Training posture adjustment method and device, electronic equipment and storage medium

Abstract

The invention provides a training posture adjustment method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the human body key point detection of an image in a target object training process, and determining the human body key point data of a target object; based on the human body key point data, determining a training posture of the target object; determining an orientation vector of the target object based on the human body key point data, and correcting the training posture based on the orientation vector to obtain a corrected training posture; and determining posture adjustment information of the target object based on the corrected training posture and a standard posture in the training process. The attitude error can be effectively eliminated by calculating the attitude of the target object for correction. After the corrected training posture is obtained, the corrected training posture is compared with the standard posture in the training process, posture adjustment information capable of being used for accurately guiding the process is quantitatively determined, and the posture adjustment accuracy is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a training posture adjustment method and device, electronic equipment and storage medium. BACKGROUND

[0002] In the prior art, the training evaluation process of the personnel performing training relies on artificial examination and subjective judgment. The artificial examination and judgment method has certain subjectivity and one-sidedness, and the evaluation criteria are easily affected by the experience level of the examiner, visual error and cognitive preference.

[0003] After artificial examination and judgment, the artificial feedback output usually describes adjustment opinions in natural language, forming fuzzy instructions, which cannot realize accurate guidance of posture adjustment, resulting in low accuracy of training posture adjustment. SUMMARY

[0004] The present application provides a training posture adjustment method and device, electronic equipment and storage medium, which can realize accurate guidance of posture adjustment and improve the accuracy of training posture adjustment.

[0005] The present application provides a training posture adjustment method, comprising the following steps: detecting human key points of images of a target object training process to determine human key point data of the target object; determining a training posture of the target object based on the human key point data; determining a direction vector of a line connecting a shoulder midpoint and a hip midpoint of the target object based on the human key point data, determining an orientation vector of the target object based on the direction vector, and correcting the training posture based on the orientation vector to obtain a corrected training posture; determining posture adjustment information of the target object based on the corrected training posture and a standard posture of the training process, the posture adjustment information being used for training posture adjustment of the target object.

[0006] According to the training posture adjustment method provided by the present application, the training posture is corrected based on the orientation vector to obtain a corrected training posture, which comprises: determining an angle between the orientation vector and the Y-axis direction of the camera coordinate system; constructing a rotation matrix based on the angle, and transforming the training posture of the target object based on the rotation matrix to obtain a matrix-transformed training posture; the matrix-transformed training posture is taken as the corrected training posture.

[0007] The training posture adjustment method provided in the present application comprises the following steps: The trunk key point of the target object is determined based on the human key point data, and the trunk key point is connected to obtain the trunk connection line of the target object. The trunk connection line is taken as the training posture of the target object.

[0008] The training posture adjustment method provided in the present application comprises the following steps: The difference angle between the corrected trunk connection line and the trunk connection line corresponding to the standard posture of the training process is determined. The posture adjustment information of the target object is determined based on the difference angle.

[0009] The training posture adjustment method provided in the present application comprises the following steps: The dynamic characteristic information of the target object in the training process is obtained, and the dynamic characteristic information comprises speed information and acceleration information. The corrected training posture is further corrected based on the dynamic characteristic information. The training posture adjustment method provided in the present application comprises the following steps: The inertia compensation vector of the target object in the training process is determined based on the dynamic characteristic information. The corrected training posture is further corrected based on the inertia compensation vector.

[0010] The standard posture of the training process is obtained by connecting the trunk key point of the standard execution object in the training process.

[0011] The present application further provides a training posture adjustment method, which comprises the following modules: The key point extraction module is used for human key point detection on the image of the target object in the training process, and the human key point data of the target object is determined. The posture determination module is used for determining the training posture of the target object based on the human key point data. The correction module is configured to determine a line direction vector of a midpoint of a shoulder and a midpoint of a hip of the target object based on the human key point data, determine an orientation vector of the target object based on the line direction vector, and correct the training posture based on the orientation vector to obtain a corrected training posture. The adjustment module is configured to determine posture adjustment information of the target object based on the corrected training posture and a standard posture of the training process, where the posture adjustment information is used for adjusting the training posture of the target object.

[0012] The present application also provides an electronic device including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the training posture adjustment method according to any one of the above when executing the program.

[0013] The present application also provides a non-transitory computer readable storage medium having a computer program stored thereon, where the computer program, when executed by a processor, implements the training posture adjustment method according to any one of the above.

