Method and apparatus for determining exercise intensity across individuals

Abstract

The application provides a method and device for determining the intensity of an individual's exercise, which comprises: obtaining individual exercise information; inputting the individual exercise information into an individual exercise intensity determination model to obtain an individual exercise intensity grade; wherein the individual exercise intensity determination model is obtained based on supervised source domain samples and is tested for performance by unsupervised target domain samples; the source domain samples comprise an individual exercise information sample training set and corresponding individual exercise intensity grades, and the target domain samples comprise an individual exercise information sample test set and corresponding individual exercise intensity grades. The method and device are used to solve the defects of low accuracy and high data collection cost in the prior art, and achieve high accuracy and low cost in the identification of the intensity of an individual's exercise.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sports data analysis, in particular to a cross-individual exercise intensity determination method and device. BACKGROUND

[0002] With the continuous improvement of living standards, the development of convenience, and the increasing length of sedentary behavior, people are severely lacking in physical activity, and the incidence of various diseases is also rising, posing a great threat to people's health. Physical activity not only effectively improves physical fitness, but also promotes mental health, reduces sedentary behavior, and injects vitality into the lives of the general public, reducing the incidence of diseases caused by insufficient daily physical activity.

[0003] In daily life, people often engage in different speed walking, running and other full-body movements. These activities have different stimulating abilities on the body, and a certain amount of physical activity can effectively promote the healthy development of the body. However, due to individual differences, different bodies have different tolerances for these activities, so adjustments need to be made according to actual conditions when engaging in these physical activities. Exercise intensity is an important measure of the normal implementation of exercise prescriptions, and it is one of the important factors for whether to continue or adjust the exercise prescription. During exercise, some athletes cannot intuitively understand the intensity of the exercise, which can lead to ineffective exercise or exercise injuries and sudden death. Research has found that different exercise intensities can affect the incidence of depression and anxiety, and long-term low-intensity exercise (such as sitting or lying) can increase the probability of depression and anxiety. Therefore, real-time monitoring of exercise intensity is particularly important, as it is an effective guarantee for life safety and physical health.

[0004] Due to individual differences, each person's physical indicators are different, resulting in different exercise intensities for different bodies when performing the same exercise. The collected data will greatly reduce the accuracy of algorithm recognition when facing cross-individuals. Therefore, a large amount of data is usually collected for the research object to train the model to pursue the generalization ability of the model. Although this method can improve the recognition rate, it requires collecting data from numerous user groups and processing it, which will cost a huge amount of time and money. SUMMARY

[0005] The present application provides a cross-individual exercise intensity determination method and device to solve the problem of low accuracy and high data collection cost in the prior art, achieving high accuracy and low cost in cross-individual exercise intensity recognition.

[0006] The present application provides a cross-individual exercise intensity determination method, comprising:

[0007] Obtaining individual exercise information;

[0008] inputting the individual motion information into an individual motion intensity determination model to obtain an individual motion intensity level;

[0009] The individual motion intensity determination model is trained based on supervised source domain samples and tested in performance by unsupervised target domain samples.

[0010] According to the method for determining the motion intensity across individuals, the training of the individual motion intensity determination model comprises:

[0011] obtaining individual motion information samples corresponding to individual motion intensity levels respectively;

[0012] The individual motion information samples are divided into supervised source domain samples and unsupervised target domain samples.

[0013] The individual motion intensity determination model is trained based on the individual motion information sample training set of the source domain samples and the corresponding individual motion intensity levels as labels.

[0014] The individual motion information sample test set of the target domain samples is input into the individual motion intensity determination model to obtain an individual motion intensity level test result.

[0015] The individual motion intensity level test result is compared with the individual motion intensity level corresponding to the individual motion information sample test set to test the performance of the individual motion intensity determination model.

[0016] According to the method for determining the motion intensity across individuals, the individual motion information samples comprise human body basic data and motion data.

