CN114429554A

CN114429554A - Data enhancement method and device for human body posture estimation and terminal equipment

Info

Publication number: CN114429554A
Application number: CN202111629715.XA
Authority: CN
Inventors: 林灿然; 张雯圆; 邵池; 庞建新
Original assignee: Ubtech Robotics Corp
Current assignee: Ubtech Robotics Corp
Priority date: 2021-12-28
Filing date: 2021-12-28
Publication date: 2022-05-03

Abstract

The application is applicable to the technical field of data enhancement, and provides a data enhancement method and device for human body posture estimation and terminal equipment. In the embodiment of the application, a target image is obtained, and posture estimation is carried out on a human body in the target image to obtain an initial heat map of the human body; determining target key points with confidence degrees meeting a preset threshold range from the initial heat map; determining local parts of the human body corresponding to the target key points, and acquiring a preset number of part graphs corresponding to the local parts from a preset part library; and performing data enhancement on the initial heat map according to the preset number of the part maps to obtain an enhanced human body heat map, thereby improving the accuracy of extracting the human body key points.

Description

Data enhancement method and device for human body posture estimation and terminal equipment

Technical Field

The application belongs to the technical field of data enhancement, and particularly relates to a data enhancement method and device for human body posture estimation and a terminal device.

Background

With the development of society, key points of a human body can be extracted by estimating the posture of the human body, so that visual perception or other tasks are realized according to the key points, for example, somatosensory game interaction is realized through key point extraction, target tracking is realized through key point extraction, behavior recognition is realized through key point extraction, and the like.

However, since the structure of the human body is a non-rigid body, there are various changes compared to rigid bodies such as automobiles and tables and chairs, and for example, a part of the human body performs a rotational movement around a joint, thereby performing a complicated behavior or action. In addition, there are some special scenes with serious shielding or more obstacles, so the pose estimation of the human body cannot be well performed, and the accuracy of extracting the key points of the human body is low.

Disclosure of Invention

The embodiment of the application provides a data enhancement method and device for human body posture estimation and terminal equipment, and can solve the problem of low precision of extracting human body key points.

In a first aspect, an embodiment of the present application provides a data enhancement method for human body posture estimation, including:

acquiring a target image, and performing posture estimation on a human body in the target image to obtain an initial heat map of the human body;

determining a target key point with a confidence coefficient meeting a preset threshold range from the initial heat map;

determining local parts of the human body corresponding to the target key points, and acquiring a preset number of part graphs corresponding to the local parts from a preset part library;

and performing data enhancement on the initial heat map according to the preset number of the part maps to obtain an enhanced human body heat map.

In a second aspect, an embodiment of the present application provides a data enhancement device for human body posture estimation, including:

the image acquisition module is used for acquiring a target image and carrying out posture estimation on the human body in the target image to obtain an initial heat map of the human body;

a key point determining module, configured to determine, from the initial heat map, a target key point whose confidence level meets a preset threshold range;

a component diagram acquiring module, configured to determine local components of the human body corresponding to the target key points, and acquire a preset number of component diagrams corresponding to the local components from a preset component library;

and the data enhancement module is used for carrying out data enhancement on the initial heat map according to the preset number of the part maps to obtain an enhanced human body heat map.

In a third aspect, an embodiment of the present application provides a terminal device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements any of the steps of the data enhancement method for human body posture estimation when executing the computer program.

In a fourth aspect, the present application provides a computer-readable storage medium, where a computer program is stored, and the computer program, when executed by a processor, implements the steps of any one of the above-mentioned data enhancement methods for human body posture estimation.

In a fifth aspect, the present application provides a computer program product, which when run on a terminal device, causes the terminal device to execute any one of the above-mentioned data enhancement methods for human body posture estimation in the first aspect.