[0014] The training posture adjustment method, device, electronic device, and storage medium provided by the present application can effectively eliminate posture errors caused by camera visual angle or target object orientation deviation by obtaining human key point data of a target object in a training process, determining a training posture of the target object based on the human key point data, correcting the posture by calculating an orientation vector of the target object, and significantly improving the accuracy of posture analysis. After obtaining the corrected training posture, the corrected training posture is compared with a standard posture of the training process to quantitatively determine posture adjustment information that can be used to accurately guide the process, thereby improving the accuracy of posture adjustment. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0016] Figure 1 is a flowchart of the training posture adjustment method provided by the present application.

[0017] Figure 2 is a structural schematic diagram of the training posture adjustment device provided by the present application.

[0018] Figure 3 is a structural schematic diagram of the electronic device provided by the present application. DETAILED DESCRIPTION

[0019] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0020] Figure 1 is a flowchart of the training posture adjustment method provided by the present application, as shown in Figure 1 The method comprises the following steps: Step 110, performing human key point detection on images of a training process of a target object to determine human key point data of the target object; Step 120, determining a training posture of the target object based on the human key point data; Step 130, determining a direction vector of a line connecting a midpoint of a shoulder and a midpoint of a hip of the target object based on the human key point data, determining an orientation vector of the target object based on the direction vector, and correcting the training posture based on the orientation vector to obtain a corrected training posture; Step 140, determining posture adjustment information of the target object based on the corrected training posture and a standard posture of the training process, the posture adjustment information being used for training posture adjustment of the target object.

[0021] The execution subject of the training posture adjustment method provided by the present application can be an electronic device, a component in the electronic device, an integrated circuit, or a chip. The electronic device can be a mobile electronic device or a non-mobile electronic device. Exemplarily, the mobile electronic device can be a mobile phone, a tablet computer, a notebook computer, a palm computer, an ultra-mobile personal computer (UMPC), a netbook, or a personal digital assistant (PDA), and the non-mobile electronic device can be a server, a Network Attached Storage (NAS), or a personal computer (PC), without specific limitation of the present application.

[0022] The technical solutions of the present application will be described in detail below taking a computer executing the training posture adjustment method provided by the present application as an example.

[0023] In step 110, human key point detection is performed on images of a training process of a target object to determine human key point data of the target object.

[0024] In the training posture adjustment process, first, the image of the target object training process needs to be detected for human key points.

[0025] The accuracy of human key point detection directly affects the effectiveness of subsequent posture analysis and adjustment. Human key point detection usually uses deep learning algorithms such as recurrent neural networks (RNN) and convolutional neural networks (CNN), which can efficiently extract features from images. Through training, these networks can accurately identify and locate key points in human images, such as head, shoulders, elbows, wrists, hips, knees, and ankles.

[0026] These key points form the basic framework of human posture, and by detecting these key points, accurate position information of the target object during the training process can be obtained. These information not only provides basic data for subsequent posture analysis, but also helps to understand the action patterns and motion trajectories of the target object.

[0027] In step 120, based on the human key point data, the training posture of the target object is determined.

[0028] After obtaining the human key point data of the target object, the training posture of the target object is determined based on the human key point data.

[0029] Specifically, the trunk line can be formed by connecting the key points, thereby directly representing the trunk position and direction of the target object, and thereby obtaining the training posture of the target object.

[0030] In step 130, based on the human key point data, the direction vector of the line connecting the midpoint of the shoulder and the midpoint of the hip of the target object is determined, and based on the direction vector, the orientation vector of the target object is determined, and based on the orientation vector, the training posture is corrected to obtain the corrected training posture.

[0031] After determining the training posture of the target object, the posture is further corrected to ensure its accuracy.

[0032] Based on the human key point data, the midpoint of the shoulder and the midpoint of the hip are determined. By locating the direction vector of the line connecting the midpoint of the shoulder and the midpoint of the hip of the target object, the orientation vector of the target object can be determined. The orientation vector reflects the actual orientation of the target object in the world coordinate system.

[0033] The correctness of the orientation of the target object can be evaluated by calculating the included angle of the orientation vector relative to the Y-axis direction of the camera coordinate system. Due to the difference between the camera angle of view and the movement direction of the target object, the posture detection may deviate. Therefore, by calculating the included angle of the orientation vector of the target object and the Y-axis direction of the camera coordinate system, the deviation can be quantified, and a rotation matrix can be constructed based on the included angle. By transforming the training posture using the rotation matrix, the posture of the target object can be adjusted to a state aligned with the camera coordinate system, thereby obtaining the corrected training posture. The correction process of the training posture can eliminate the posture deviation caused by the camera angle or the orientation of the target object, and improve the accuracy of the posture analysis.

