availability of data by preprocessing the pressure data, optimizes the efficiency and accuracy of subsequent analysis, and makes posture reconstruction more accurate.

[0121] Optionally, the method also includes: training the feature extraction model through the following steps: constructing training samples, wherein the training samples include sample pressure data and corresponding sample joint point position information; and training the feature extraction model according to the training samples using a minimum square error loss function and a cosine similarity loss function.

[0122] Specifically, the human posture reconstruction system also includes a training module for training the feature extraction model. The training module obtains sample pressure data and corresponding sample joint point position information as training samples. The sample joint point position information is used to indicate the position of each joint point under the posture corresponding to the sample pressure data. And for the sample joint point position information, the training module can determine the position information of each joint point according to a preset joint point positioning model, or determine the position information of each joint point according to a camera-based linear triangulation method, which is not specifically limited here.

[0123] More specifically, the training module pre-sets multiple joints of the target human body. For example, 15 joints are the center of the buttocks, the neck, the nose, the left shoulder, the right shoulder, the left elbow, the right elbow, the left wrist, the right wrist, the left hip, the right hip, the left knee, the right knee, the left ankle, and the right ankle. Then, the training module obtains the sample pressure data and the sample joint point position information of the multiple joint points corresponding to the sample pressure data as a set of training samples. Thus, the training module obtains multiple sets of training samples according to the above method for training the feature extraction model.

[0124] Furthermore, the training module inputs the sample pressure data into the initial model corresponding to the feature extraction model, so that the initial model outputs the sample feature vector. Then the training module generates the sample rotation information according to the sample feature vector using the mapping algorithm. Then the training module generates the predicted joint point position information according to the sample rotation information using the forward dynamics algorithm.

[0125] The training module generates sample rotation information based on the sample feature vector using a mapping algorithm, so that the feature vector output after the feature extraction model training is completed satisfies the orthogonal relationship with the corresponding rotation information.

[0126] Furthermore, the training module uses a preset minimum square error loss function and a cosine similarity loss function to train the feature extraction model according to the training samples, wherein the minimum square error loss function L2 is:

[0127]

[0128] Where N represents the number of training samples. j and They are the predicted joint point position information and the sample joint point position information respectively.

[0129] The cosine similarity loss function L c for:

[0130]

[0131] Among them, g pa(j) and They are respectively the predicted position information of the previous level joint point and the sample position information of the previous level joint point.

[0132] Therefore, this technical solution combines the minimum square error loss function and the cosine similarity loss function to train the feature extraction model, thereby optimizing the model performance from different angles.

[0133] In summary, Fig.13 As shown, the sequential steps of the computing device 20 to reconstruct the posture according to the pressure data are as follows:

[0134] S1301: the pressure smart clothing 10 generates pressure data in response to the posture of the target human body, and sends the pressure data to the computing device 20;

[0135] S1302: After receiving the pressure data sent by the pressure smart clothing 10, the computing device 20 uses a feature extraction model to extract features from the pressure data and generates feature vectors corresponding to multiple joint points;

[0136] S1303: The computing device 20 generates corresponding rotation information according to the feature vector through a mapping algorithm;

[0137] S1304: The computing device 20 sequentially uses the multiple joint points as target joint points;

[0138] S1304: The computing device 20 determines first position deviation information between the target joint point and the upper-level joint point according to the second position information of the target joint point and the upper-level joint point of the target joint point;

[0139] S1305: The computing device 20 determines second position deviation information of the target joint point after it is rotated relative to the previous joint point according to the rotation matrix and the first position deviation information;

[0140] S1306: The computing device 20 determines the first position information of the target joint point according to the second position deviation information and the first position information of the previous joint point;

[0141] S1307: The computing device 20 generates a graphic corresponding to the gesture according to the first position information.

[0142] In addition, according to the second aspect of this embodiment, a method for reconstructing a human body posture is also provided, including: acquiring pressure data generated by the pressure smart clothing in response to the posture of the target human body; and generating a graphic corresponding to the posture according to the pressure data.