In the embodiment of the application, the target image is obtained, the posture of the human body in the target image is estimated, so as to obtain the initial heat map of the human body, then determining the target key points with confidence degrees meeting the preset threshold range from the initial heat map, to determine the key points with lower precision, and then to determine the local parts of the human body corresponding to the target key points, the method comprises acquiring a preset number of component drawings corresponding to the local components from a preset component library, performing data enhancement on the initial heat map according to the preset number of component drawings to improve the accuracy of posture estimation of the human body, finally obtaining an enhanced human body heat map so as to obtain key points with higher human body precision through the enhanced human body heat map, therefore, the accuracy of extracting the key points of the human body is improved by performing data enhancement on the heat map obtained by posture estimation.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.

Fig. 1 is a first flowchart illustrating a data enhancement method for human body posture estimation according to an embodiment of the present disclosure;

FIG. 2 is a second flowchart of a method for enhancing data of human body posture estimation according to an embodiment of the present disclosure;

FIG. 3 is a schematic structural diagram of a data enhancement apparatus for human body posture estimation according to an embodiment of the present application;

fig. 4 is a schematic structural diagram of a terminal device according to an embodiment of the present application.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It should also be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

As used in this specification and the appended claims, the term "if" may be interpreted contextually as "when", "upon" or "in response to" determining "or" in response to detecting ". Similarly, the phrase "if it is determined" or "if a [ described condition or event ] is detected" may be interpreted contextually to mean "upon determining" or "in response to determining" or "upon detecting [ described condition or event ]" or "in response to detecting [ described condition or event ]".

Furthermore, in the description of the present application and the appended claims, the terms "first," "second," "third," and the like are used for distinguishing between descriptions and not necessarily for describing or implying relative importance.

Fig. 1 is a schematic flow chart of a data enhancement method for human body posture estimation in an embodiment of the present application, where an execution subject of the method may be a terminal device, as shown in fig. 1, the data enhancement method for human body posture estimation may include the following steps:

and S101, acquiring a target image, and performing posture estimation on the human body in the target image to obtain an initial heat map of the human body.

In this embodiment, the terminal device may perform pose estimation on the human body in the target image by using a preset pose estimation algorithm to obtain an initial heat map in the target image without performing any processing on the human body. The attitude estimation algorithm may be implemented by an attitude estimation model, and the attitude estimation algorithm includes, but is not limited to, a fastpos algorithm, an LPN algorithm, and the like.

In an embodiment, before performing the pose estimation on the human body, the terminal device may first detect the human body in the image through a preset human body detection algorithm, so as to obtain a human body detection frame for a single human body, for example, if three human bodies exist in the current target image, three human body detection frames correspondingly appear, and then perform the pose estimation on the human body in the human body detection frame, so as to obtain an initial heat map of the human body corresponding to the human body detection frame. The human body detection algorithm includes, but is not limited to, a YOLO algorithm, an SSD algorithm, and the like.

And S102, determining target key points with confidence degrees meeting a preset threshold range from the initial heat map.

In this embodiment, for the initial heat map obtained by the terminal device through the posture estimation model, the terminal device may process through an Argmax function to obtain a coordinate of a key point corresponding to the initial heat map, and then process the obtained key point through a Softmax function to obtain a confidence level of each key point, so as to determine the confidence level of each key point, select a target key point whose confidence level meets a preset threshold range from the key points, that is, a key point easy to identify an error, and perform data enhancement on the key point easy to identify the error. Wherein the threshold range may be 0.4 to 0.85.

It can be understood that, if the confidence of the key point is smaller than the minimum value in the threshold range, which indicates that the key point is an invisible point, the key point is filtered without being displayed or processed correspondingly; if the confidence of the key point is greater than the maximum value in the threshold range, the key point is a point easy to learn, and the data of the point easy to learn is not required to be enhanced, and the point easy to learn is directly displayed or correspondingly processed.

In one embodiment, the number of the above-mentioned keypoints may vary according to the category of the dataset stored by the image, for example, the MPII dataset corresponds to 16 types of keypoints, and the COCO dataset corresponds to 17 types of keypoints.

And S103, determining local parts of the human body corresponding to the target key points, and acquiring a preset number of part graphs corresponding to the local parts from a preset part library.