[0017] According to the method for determining the motion intensity across individuals, before the individual motion information samples are divided into a training set and a test set, the method further comprises:

[0018] The high-frequency noise of the three-axis acceleration and the three-axis angular velocity in the individual motion information samples is filtered out, and the features of the three-axis acceleration and the three-axis angular velocity are extracted.

[0019] According to the application, a cross-individual exercise intensity determination method is provided, wherein the individual exercise intensity level is divided by measuring the percentage of maximum oxygen uptake.

[0020] According to the application, a cross-individual exercise intensity determination method is provided, wherein the individual exercise intensity determination model is trained by the DSAN algorithm.

[0021] According to the application, a cross-individual exercise intensity determination method is provided, wherein the loss function of the individual exercise intensity determination model is BCELoss, the optimizer is AdamW, the batch_size is set to 32, the activation function is set to leaky_relu, the epoch is set to 50, the learning rate of the individual exercise intensity determination model is adjusted by the self-adaptive learning rate method, and the weight is adjusted by the self-adaptive decay weight method.

[0022] The application further provides a cross-individual exercise intensity determination device, comprising:

[0023] An information acquisition module is configured to acquire individual exercise information.

[0024] An intensity determination module is configured to input the individual exercise information into an individual exercise intensity determination model to obtain individual exercise intensity.

[0025] The individual exercise intensity determination model is trained based on supervised source domain samples and tested based on unsupervised target domain samples.

[0026] The application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the cross-individual exercise intensity determination method according to any one of the above.

[0027] The application further provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable on a processor to implement the cross-individual exercise intensity determination method according to any one of the above.

[0028] This invention provides a method and apparatus for determining cross-individual movement intensity. It acquires individual movement information and inputs it into an individual movement intensity determination model to obtain the individual's movement intensity level, thereby determining the current individual's movement intensity. The individual movement intensity determination model is trained using supervised source domain samples and its performance is tested using unsupervised target domain samples. The source domain samples include: a training set of individual movement information samples and their corresponding individual movement intensity levels; the target domain samples include: a test set of individual movement information samples and their corresponding individual movement intensity levels. This invention uses the individual movement information sample training set and its corresponding individual movement intensity level labels as the supervised source domain to establish the individual movement intensity determination model. By directly establishing the determination model on an individual basis for identification, it solves the problem of low cross-individual movement intensity recognition rate caused by individual differences and improves the accuracy of cross-individual movement intensity recognition. Meanwhile, the individual motion information sample test set and its corresponding individual motion intensity level are used as the unsupervised target domain to test the individual motion intensity judgment model. That is, the individual motion intensity judgment model outputs the test results of the individual motion information sample test set, and then the test results are compared with the individual motion intensity level corresponding to the individual motion information sample test set to test the performance of the individual motion information sample test set. In this way, the present invention does not require a lot of data annotation during the model construction process, which greatly reduces the cost of data collection. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0030] Figure 1 This is a flowchart illustrating the cross-individual motion intensity determination method provided by the present invention;

[0031] Figure 2 This is a schematic diagram of the cross-individual motion intensity determination device provided by the present invention;

[0032] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0034] The following is combined Figure 1 The present invention describes a method for determining cross-individual motion intensity according to a first embodiment.

[0035] like Figure 1 As shown, this embodiment of a method for determining cross-individual exercise intensity specifically includes the following steps (the numbering of each step in this invention is only for distinguishing steps and does not limit the specific execution order of each step):

[0036] Step S1: Obtain individual motion information.

[0037] Acquire various data points related to individual movement and use them as input for the individual movement intensity determination model.

[0038] Step S2: Input the individual movement information into the individual movement intensity determination model to obtain the individual movement intensity level.

[0039] The acquired individual movement information data is input into the individual movement intensity judgment model to determine the individual movement intensity level, which refers to the degree of physiological stimulation to the body during the subject's exercise.

[0040] The individual motion intensity determination model is trained based on supervised source domain samples and its performance is tested using unsupervised target domain samples. The source domain samples include: an individual motion information sample training set and its corresponding individual motion intensity level. The target domain samples include: an individual motion information sample test set and its corresponding individual motion intensity level.