[0143] In addition, according to the third aspect of this embodiment, a method for locating human joints is also provided, including: obtaining pressure data generated by the pressure smart clothing in response to the posture of the target human body; and determining position information corresponding to multiple joints of the target human body based on the pressure data.

[0144] In addition, reference Figure 1 As shown, according to the fourth aspect of this embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0145] In addition, reference Figure 1 As shown, according to the fifth aspect of this embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0146] Thus, according to this embodiment, the pressure smart clothing generates pressure data, and the computing device determines the position information corresponding to multiple joints of the target human body according to the pressure data generated by the pressure smart clothing, and then generates a graph corresponding to the posture according to the position information, thereby reconstructing the posture. Compared with the prior art, this technical solution generates pressure data corresponding to the target human body through the pressure smart clothing, reduces interference to the target human body, improves the comfort and acceptance of use, and reduces the limitation of not being able to obtain pressure data anytime and anywhere to reconstruct the posture of the target human body. This solves the technical problem of high limitations in reconstructing human body posture in the prior art.

[0147] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0148] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0149] Example 2

[0150] Fig.14 The human body posture reconstruction device 1400 according to the first aspect of this embodiment is shown, and the device 1400 corresponds to the method according to the first aspect of embodiment 1. Fig.14 As shown, the device 1400 includes: a data acquisition module 1410, used to acquire pressure data generated by the pressure smart clothing in response to the posture of the target body; an information determination module 1420, used to determine first position information corresponding to multiple joints of the target body according to the pressure data; and a graphics generation module 1430, used to generate a graphic corresponding to the posture according to the first position information.

[0151] Optionally, the information determination module 1420 includes: a first determination submodule, used to determine the rotation information of multiple joint points according to the pressure data; and a second determination submodule, used to determine the first position information of the multiple joint points according to the rotation information using a forward dynamics algorithm.

[0152] Optionally, the first determination submodule includes: a first generation unit, used to extract features from pressure data using a neural network-based feature extraction model, and generate six-dimensional feature vectors corresponding to multiple joint points respectively, wherein the feature vectors are associated with the rotation information of the corresponding joint points; and a first determination unit, used to determine the corresponding rotation matrix as rotation information based on the feature vector through a preset mapping algorithm.

[0153] Optionally, the first generating unit includes: performing a three-dimensional convolution operation on the pressure data through a convolution layer to generate a first feature map corresponding to the pressure data; encoding the first feature map through a residual network to generate a second feature map; and extracting features from the second feature map through a fully connected layer to generate a feature vector.

[0154] Optionally, the second determination submodule includes: a joint point selection unit, used to select a target joint point from multiple joint points; a second determination unit, used to determine the first position deviation information between the target joint point and the previous level joint point based on the second position information of the target joint point and the previous level joint point of the target joint point, wherein the second position information indicates the reference position of the target joint point and the previous level joint point; a third determination unit, used to determine the second position deviation information of the target joint point after rotation relative to the previous level joint point based on the rotation matrix and the first position deviation information; and a fourth determination unit, used to determine the first position information of the target joint point based on the second position deviation information and the first position information of the previous level joint point.

[0155] Optionally, the device 1400 also includes: a first preprocessing module, which formats the pressure data into pressure data that reflects the pressure distribution of the target human body, wherein the pressure data is in matrix form; a second preprocessing module, which is used to adjust the value of the pressure data greater than the first threshold and less than the second threshold to the first threshold, and adjust the value of the pressure data greater than the second threshold to 0; and a third preprocessing module, which is used to repair bad sectors on the pressure data.

[0156] Optionally, the device 1400 also includes: a sample acquisition module, used to construct training samples, wherein the training samples include sample pressure data and corresponding sample joint point position information; and a model training module, used to train the feature extraction model according to the training samples using a minimum square error loss function and a cosine similarity loss function.