In this embodiment, the terminal device may process the position of the part by acquiring a certain number of part diagrams, which are the same as the local part of the human body corresponding to the target key point, so as to improve the accuracy of the processed human body heat diagram, and enhance the data of the position by using the plurality of part diagrams, thereby accurately extracting the human body key point.

Specifically, the step S103 may include: the terminal equipment can determine the human body local part corresponding to the target key point from a preset key point comparison table according to the target key point. For example, if the current target key point is located at the center point of the left wrist, the body part may be a part between the left wrist and the left elbow of the human body. The keypoint comparison table may be a table generated according to the positions of the keypoints that may appear on the human body and the components with the lowest confidence corresponding to the positions, and the components with the lowest confidence among the components to which the keypoints are connectable may be determined by means of data analysis.

Specifically, the step S103 may include: the terminal device may determine an adjacent keypoint connected to the target keypoint, thereby determining the local component from the target keypoint and the adjacent keypoint. For example, the part where the target key point and the adjacent key points are connected is determined as the local part to be subjected to data enhancement.

In one embodiment, when there are at least two adjacent keypoints, the terminal device may select an adjacent keypoint with the minimum confidence from the at least two adjacent keypoints, and determine a component between the target keypoint and the adjacent keypoint with the minimum confidence as a local component. It can be understood that, for a component consisting of an adjacent key point with the minimum confidence and a target key point, and a component corresponding to other key points simultaneously connected with the target key point, the confidence of the component consisting of the adjacent key point with the minimum confidence and the target key point is also minimum, so that a local component needing enhancement can be accurately required.

In one embodiment, before step S103, the method may further include: the terminal device obtains a global image of the human body, namely an image only containing the human body, cuts the global image according to a preset segmentation network to obtain each local part image, calculates the confidence of each local part image, and stores the local part image with the confidence greater than a preset confidence threshold in a corresponding position in a part library, for example, according to the name of the local part image. Wherein, the split network may be a PSPNet network. It is understood that the confidence threshold is a confidence threshold corresponding to a point easy to learn, so the confidence threshold may be the maximum value in the threshold range.

And step S104, performing data enhancement on the initial heat map according to the preset number of the part maps to obtain an enhanced human body heat map.

In this embodiment, the terminal device performs local data enhancement on the human body components with poor estimation effect in the initial heat map obtained by the estimation of the posture estimation model according to the preset number of component maps, so as to obtain a preset number of enhanced human body heat maps, and thus obtain key points with high human body precision through the enhanced human body heat maps, and perform corresponding tasks.

For example, if there are currently 2 target key points, which are point a and point B, respectively, and the preset number is set to be 3, after data enhancement is performed on 3 component maps corresponding to point a and point B, human body heat maps a1, a2, and A3 are obtained after the point a corresponds to the enhanced data, and human body heat maps B1, B2, and B3 after the point B corresponds to the enhanced data.

In one embodiment, step S104 includes: and randomly pasting a preset number of part images on the human body in the target image, and if the human body detection frame exists, randomly pasting the part images in the human body detection frame to obtain a preset number of human body sample images. And processing a preset number of human body sample images by a preset posture estimation algorithm, for example, processing by using the posture estimation algorithm to obtain an enhanced human body heat image corresponding to the human body enhancement data in the target image. In addition, the preset number of component maps may be subjected to data enhancement based on global data of the human body, for example, rotation by a certain angle in a certain rotation direction, change of brightness contrast.

In one embodiment, before step S104, the method may further include: the terminal equipment calculates the part length of the local part according to the target key point, namely the length of the local limb of the human body, and then normalizes the part images in preset number according to the part length, so that the scale of the local part to be enhanced in the initial heat image is consistent with that of the local part corresponding to the enhanced human body heat image. For example, if the length of a component is 250 pixels, the lengths of a preset number of component maps are normalized to be 250 pixels.