[0041] An individual motion intensity determination model is established using a training set of individual motion information samples and their corresponding individual motion intensity level labels as the supervised source domain. Simultaneously, the individual motion intensity determination model is tested using a test set of individual motion information samples and their corresponding individual motion intensity levels as the unsupervised target domain. Specifically, the model outputs test results from the individual motion information sample test set, and then compares these results with the individual motion intensity levels corresponding to the individual motion information sample test set to evaluate the performance of the individual motion information sample test set.

[0042] The first embodiment of the present application provides a cross-individual exercise intensity determination method. Individual exercise information is obtained and input into an individual exercise intensity determination model to obtain an individual exercise intensity level to determine the current individual exercise intensity. The individual exercise intensity determination model is trained based on supervised source domain samples and tested for performance based on unsupervised target domain samples. The source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity levels. The present application uses the individual exercise information sample training set and corresponding individual exercise intensity level labels as a supervised source domain to establish an individual exercise intensity determination model, directly establishes a determination model for identification based on individuals, solves the problem of low cross-individual exercise intensity identification rate caused by individual differences, and improves the accuracy of cross-individual exercise intensity identification. At the same time, the individual exercise information sample test set and corresponding individual exercise intensity levels are used as an unsupervised target domain to test the individual exercise intensity determination model, i.e., the test results of the individual exercise information sample test set are output by the individual exercise intensity determination model, and then the test results are compared based on the corresponding individual exercise intensity levels of the individual exercise information sample test set to test the performance of the individual exercise information sample test set. In the above manner, the present application does not need to perform a large amount of data labeling in the model construction process, thereby greatly reducing the data acquisition cost.

[0043] In the present embodiment, the training of the individual exercise intensity determination model includes:

[0044] Obtaining individual exercise information samples corresponding to individual exercise intensity levels;

[0045] Dividing the individual exercise information samples into supervised source domain samples and unsupervised target domain samples. The source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity levels;

[0046] Training an individual exercise intensity determination model based on the individual exercise information sample training set and corresponding individual exercise intensity levels of the source domain samples as labels;

[0047] Inputting the individual exercise information sample test set of the target domain samples into the individual exercise intensity determination model to obtain individual exercise intensity level test results;

[0048] Comparing the individual exercise intensity level test results based on the corresponding individual exercise intensity levels of the individual exercise information sample test set to test the performance of the individual exercise intensity determination model.

[0049] First, the individual motion information samples required for model establishment are acquired, and each individual motion information sample has a corresponding individual motion intensity level, that is, the individual motion information sample is used as the input of the model, and the corresponding individual motion intensity level is used as the output of the model. Then, the individual motion information samples are divided into supervised source domain samples and unsupervised target domain samples according to the functional use, the source domain samples include the individual motion information sample training set and the corresponding individual motion intensity level, and the target domain samples include the individual motion information sample test set and the corresponding individual motion intensity level. The source domain samples are used for model training, and the individual motion information sample training set and the corresponding individual motion intensity level based on the source domain samples are used as labels to train the individual motion intensity determination model. After obtaining the individual motion intensity determination model, the target domain samples are used for model testing, the individual motion information sample test set of the target domain samples is input into the individual motion intensity determination model, the individual motion intensity level test result is obtained, then the individual motion intensity level test result is compared according to the corresponding individual motion intensity level of the individual motion information sample test set, so as to test the performance of the individual motion intensity determination model, and the model construction is completed after the test result reaches the preset standard. The individual motion intensity determination model constructed by the above process directly establishes a determination model for identification in units of individuals, solves the problem of low cross-individual motion intensity identification rate caused by individual differences, and improves the accuracy of cross-individual motion intensity identification. By combining the supervised source domain samples and the unsupervised target domain, a large amount of data labeling is not required in the model construction process, and the cost of data acquisition is reduced.