[0157] Thus, according to this embodiment, the pressure smart clothing generates pressure data, and the computing device determines the position information corresponding to multiple joints of the target human body according to the pressure data generated by the pressure smart clothing, and then generates a graph corresponding to the posture according to the position information, thereby reconstructing the posture. Compared with the prior art, this technical solution generates pressure data corresponding to the target human body through the pressure smart clothing, reduces interference to the target human body, improves the comfort and acceptance of use, and reduces the limitation of not being able to obtain pressure data anytime and anywhere to reconstruct the posture of the target human body. This solves the technical problem of high limitations in reconstructing human body posture in the prior art.

[0158] Example 3

[0159] Fig.15 The human body posture reconstruction device 1500 according to the first aspect of this embodiment is shown, and the device 1500 corresponds to the method according to the first aspect of embodiment 1. Fig.15As shown, the device 1500 includes: a processor 1510; and a memory 1520, which is connected to the processor 1510 and is used to provide the processor 1510 with instructions for processing the following processing steps: obtaining pressure data generated by the pressure smart clothing in response to the posture of the target body; determining first position information corresponding to multiple joints of the target body based on the pressure data; and generating a graphic corresponding to the posture based on the first position information.

[0160] Optionally, the operation of determining first position information corresponding to multiple joints of the target human body based on pressure data includes: determining rotation information of the multiple joints based on the pressure data; and determining the first position information of the multiple joints based on the rotation information using a forward dynamics algorithm.

[0161] Optionally, the operation of determining the rotation information of multiple joint points based on pressure data includes: performing feature extraction on the pressure data using a neural network-based feature extraction model to generate six-dimensional feature vectors corresponding to the multiple joint points respectively, wherein the feature vectors are associated with the rotation information of the corresponding joint points; and determining a corresponding rotation matrix as the rotation information based on the feature vectors through a preset mapping algorithm.

[0162] Optionally, the pressure data is subjected to feature extraction using a neural network-based feature extraction model to generate a six-dimensional feature vector corresponding to a plurality of joint points, including: performing a three-dimensional convolution operation on the pressure data through a convolution layer to generate a first feature map corresponding to the pressure data; encoding the first feature map through a residual network to generate a second feature map; and extracting features from the second feature map through a fully connected layer to generate a feature vector.

[0163] Optionally, the operation of determining the first position information of multiple joints based on rotation information using a forward dynamics algorithm includes: selecting a target joint from multiple joints; determining first position deviation information between the target joint and the previous-level joint based on second position information of the target joint and the previous-level joint of the target joint, wherein the second position information indicates a reference position of the target joint and the previous-level joint; determining second position deviation information of the target joint after rotation relative to the previous-level joint based on the rotation matrix and the first position deviation information; and determining the first position information of the target joint based on the second position deviation information and the first position information of the previous-level joint.

[0164] Optionally, the pressure data is data in matrix form, and before determining the position information corresponding to multiple joints of the target body based on the pressure data, it also includes at least one of the following operations: organizing the format of the pressure data into pressure data that reflects the pressure distribution of the target body, wherein the pressure data is data in matrix form; adjusting the value of the pressure data greater than a first threshold and less than a second threshold to the first threshold, and adjusting the value of the pressure data greater than the second threshold to 0; and repairing bad sectors on the pressure data.

[0165] Optionally, the memory 1520 is also used to provide instructions for the processor 1510 to process the following processing steps: training the feature extraction model through the following steps: constructing training samples, wherein the training samples include sample pressure data and corresponding sample joint point position information; and training the feature extraction model based on the training samples using a minimum square error loss function and a cosine similarity loss function.

[0166] Thus, according to this embodiment, the pressure smart clothing generates pressure data, and the computing device determines the position information corresponding to multiple joints of the target human body according to the pressure data generated by the pressure smart clothing, and then generates a graph corresponding to the posture according to the position information, thereby reconstructing the posture. Compared with the prior art, this technical solution generates pressure data corresponding to the target human body through the pressure smart clothing, reduces interference to the target human body, improves the comfort and acceptance of use, and reduces the limitation of not being able to obtain pressure data anytime and anywhere to reconstruct the posture of the target human body. This solves the technical problem of high limitations in reconstructing human body posture in the prior art.