Specifically, if the local component is determined only according to the target key point, the length of the local component can be estimated based on the length of the current human body, so that the length of the local component can be estimated. And if the local part is determined according to the target key point and the adjacent key point adjacent to the target key point, determining the part length of the local part according to the horizontal and vertical coordinates of the target key point and the adjacent key point.

In one embodiment, as shown in fig. 2, after step S104, the method further includes:

step S201, a tag heat map of a human body is obtained, and a first loss value is determined according to the tag heat map and the initial heat map.

In this embodiment, the terminal device may calculate the tag heat map and the initial heat map by using the mean square error, and obtain a loss value between the tag heat map and the initial heat map, that is, a difference between the tag heat map and the initial heat map, that is, the first loss value.

Step S202, determining a second loss value according to the enhanced human body heat map and the initial heat map.

In this embodiment, the terminal device may calculate the enhanced human body heat map and the initial heat map by using the mean square error, so as to obtain a loss value between the enhanced human body heat map and the initial heat map, that is, a difference between the enhanced human body heat map and the initial heat map. It will be appreciated that the second loss value is the sum of the loss values corresponding to a predetermined number of enhanced human body heat maps, since the number of enhanced human body heat maps is the same as the number of component maps used to enhance the data.

Illustratively, based on the human body heat maps a1, a2, A3 obtained in the above example after the point a corresponds to the enhanced data and the human body heat maps B1, B2, B3 obtained in the above example after the point B corresponds to the enhanced data, the loss values of the initial heat map and a1, a2, A3, B1, B2, B3 are calculated respectively, the loss values of the initial heat map and a1, the loss values of the initial heat map and a2, the loss values of the initial heat map and A3, the loss values of the initial heat map and B361, the loss values of the initial heat map and B2, the loss values of the initial heat map and B3, the sum of the loss values of the initial heat map and B1, the loss values of the initial heat map and B363, the second loss value of the loss values of the points a1, a2, B3 and B3 are calculated again, and the second loss values of the points₂。

Step S203, performing weighting processing on the first loss value and the second loss value, and determining a total loss value according to the weighted first loss value and the weighted second loss value. Wherein, the calculation formula of the total loss value is as follows:

Loss_a＝α*Loss₂+(1-α)Loss₁

wherein, the alpha is a weighted value which can be selected from 0.3 to 0.5; loss as described above₁Is a first loss value; loss as described above₂A second loss value; loss as described above_aIs the total loss value.

And step S204, updating model parameters of the posture estimation model of the human body according to the total loss value.

In this embodiment, the terminal device performs gradient pass-back on the total loss value to act on the optimizer to update the model parameters of the attitude estimation model, so that the attitude estimation model is better learned, the estimation effect of the attitude estimation model is improved, and the recognition accuracy of the model on the human key points is further improved.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

Corresponding to the above-mentioned data enhancement method for human body posture estimation, fig. 3 is a schematic structural diagram of a data enhancement device for human body posture estimation in an embodiment of the present application, and as shown in fig. 3, the data enhancement device for human body posture estimation may include:

the image acquisition module 301 is configured to acquire a target image, and perform posture estimation on a human body in the target image to obtain an initial heat map of the human body.

A key point determining module 302, configured to determine a target key point whose confidence level meets a preset threshold range from the initial heat map.

The component diagram acquiring module 303 is configured to determine local components of the human body corresponding to the target key points, and acquire a preset number of component diagrams corresponding to the local components from a preset component library.

And the data enhancement module 304 is configured to perform data enhancement on the initial heat map according to a preset number of component maps to obtain an enhanced human body heat map.

In one embodiment, the data enhancement device for estimating the human body posture may further include:

and the heat map acquisition module is used for acquiring a tag heat map of the human body and determining a first loss value according to the tag heat map and the initial heat map.

A loss value determination module for determining a second loss value based on the enhanced human body heat map and the initial heat map.

And the processing module is used for weighting the first loss value and the second loss value and determining a total loss value according to the weighted first loss value and the weighted second loss value.

And the parameter updating module is used for updating the model parameters of the posture estimation model of the human body according to the total loss value.