[0034] In step 140, based on the corrected training posture and the standard posture of the training process, the posture adjustment information of the target object is determined, which is used for the training posture adjustment of the target object.

[0035] Based on the corrected training posture and the standard posture of the training process, the posture adjustment information of the target object is determined. The goal is to evaluate whether the posture of the target object meets the requirements by comparing the differences between the corrected training posture and the standard posture, and to give specific adjustment suggestions.

[0036] Specifically, the difference angle between the corrected torso connecting line and the torso connecting line corresponding to the standard posture can be calculated to determine the posture adjustment information. The posture adjustment information not only indicates the deviation between the current posture of the target object and the standard posture, but also provides the direction and amplitude of the adjustment.

[0037] The posture adjustment information can be fed back to the target object or the training system, so as to guide the target object to continuously optimize the posture in the training process and improve the training effect. At the same time, these information can also be used for adaptive adjustment of the training system to better adapt to the posture characteristics of different target objects.

[0038] Specifically, the above implementation process can be composed of four levels of data acquisition, transmission, storage processing and analysis application. The data acquisition layer realizes real-time acquisition of multi-source data such as training action, equipment state and personnel information by deploying sensor network, video acquisition system and accessing training management system; the data transmission layer relies on 5G wireless communication network and wired network complementarily to ensure fast and stable data transmission; the data storage and processing layer adopts distributed storage system and big data processing framework to complete safe storage and efficient analysis of massive data; the data analysis and application layer constructs training evaluation, prediction model and intelligent recommendation system by means of machine learning and deep learning algorithm, and provides basis for training decision by mining data value.

[0039] In terms of function modules, the training data management realizes data collection and integration, storage query and update and maintenance, to ensure data quality; the training effect evaluation supports real-time and comprehensive evaluation, presents training indicators through visualization, and feeds back training performance in a timely manner; the training prediction and early warning can predict training effect and identify training risks such as injury and abnormality; and the personalized training recommendation customizes training plans according to individual characteristics, training targets and evaluation results, and pushes adaptive training resources.

[0040] The training posture adjustment method provided by the application can effectively eliminate posture errors caused by camera viewing angles or target object orientation deviations, significantly improve the accuracy of posture analysis, and improve the accuracy of posture adjustment.

[0041] In one embodiment, based on the orientation vector, the training posture is corrected to obtain a corrected training posture, including: determining the included angle between the orientation vector and the Y-axis direction of the camera coordinate system; based on the included angle, a rotation matrix is constructed, and the training posture of the target object is transformed based on the rotation matrix to obtain a matrix-transformed training posture; and the matrix-transformed training posture is taken as the corrected training posture.

[0042] Based on the human body key point data, the shoulder midpoint and the hip midpoint are determined. By positioning the direction vector of the line connecting the shoulder midpoint and the hip midpoint of the target object, the orientation vector of the target object can be determined. The orientation vector not only reflects the actual orientation of the target object in the world coordinate system.

[0043] The included angle between the orientation vector and the Y-axis direction of the camera coordinate system is determined. This included angle quantifies the relative angle difference between the target object and the camera. Through this included angle, the degree of deviation between the current posture of the target object and the camera coordinate system can be determined.

[0044] Based on this included angle, a rotation matrix can be constructed. This rotation matrix can accurately rotate the training posture of the target object from its current orientation to a position completely consistent with the Y-axis direction of the camera coordinate system. By applying this rotation matrix to transform the training posture of the target object, a matrix-transformed training posture can be obtained.

[0045] This transformation process not only eliminates the pose deviation caused by camera angle or target object orientation problems, but also ensures the alignment between the target object pose and the camera coordinate system, providing a foundation for accurate determination of subsequent pose adjustment information.

[0046] The training pose after matrix transformation is used as the corrected training pose. The corrected training pose not only more accurately reflects the actual orientation of the target object in the training space, but also ensures its consistency with the Y-axis direction of the camera coordinate system.