[0050] In the embodiment, the individual motion information sample includes human body basic data and motion data, the human body basic data includes height, age, gender, weight and resting heart rate, and the motion data includes three-axis acceleration, three-axis angular velocity and motion heart rate.

[0051] The individual motion information sample includes human body basic data and motion data of the individual itself, and the human body basic data includes height, age, gender, weight and resting heart rate, and the motion data includes three-axis acceleration, three-axis angular velocity and motion heart rate. The above data test method is convenient, low in cost, non-invasive, and can realize real-time monitoring, for example, some wearable devices can be used to acquire the data, and the discomfort of the monitored person is not increased. Through analysis of the above multi-dimensional data, the accuracy of cross-individual motion intensity determination can be ensured.

[0052] In the embodiment, before the individual motion information sample is divided into the training set and the test set, the method further includes:

[0053] The high-frequency noise of the three-axis acceleration and the three-axis angular velocity in the individual motion information sample is filtered out, and the features of the three-axis acceleration and the three-axis angular velocity are extracted.

[0054] Before dividing the individual motion information samples into a training set and a test set to build a model, high-frequency noise of the three-axis acceleration and the three-axis angular velocity in the individual motion information samples needs to be filtered out, and features of the three-axis acceleration and the three-axis angular velocity need to be extracted to ensure the accuracy of the data. Studies have shown that the frequency of the human body during movement is 0-20 Hz, so low-pass filtering is performed in this embodiment, and the cutoff frequency is set to 20 Hz to filter out high-frequency noise of the three-axis acceleration and the three-axis angular velocity. Then, features of the three-axis acceleration and the three-axis angular velocity are extracted from the time domain and the frequency domain by setting a sliding window of an appropriate size, and the extracted features, the basic information of the human body, and the corresponding labels are combined for subsequent model building.

[0055] In this embodiment, the individual motion intensity level is divided by measuring the percentage of maximum oxygen uptake.

[0056] In the process of determining the individual motion level, the pace of the maximum oxygen uptake of the personnel is measured, and the five levels divided by the percentage of the maximum oxygen uptake of the Sports Medicine Industry Association are re-divided into three motion intensity levels of low (37-45% of the maximum oxygen uptake), medium (46-63% of the maximum oxygen uptake), and high (64-90% of the maximum oxygen uptake) because the lowest level is normal life and the highest level is difficult for ordinary people to achieve. Finally, the data of the three motion intensity levels are collected by the sensor to determine the individual motion intensity level, so that it is more suitable for the application scenario in this embodiment.

[0057] In this embodiment, the individual motion intensity determination model is trained by the DSAN algorithm.

[0058] The DSAN (Deep Subdomain Adaption Network) algorithm, unlike the traditional domain adaptation algorithm, is a more fine-grained algorithm that divides different categories into sub-domains, aligns the sub-domains in the target domain and the source domain, and then discriminates. The DSAN algorithm is simple and efficient, does not require adversarial training, has fast convergence speed, and can achieve the effects of high accuracy and low cost in this embodiment.

[0059] In this embodiment, the loss function of the individual motion intensity determination model is BCELoss, the optimizer is AdamW, the batch_size is set to 32, the activation function is set to leaky_relu, the epoch is set to 50, the individual motion intensity determination model adjusts the learning rate by the self-adaptive learning rate method, and adjusts the weight by the self-adaptive weight decay method.

[0060] In the process of setting the model parameters, the loss function of the individual exercise intensity determination model is BCELoss (Binary Cross Entropy Loss), and the optimizer is AdamW (Adam Weight Decay) to solve the overfitting problem that may occur when training a deep neural network. The batch_size, i.e., the number of data samples fetched at a time, is set to 32 to ensure the training speed and model optimization. The activation function is set to leaky_relu to solve the zero gradient problem in the negative input case and alleviate the dead rule problem. The number of epochs is 50, the individual exercise intensity determination model adjusts the learning rate through an adaptive learning rate method and adjusts the weight through an adaptive weight decay method to ensure the model training effect.