In one embodiment, the component diagram obtaining module 303 may include:

and the key point determining unit is used for determining adjacent key points connected with the target key point.

And the component determining unit is used for determining the local component according to the target key point and the adjacent key points.

In one embodiment, the component determination unit may include:

and the key point selecting subunit is used for selecting the adjacent key point with the minimum confidence level from the at least two adjacent key points when at least two adjacent key points exist.

A component determination subunit, configured to determine a component between the target keypoint and the adjacent keypoint with the smallest confidence as a local component.

In one embodiment, the data enhancement module 304 may include:

and the part map pasting unit is used for pasting a preset number of part maps on the human body in the target image at random to obtain a preset number of human body sample maps.

And the processing unit is used for processing a preset number of human body sample images through a preset posture estimation algorithm to obtain the enhanced human body heat image.

In one embodiment, the data enhancement device for estimating the posture of the human body may further include:

and the length calculating unit is used for calculating the part length of the local part according to the target key point.

And the normalization processing unit is used for performing normalization processing on the preset number of component graphs according to the length of the components.

and the cutting unit is used for obtaining the global image of the human body and cutting the global image according to a preset segmentation network to obtain each local part image.

And the confidence coefficient calculation unit is used for calculating the confidence coefficient of each local part graph and storing the local part graph with the confidence coefficient larger than a preset confidence coefficient threshold value in the corresponding position in the part library.

It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the apparatus and the module described above may refer to corresponding processes in the foregoing system embodiments and method embodiments, and are not described herein again.

Fig. 4 is a schematic structural diagram of a terminal device according to an embodiment of the present application. For ease of illustration, only portions relevant to the embodiments of the present application are shown.

As shown in fig. 4, the terminal device 4 of this embodiment includes: at least one processor 400 (only one shown in fig. 4), a memory 401 connected to the processor 400, and a computer program 402, such as a data enhancement program for human body posture estimation, stored in the memory 401 and executable on the at least one processor 400. When the processor 400 executes the computer program 402, the steps in the data enhancement method embodiments of the human body pose estimation, such as the steps S101 to S104 shown in fig. 1, are implemented. Alternatively, the processor 400 executes the computer program 402 to implement the functions of the modules in the device embodiments, such as the modules 301 to 304 shown in fig. 3.

Illustratively, the computer program 402 may be divided into one or more modules, which are stored in the memory 401 and executed by the processor 400 to complete the present application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of the computer program 402 in the terminal device 4. For example, the computer program 402 may be divided into the image acquisition module 301, the key point determination module 302, the component diagram acquisition module 303, and the data enhancement module 304, and the specific functions of the modules are as follows:

the image acquisition module 301 is configured to acquire a target image, perform posture estimation on a human body in the target image, and obtain an initial heat map of the human body;

a key point determining module 302, configured to determine, from the initial heat map, a target key point whose confidence level meets a preset threshold range;

a part map acquiring module 303, configured to determine local parts of the human body corresponding to the target key points, and acquire a preset number of part maps corresponding to the local parts from a preset part library;

The terminal device 4 may include, but is not limited to, a processor 400 and a memory 401. Those skilled in the art will appreciate that fig. 4 is merely an example of the terminal device 4, and does not constitute a limitation of the terminal device 4, and may include more or less components than those shown, or combine some of the components, or different components, such as an input-output device, a network access device, a bus, etc.

The Processor 400 may be a Central Processing Unit (CPU), and the Processor 400 may be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate arrays (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 or the like.