[0047] In one embodiment, based on the human key point data, the training pose of the target object is determined, including: determining the torso key points of the target object based on the human key point data, and connecting the torso key points to obtain the torso connection of the target object; taking the torso connection as the training pose of the target object.

[0048] From the human key point data, the torso key points are selected, which usually cover the key coordinate points of the core positions such as shoulders, chest, waist and hips, and they jointly constitute the basic framework of the torso pose.

[0049] After selecting the torso key points, these key points need to be connected. By connecting the key points such as shoulders, chest, waist and hips, one or more line segments representing the torso of the target object are formed. These line segments depict the pose and direction of the torso, thereby combining into the torso connection of the target object.

[0050] The torso connection is taken as the training pose of the target object. Through the torso connection, the key pose information such as the inclination angle and the bending degree of the torso can be clearly observed, providing an explicit and quantifiable basis for subsequent pose correction and adjustment. Based on the training pose representation of the torso connection, the training pose adjustment method can more accurately identify the pose deviation.

[0051] In one embodiment, based on the corrected training pose and the standard pose of the training process, the pose adjustment information of the target object is determined, including: determining the difference angle between the corrected torso connection and the torso connection corresponding to the standard pose of the training process; based on the difference angle, determining the pose adjustment information of the target object.

[0052] In the training pose adjustment method, the difference angle between the corrected torso connection and the torso connection corresponding to the standard pose of the training process is determined, which can quantify the pose deviation. Two groups of data are obtained, including: the torso connection data after correction, which reflects the current torso pose of the target object after pose correction; the other group is the torso connection data corresponding to the standard pose of the training process, which represents the standard torso pose.

[0053] After obtaining these two sets of data, the angle difference between them can be calculated to determine the difference angle. The difference angle reflects the degree of deviation between the target object's current training posture and the standard posture, providing a quantitative basis for subsequent posture adjustments.

[0054] Based on this difference, the posture adjustment information of the target object can be further determined. This information can specifically include the specific body parts that need adjustment, such as the shoulders, chest, or waist; the direction of adjustment, such as forward, backward, left, or right; and the magnitude of adjustment, such as fine-tuning or large-scale adjustment.

[0055] In one embodiment, after obtaining the corrected training posture, the method further includes: acquiring dynamic feature information of the target object during the training process, the dynamic feature information including velocity information and acceleration information; and further correcting the corrected training posture based on the dynamic feature information.

[0056] During the training process of the target object, speed and acceleration information of the target object can be collected based on the sensors set in the target object.

[0057] Velocity information reveals how fast the target object moves during training, while acceleration information captures how quickly its velocity changes. This information can be used to understand the dynamic behavior of the target object.

[0058] For example, when a target object accelerates, its posture may change due to inertia. By using dynamic feature information, the method can capture this change and fine-tune the corrected training posture accordingly, ensuring that the posture always remains highly consistent with the actual motion state of the target object.

[0059] This further correction based on dynamic feature information not only improves the training effect, but also further enhances the accuracy of posture adjustment.

[0060] Specifically, for the soldier training process, various sensors such as accelerometers, gyroscopes, and pressure sensors are deployed on training grounds, weapons, and soldier training equipment to collect data in real time on the posture, force, and speed of soldiers' training movements, as well as data on the usage status and performance parameters of weapons and equipment. At the same time, high-definition cameras installed in all directions in the training area are used to automatically identify and collect data on the standardization of soldiers' training movements and teamwork through video image recognition technology. The system is also connected to the existing training management system of the military to obtain structured data such as training plans, personnel information, and assessment results.

[0061] In one embodiment, further correction of the corrected training posture based on the dynamic feature information includes: determining the inertia compensation vector of the target object during the training process based on the dynamic feature information; and further correcting the corrected training posture based on the inertia compensation vector.

[0062] Based on sensors installed in the target object, dynamic feature information of the target object during the training process can be captured, such as velocity information and acceleration information.

[0063] By analyzing these dynamic feature information, the potential attitude shift or adjustment requirements of the target object due to motion inertia can be calculated, and the inertia compensation vector can be further determined.

[0064] Based on a defined inertia compensation vector, the previously corrected training posture is further fine-tuned. This enables the training posture adjustment method to intelligently adapt to various changes in the target object during training, whether it's changes in velocity or fluctuations in acceleration, resulting in timely posture adjustments and significantly improving the accuracy and effectiveness of posture adjustment.

[0065] In one embodiment, the standard posture of the training process is obtained by connecting the key points of the torso of the standard execution object during the training process.