[0061] The cross-individual exercise intensity determination device provided by the present application is described below. The cross-individual exercise intensity determination device described below can be referred to in conjunction with the cross-individual exercise intensity determination method described above.

[0062] As shown in Figure 2 The second embodiment of the present application also provides a multi-cross-individual exercise intensity determination device, which comprises:

[0063] The information acquisition module 210 is configured to acquire individual exercise information.

[0064] The intensity determination module 220 is configured to input the individual exercise information into an individual exercise intensity determination model to obtain individual exercise intensity.

[0065] The individual exercise intensity determination model is trained based on supervised source domain samples and tested for performance through unsupervised target domain samples. The source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity levels.

[0066] The second embodiment of the present application provides a cross-individual exercise intensity determination device. Individual exercise information is obtained and input into an individual exercise intensity determination model to obtain an individual exercise intensity level to determine the current individual exercise intensity. The individual exercise intensity determination model is trained based on supervised source domain samples and tested for performance based on unsupervised target domain samples. The source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity levels. The present application uses the individual exercise information sample training set and corresponding individual exercise intensity level labels as a supervised source domain to establish an individual exercise intensity determination model, directly establishes a determination model for individual units for identification, solves the problem of low cross-individual exercise intensity identification rate caused by individual differences, and improves the accuracy of cross-individual exercise intensity identification. At the same time, the individual exercise information sample test set and corresponding individual exercise intensity levels are used as an unsupervised target domain to test the individual exercise intensity determination model, i.e. the test results of the individual exercise information sample test set are output by the individual exercise intensity determination model, and then the test results are compared by the corresponding individual exercise intensity levels of the individual exercise information sample test set to test the performance of the individual exercise information sample test set. Through the above method, the present application does not need to perform a large amount of data labeling in the model construction process, greatly reducing the cost of data acquisition.

[0067] Figure 3 An example of a schematic diagram of the physical 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. The processor 310, the communications interface 320, and the memory 330 can communicate with each other through the communications bus 340. The processor 310 can invoke logical instructions in the memory 330 to execute a cross-individual exercise intensity determination method. The method includes:

[0068] Obtaining individual exercise information.

[0069] Inputting the individual exercise information into an individual exercise intensity determination model to obtain an individual exercise intensity level.

[0070] The individual exercise intensity determination model is trained based on supervised source domain samples and tested for performance based on unsupervised target domain samples. The source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity levels.

[0071] In addition, the logic 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, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The foregoing 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.

[0072] In another aspect, 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 can be executed by a processor to enable a computer to execute the cross-individual exercise intensity determination method provided by the above-mentioned methods, which comprises:

[0073] Obtaining individual exercise information.

[0074] Inputting the individual exercise information into an individual exercise intensity determination model to obtain an individual exercise intensity level.

[0075] The individual exercise intensity determination model is trained based on supervised source domain samples and tested for performance by unsupervised target domain samples; the source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity levels.

[0076] In another aspect, the present application also provides a non-transitory computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the cross-individual exercise intensity determination method provided by the above-mentioned methods, which comprises:

[0077] Obtaining individual exercise information.

[0078] Inputting the individual exercise information into an individual exercise intensity determination model to obtain an individual exercise intensity level.

[0079] The individual movement intensity determination model is trained based on supervised source domain samples and tested by unsupervised target domain samples; the source domain samples include an individual movement information sample training set and corresponding individual movement intensity grades, and the target domain samples include an individual movement information sample test set and corresponding individual movement intensity grades.

[0080] The device embodiments described above are merely illustrative, wherein the units illustrated as separate components can or can not be physically separated, and the components illustrated as units can or can not be physical units, i.e., 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.

[0081] 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 and necessary universal 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 software products, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and include a number 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.