The storage 401 may be an internal storage unit of the terminal device 4 in some embodiments, for example, a hard disk or a memory of the terminal device 4. In other embodiments, the memory 401 may also be an external storage device of the terminal device 4, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the terminal device 4. Further, the memory 401 may include both an internal storage unit and an external storage device of the terminal device 4. The memory 401 is used for storing an operating system, an application program, a Boot Loader (Boot Loader), data, and other programs, such as program codes of the computer programs. The memory 401 described above may also be used to temporarily store data that has been output or is to be output.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned functions may be distributed as different functional units and modules according to needs, that is, the internal structure of the apparatus may be divided into different functional units or modules to implement all or part of the above-mentioned functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the technical solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other ways. For example, the above-described embodiments of the apparatus/terminal device are merely illustrative, and for example, the division of the above modules or units is only one logical function division, and there may be other division manners in actual implementation, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

The integrated unit may be stored in a computer-readable storage medium if it is implemented in the form of a software functional unit and sold or used as a separate product. Based on such understanding, all or part of the processes in the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium and can implement the steps of the embodiments of the methods described above when the computer program is executed by a processor. The computer program includes computer program code, and the computer program code may be in a source code form, an object code form, an executable file or some intermediate form. The computer-readable medium may include at least: any entity or device capable of carrying computer program code to a photographing apparatus/terminal apparatus, a recording medium, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), an electrical carrier signal, a telecommunications signal, and a software distribution medium. Such as a usb-drive, a removable hard drive, a magnetic or optical disk, etc. In certain jurisdictions, computer-readable media may not be an electrical carrier signal or a telecommunications signal in accordance with legislative and patent practice.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the embodiments of the present application, and they should be construed as being included in the present application.

Claims

1. A data enhancement method for human body posture estimation is characterized by comprising the following steps:

determining target key points with confidence degrees meeting a preset threshold range from the initial heat map;

2. The method for data enhancement of human pose estimation according to claim 1, further comprising, after obtaining the enhanced human heat map:

obtaining a tag heat map of the human body, determining a first loss value from the tag heat map and the initial heat map;

determining a second loss value from the enhanced human body heat map and the initial heat map;

weighting the first loss value and the second loss value, and determining a total loss value according to the weighted first loss value and the weighted second loss value;

and updating the model parameters of the posture estimation model of the human body according to the total loss value.

3. The method of data enhancement of human pose estimation according to claim 1, wherein said determining local parts of the human body corresponding to target keypoints comprises:

determining adjacent key points connected with the target key point;

determining the local component according to the target keypoint and the adjacent keypoints.

4. The method of data enhancement of human pose estimation according to claim 3, wherein said determining said local component from said target keypoints and said neighboring keypoints comprises:

when at least two adjacent key points exist, selecting the adjacent key point with the minimum confidence level from the at least two adjacent key points;

determining the part between the target keypoint and the neighboring keypoint with the lowest confidence as the local part.

5. The method for enhancing data of human body pose estimation according to claim 1, wherein said data enhancing the initial heat map according to the preset number of component maps to obtain an enhanced human body heat map comprises:

randomly pasting the preset number of part images on the human body in the target image to obtain a preset number of human body sample images;

and processing the human body sample images in the preset number through a preset posture estimation algorithm to obtain an enhanced human body heat image.

6. The method of data enhancement of human pose estimation according to claim 1, further comprising, prior to data enhancing said initial heat map according to said preset number of component maps:

calculating the part length of the local part according to the target key point;

and carrying out normalization processing on the preset number of component graphs according to the length of the components.

7. The method of enhancing data on human pose estimation according to claim 1, further comprising, before obtaining a preset number of part maps corresponding to said local parts from a preset part library:

obtaining a global image of the human body, and cutting the global image according to a preset segmentation network to obtain each local component image;

calculating the confidence of each local part map, and storing the local part map with the confidence greater than a preset confidence threshold value in the corresponding position in the part library.

8. A data enhancement device for human body posture estimation is characterized by comprising:

the image acquisition module is used for acquiring a target image and estimating the posture of a human body in the target image to obtain an initial heat map of the human body;

the part drawing acquisition module is used for determining local parts of the human body corresponding to the target key points and acquiring a preset number of part drawings corresponding to the local parts from a preset part library;

and the data enhancement module is used for performing data enhancement on the initial heat map according to the preset number of the part maps to obtain an enhanced human body heat map.

9. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor when executing the computer program implements the steps of a data enhancement method of human pose estimation as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the steps of a method for data enhancement of human body pose estimation according to any one of claims 1 to 7.