[0066] The standard posture during training is constructed by capturing and analyzing the key points of the torso of the target subject during training and connecting these key points. The standard posture provides a crucial benchmark for subsequent posture comparison and adjustment of the target subject, ensuring the accuracy and standardization of training.

[0067] The training posture adjustment device provided by the present invention is described below. The training posture adjustment device described below can be referred to in correspondence with the training posture adjustment method described above.

[0068] like Figure 2 As shown, the device includes: Key point extraction module 210 is used to detect human key points in the images of the target object during the training process and determine the human key point data of the target object; The pose determination module 220 is used to determine the training pose of the target object based on the human body key point data. The correction module 230 is used to determine the direction vector of the line connecting the midpoint of the shoulder and the midpoint of the hip of the target object based on the human body key point data, and to determine the orientation vector of the target object based on the direction vector of the line, and to correct the training posture based on the orientation vector to obtain the corrected training posture. The adjustment module 240 is configured to determine the posture adjustment information of the target object based on the corrected training posture and the standard posture of the training process, and the posture adjustment information is used for adjusting the training posture of the target object.

[0069] The training posture adjustment device provided by the application can obtain the human body key point data of the target object in the training process, and determine the training posture of the target object based on the human body key point data. The posture is corrected by calculating the orientation vector of the target object, which can effectively eliminate the posture error caused by the camera angle or the orientation deviation of the target object, and significantly improve the accuracy of the posture analysis. After obtaining the corrected training posture, the corrected training posture is compared with the standard posture of the training process, so as to quantitatively determine the posture adjustment information which can be used to accurately guide the process, and improve the accuracy of the posture adjustment.

[0070] In one embodiment, the correction module 230 is specifically configured to: correct the training posture based on the orientation vector to obtain a corrected training posture, including: determining the included angle between the orientation vector and the Y-axis direction of the camera coordinate system; constructing a rotation matrix based on the included angle, and transforming the training posture of the target object based on the rotation matrix to obtain a matrix-transformed training posture; the matrix-transformed training posture is taken as the corrected training posture.

[0071] In one embodiment, the posture determination module 220 is specifically configured to: determine the training posture of the target object based on the human body key point data, including: determining the trunk key point of the target object based on the human body key point data, and connecting the trunk key point to obtain a trunk line of the target object; the trunk line is taken as the training posture of the target object.

[0072] In one embodiment, the adjustment module 240 is specifically configured to: determine the posture adjustment information of the target object based on the corrected training posture and the standard posture of the training process, including: determining the difference angle between the corrected trunk line and the trunk line corresponding to the standard posture of the training process; determine the posture adjustment information of the target object based on the difference angle.

[0073] In one embodiment, the correction module 230 is further specifically configured to: after obtaining the corrected training posture, further including: acquire dynamic characteristic information of the target object in the training process, the dynamic characteristic information including speed information and acceleration information; correct the corrected training posture further based on the dynamic characteristic information.

[0074] In one embodiment, the correction module 230 is further specifically used for: correct the corrected training posture further based on the dynamic characteristic information, including: determine an inertia compensation vector of the target object in the training process based on the dynamic characteristic information; correct the corrected training posture further based on the inertia compensation vector.

[0075] In one embodiment, the key point extraction module 210 is specifically used for: perform human key point detection on images of the target object in the training process to obtain human key point data of the target object in the training process, including: perform feature extraction on the images of the target object in the training process based on a recurrent neural network to obtain a feature map of the target object; perform further feature extraction on the feature map based on a convolutional neural network to determine human key points of the target object.

[0076] In one embodiment, the posture determination module 220 is specifically used for: determine the standard posture of the training process based on connection of trunk key points of a standard execution object in the training process.

[0077] Figure 3 An example of a schematic diagram of an entity structure of an electronic device is shown in Figure 3 The electronic device can include a processor 310, a communications interface 320, a memory 330, and a communications bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other through the communications bus 340. The processor 310 can invoke logical instructions in the memory 330 to execute a training posture adjustment method, which includes performing human key point detection on images of a target object in a training process to determine human key point data of the target object; determine a training posture of the target object based on the human key point data; determine a line direction vector of a midpoint of a shoulder and a midpoint of a hip of the target object based on the human key point data, determine an orientation vector of the target object based on the line direction vector, and correct the training posture based on the orientation vector to obtain a corrected training posture; determine posture adjustment information of the target object based on the corrected training posture and the standard posture of the training process, the posture adjustment information being used for training posture adjustment of the target object.