[0082] 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 part of the 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 method of determining exercise intensity across individuals, characterized by, The method comprises the following steps: obtaining individual exercise information; inputting the individual exercise information into an individual exercise intensity determination model to obtain an individual exercise intensity level; wherein the individual exercise intensity determination model is trained based on supervised source domain samples and tested for performance by unsupervised target domain samples; the source domain samples comprise an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples comprise an individual exercise information sample test set and corresponding individual exercise intensity levels; the individual exercise information sample comprises human body basic data and exercise data, the human body basic data comprises height, age, gender, weight and resting heart rate, and the exercise data comprises three-axis acceleration, three-axis angular velocity and exercise heart rate; the training of the individual exercise intensity determination model comprises the following steps: obtaining individual exercise information samples, each of which corresponds to an individual exercise intensity level; dividing the individual exercise information samples into supervised source domain samples and unsupervised target domain samples, the source domain samples comprising an individual exercise information sample training set and corresponding individual exercise intensity levels, and the target domain samples comprising an individual exercise information sample test set and corresponding individual exercise intensity levels; training an individual exercise intensity determination model based on the individual exercise information sample training set of the source domain samples and the corresponding individual exercise intensity levels as labels; inputting the individual exercise information sample test set of the target domain samples into the individual exercise intensity determination model to obtain individual exercise intensity level test results; comparing the individual exercise intensity level test results according to the individual exercise intensity levels corresponding to the individual exercise information sample test set to test the performance of the individual exercise intensity determination model.

2. The method according to claim 1, wherein The individual exercise information sample comprises human body basic data and exercise data, the human body basic data comprises height, age, gender, weight and resting heart rate, and the exercise data comprises three-axis acceleration, three-axis angular velocity and exercise heart rate.

3. The method according to claim 2, wherein Before dividing the individual exercise information samples into a training set and a test set, the method further comprises the following steps: filtering out high-frequency noise of three-axis acceleration and three-axis angular velocity in the individual exercise information samples and extracting features of the three-axis acceleration and three-axis angular velocity.

4. The method according to claim 2, wherein The individual exercise intensity level is divided by measuring the percentage of maximum oxygen uptake.

5. The method according to any one of claims 1 to 4, wherein The individual exercise intensity determination model is trained by a DSAN algorithm.

6. The method according to claim 5, wherein The loss function of the individual exercise intensity determination model is BCELoss, the optimizer is AdamW, the batch_size is set to 32, the activation function is set to leaky_relu, the epoch is set to 50, the learning rate of the individual exercise intensity determination model is adjusted by an adaptive learning rate method, and the weight is adjusted by an adaptive decay weight method.

7. An exercise intensity determination device across individuals, characterized by, The method comprises the following steps: an information acquisition module for acquiring individual exercise information; an intensity determination module for inputting the individual exercise information into an individual exercise intensity determination model to obtain an individual exercise intensity. The individual exercise intensity determination model is trained based on supervised source domain samples and tested by unsupervised target domain samples; the source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity grades, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity grades; the individual exercise information sample includes human body basic data and exercise data, the human body basic data includes height, age, gender, weight, and resting heart rate, and the exercise data includes three-axis acceleration, three-axis angular velocity, and exercise heart rate; The device is further configured to: obtain individual exercise information samples corresponding to individual exercise intensity grades; divide the individual exercise information samples into supervised source domain samples and unsupervised target domain samples, wherein the source domain samples include an individual exercise information sample training set and corresponding individual exercise intensity grades, and the target domain samples include an individual exercise information sample test set and corresponding individual exercise intensity grades; train an individual exercise intensity determination model based on the individual exercise information sample training set and corresponding individual exercise intensity grades of the source domain samples as labels; input the individual exercise information sample test set of the target domain samples into the individual exercise intensity determination model to obtain individual exercise intensity grade test results; and compare the individual exercise intensity grade test results with the individual exercise intensity grades corresponding to the individual exercise information sample test set to test the performance of the individual exercise intensity determination model.

8. 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 executes the program to implement the cross-individual exercise intensity determination method of any one of claims 1 to 6. 9.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 cross-individual exercise intensity determination method of any one of claims 1 to 6.

Images