[0167] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0168] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0169] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0170] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0171] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0172] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0173] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A human body posture reconstruction method, characterized in that: include: Acquire pressure data generated by the pressure smart clothing in response to the posture of the target human body; Determine first position information corresponding to multiple joints of the target human body according to the pressure data; as well as A graphic corresponding to the posture is generated according to the first position information.

2. The method according to claim 1, characterized in that The operation of determining first position information corresponding to a plurality of joint points of the target human body according to the pressure data includes: Determining rotation information of the plurality of joint points according to the pressure data; and The first position information of the plurality of joint points is determined according to the rotation information using a forward dynamics algorithm, wherein The operation of determining the rotation information of the plurality of joint points according to the pressure data comprises: extracting features from the pressure data using a feature extraction model based on a neural network to generate six-dimensional feature vectors corresponding to the plurality of joint points, wherein the feature vectors are associated with the rotation information of the corresponding joint points; and determining a corresponding rotation matrix as the rotation information according to the feature vectors through a preset mapping algorithm, wherein The operation of extracting features from the pressure data using a feature extraction model based on a neural network to generate six-dimensional feature vectors corresponding to the multiple joint points respectively includes: performing a three-dimensional convolution operation on the pressure data through a convolution layer to generate a first feature map corresponding to the pressure data; encoding the first feature map through a residual network to generate a second feature map; and extracting features from the second feature map through a fully connected layer to generate the feature vector, and The operation of determining the first position information of the multiple joint points according to the rotation information using a forward dynamics algorithm includes: selecting a target joint point from the multiple joint points; determining the first position deviation information between the target joint point and the upper-level joint point according to the second position information of the target joint point and the upper-level joint point of the target joint point, wherein the second position information indicates the reference position of the target joint point and the upper-level joint point; determining the second position deviation information of the target joint point after rotation relative to the upper-level joint point according to the rotation matrix and the first position deviation information; and determining the first position information of the target joint point according to the second position deviation information and the first position information of the upper-level joint point.

3. The method according to claim 1, characterized in that The pressure data is data in a matrix form, and before determining the position information corresponding to the multiple joints of the target human body according to the pressure data, at least one of the following operations is also included: Arranging the format of the pressure data into pressure data reflecting the pressure distribution of the target human body, wherein the pressure data is data in a matrix form; Adjusting the value of the pressure data that is greater than a first threshold and less than a second threshold to the first threshold, and adjusting the value of the pressure data that is greater than the second threshold to 0; as well as Repair bad sectors on the pressure data.

4. The method according to claim 2, characterized in that: Also includes: The feature extraction model is trained by the following steps: Constructing a training sample, wherein the training sample includes sample pressure data and corresponding sample joint point position information; as well as The feature extraction model is trained according to the training samples using a minimum square error loss function and a cosine similarity loss function.

5. A method for reconstructing a human body posture, characterized in that: include: Acquire pressure data generated by the pressure smart clothing in response to the posture of the target human body; as well as A graphic corresponding to the posture is generated according to the pressure data.

6. A method for locating human joints, characterized in that: include: Acquire pressure data generated by the pressure smart clothing in response to the posture of the target human body; as well as The position information corresponding to the multiple joints of the target human body is determined according to the pressure data.

7. A human body posture reconstruction device, characterized in that: include: A data acquisition module, used to acquire pressure data generated by the pressure smart clothing in response to the posture of the target human body; An information determination module, used to determine first position information corresponding to a plurality of joints of the target human body according to the pressure data; as well as A graphics generation module is used to generate a graphic corresponding to the posture according to the first position information.

8. A human body posture reconstruction device, characterized in that: include: processor; as well as A memory, connected to the processor, configured to provide the processor with instructions for processing the following processing steps: Acquire pressure data generated by the pressure smart clothing in response to the posture of the target human body; Determine first position information corresponding to multiple joints of the target human body according to the pressure data; as well as A graphic corresponding to the posture is generated according to the first position information.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to claim 1 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to claim 1 are implemented.