[0078] In addition, the logical instructions in the memory 330 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the parts that contribute to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0079] On the other hand, the present application also provides a computer program product, which comprises a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program is executed by a processor, so that the computer can execute the training posture adjustment method provided by the above-mentioned methods. The method comprises: performing human key point detection on the image of the target object training process to determine the human key point data of the target object; determine the training posture of the target object based on the human key point data; determine a line direction vector of a midpoint of a shoulder and a midpoint of a hip of the target object based on the human key point data, determine an orientation vector of the target object based on the line direction vector, and correct the training posture based on the orientation vector to obtain a corrected training posture; determine posture adjustment information of the target object based on the corrected training posture and the standard posture of the training process, the posture adjustment information being used for training posture adjustment of the target object.

[0080] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the training posture adjustment method provided by any of the above methods, and the method comprises: performing human key point detection on an image of a target object training process to determine human key point data of the target object; Based on the human key point data, a training posture of the target object is determined; Based on the human key point data, a direction vector of a line connecting a midpoint of a shoulder and a midpoint of a hip of the target object is determined, and based on the direction vector, an orientation vector of the target object is determined, and based on the orientation vector, the training posture is corrected to obtain a corrected training posture; Based on the corrected training posture and a standard posture of the training process, posture adjustment information of the target object is determined, and the posture adjustment information is used for training posture adjustment of the target object.

[0081] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0082] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus necessary general hardware platforms, and of course, can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0083] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A training posture adjustment method, characterized by, The method comprises: human key point detection is performed on an image of a target object training process to determine human key point data of the target object; based on the human key point data, a training posture of the target object is determined; based on the human key point data, a direction vector of a line connecting a shoulder midpoint and a hip midpoint of the target object is determined, and based on the direction vector, an orientation vector of the target object is determined, and based on the orientation vector, the training posture is corrected to obtain a corrected training posture; based on the corrected training posture and a standard posture of the training process, posture adjustment information of the target object is determined, which is used for training posture adjustment of the target object.

2. The training posture adjustment method of claim 1, wherein, The method comprises: determining an included angle between the orientation vector and a Y-axis direction of a camera coordinate system; based on the included angle, a rotation matrix is constructed, and the training posture of the target object is transformed based on the rotation matrix to obtain a matrix-transformed training posture; the matrix-transformed training posture is taken as the corrected training posture.

3. The training posture adjustment method of claim 1, wherein, The method comprises: based on the human key point data, a trunk key point of the target object is determined, and the trunk key point is connected to obtain a trunk line of the target object; the trunk line is taken as the training posture of the target object.

4. The training posture adjustment method of claim 3, wherein, The method comprises: determining a difference angle between the corrected trunk line and a trunk line corresponding to the standard posture of the training process; based on the difference angle, the posture adjustment information of the target object is determined.

5. The training posture adjustment method of claim 1, wherein, After obtaining the corrected training posture, the method further comprises: obtaining dynamic characteristic information in the training process of the target object, the dynamic characteristic information comprising speed information and acceleration information; based on the dynamic characteristic information, the corrected training posture is further corrected.

6. The training posture adjustment method of claim 5, wherein, The method comprises: based on the dynamic characteristic information, an inertia compensation vector of the target object in the training process is determined; based on the inertia compensation vector, the corrected training posture is further corrected.

7. The training posture adjustment method of claim 1, wherein, The standard posture of the training process is obtained by connecting trunk key points of a standard execution object in the training process.

8. A training posture adjustment apparatus characterized by comprising: The method comprises: a key point extraction module configured to perform human key point detection on an image of a target object training process to determine human key point data of the target object; a posture determination module configured to determine a training posture of the target object based on the human key point data; The correction module is configured to determine a line direction vector of a midpoint of a shoulder and a midpoint of a hip of the target object based on the human key point data, determine an orientation vector of the target object based on the line direction vector, and correct the training posture based on the orientation vector to obtain a corrected training posture. The adjustment module is configured to determine posture adjustment information of the target object based on the corrected training posture and a standard posture of the training process, and the posture adjustment information is used for training posture adjustment of the target object.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the training posture adjustment method according to any one of claims 1 to 7 when executing the computer program.

10. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the training posture adjustment method according to any one of claims 1 to 7.

Images