WO2023216266A1

WO2023216266A1 - Heat consumption estimation method and device and storage medium

Info

Publication number: WO2023216266A1
Application number: PCT/CN2022/092873
Authority: WO
Inventors: 皮特; 姚丽峰; 高国松
Original assignee: 北京小米移动软件有限公司
Priority date: 2022-05-13
Filing date: 2022-05-13
Publication date: 2023-11-16
Also published as: CN117412700A

Abstract

The present disclosure relates to a heat consumption estimation method and device and a storage medium. The method comprises: acquiring the heart rate and the respiratory rate of a target user at a current moment; according to the heart rate at the current moment, determining motion state feature information of the target user at the current moment, the motion state feature information being used for representing whether the target user is in a motion steady-state stage or a motion recovery stage at the current moment; according to the heart rate at the current moment, the respiratory rate at the current moment and a heat estimation model corresponding to the motion state feature information at the current moment, obtaining a heat consumption increment of the target user in a current time period, wherein heat estimation models corresponding to different motion state feature information are different, and the current time period is a time period from the previous moment of the current moment to the current moment; and according to the heat consumption increment, determining total heat consumption information of the target user in a motion process. By means of the described technical scheme, estimation of heat consumption is more accurate.

Description

Calorie consumption estimation method, device and storage medium

Technical field

The present disclosure relates to the field of human-computer interaction, and in particular, to a method, device and storage medium for estimating heat consumption.

Background technique

With the continuous development of technology, wearable devices have gradually entered people's daily lives. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes. Its product forms are diverse, such as smart watches, smart bracelets, smart glasses, smart helmets, etc. There are many sensors in wearable devices that can collect physiological signals of the human body, such as heart rate and other signals, to monitor the user's health indicators. Among them, in the monitoring of exercise indicators, the user's caloric consumption during exercise is the calorie consumption. , has become the object of increasing attention. However, the calculation of calorie consumption in related technologies is not accurate enough.

Contents of the invention

In order to overcome problems existing in related technologies, the present disclosure provides a method, device and storage medium for estimating heat consumption.

According to a first aspect of an embodiment of the present disclosure, a caloric consumption estimation method is provided, including:

Obtain the heart rate and respiratory rate of the target user at the current moment; determine the motion state feature information of the target user at the current moment based on the heart rate at the current moment, wherein the motion state feature information is used to characterize the target Whether the user is in the steady state stage of exercise or the recovery stage of exercise at the current moment; according to the heart rate at the current moment, the respiratory frequency at the current moment, and the heat estimation model corresponding to the motion state characteristic information at the current moment, the result is obtained Calorie consumption increment of the target user in the current period, wherein the caloric estimation models corresponding to different motion state characteristic information are different, and the current period is the period from the previous moment of the current moment to the current moment. ; According to the calorie consumption increment, determine the total calorie consumption information of the target user during exercise, wherein the exercise process includes the process from the start time of exercise to the current time.

Optionally, based on the calorie estimation model corresponding to the heart rate at the current moment, the respiratory rate at the current moment, and the motion state feature information at the current moment, the increase in calorie consumption of the target user in the current period is obtained. The quantity includes: if the target user is in the steady-state stage of exercise at the current moment, then according to the heart rate at the current moment, the respiratory frequency at the current moment and the steady-state stage caloric estimation model, the target user is obtained in the steady-state stage. Calorie consumption increment in the current period; if the target user is in the exercise recovery stage at the current moment, then based on the heart rate at the current moment, the respiratory frequency at the current moment and the recovery stage caloric estimation model, the The incremental calorie consumption of the target user during the current period.

Optionally, the method further includes: obtaining the physiological characteristic information of the target user; and obtaining the target user based on the heart rate at the current moment, the respiratory frequency at the current moment and the steady-state stage heat estimation model. The increment of caloric consumption in the current period includes: inputting the heart rate at the current moment, the respiratory rate at the current moment, and the physiological characteristic information of the target user into the steady-state stage caloric estimation model to obtain the The caloric consumption increment output by the caloric estimation model in the steady-state phase; the heat consumption increment of the target user in the current period is obtained based on the heart rate at the current moment, the respiratory rate at the current moment and the caloric estimation model in the recovery phase. The increment of caloric consumption within the period includes: inputting the heart rate at the current moment, the respiratory rate at the current moment, and the physiological characteristic information of the target user into the recovery phase caloric estimation model to obtain the recovery phase caloric value. Estimate the incremental caloric expenditure output from the model.

Optionally, the steady-state stage caloric estimation model is trained by: obtaining the heart rate, respiratory rate, actual caloric consumption and physiological characteristic information of the user in the steady-state stage of exercise; The heart rate, respiratory rate, actual calorie consumption and physiological characteristic information are input into the deep learning network model, and the deep learning network model is trained to obtain the steady-state stage heat estimation model.

Optionally, the recovery phase caloric estimation model is trained in the following manner: obtaining the heart rate, breathing frequency, actual caloric consumption and physiological characteristic information of the user in the exercise recovery phase; Respiratory frequency, actual caloric consumption and physiological characteristic information are input into the deep learning network model, and the deep learning network model is trained to obtain the recovery stage caloric estimation model.

Optionally, determining the motion state characteristic information of the target user at the current moment according to the heart rate at the current moment includes: inputting the heart rate at the current moment into a motion state recognition model to obtain the The motion state feature information output by the motion state recognition model.

Optionally, the movement state recognition model is trained in the following manner: the heart rate when the user is in the steady state stage of movement and the label used to characterize that the heart rate corresponds to the steady state stage of movement, and when the user is in the recovery stage of movement. The heart rate and the label used to characterize the correspondence between the heart rate and the exercise recovery stage are used as the input of the classifier, and the classifier is trained to obtain the movement state recognition model.

Optionally, both the steady-state stage heat estimation model and the recovery stage heat estimation model are gated cycle unit models.

Optionally, the method further includes: after the target user finishes exercising, determine the current basal metabolic information of the target user based on the caloric consumption of the target user during this exercise, and display the current basal metabolic information. , wherein the current basal metabolic information is greater than the basal metabolic information of the target user before this exercise.

According to a second aspect of the embodiment of the present disclosure, a heat consumption estimating device is provided, including:

The frequency acquisition module is used to obtain the heart rate and respiratory frequency of the target user at the current moment; the state determination module is used to determine the motion state characteristic information of the target user at the current moment according to the heart rate at the current moment, wherein, The motion state characteristic information is used to characterize whether the target user is in a steady state phase or a recovery phase of motion at the current moment; an increment determination module is used to determine whether the target user is in a steady state phase or a recovery phase based on the heart rate at the current moment and the respiration at the current moment. The frequency and the caloric estimation model corresponding to the motion state characteristic information at the current moment are used to obtain the caloric consumption increment of the target user in the current period. The caloric estimation models corresponding to different motion state characteristic information are different. The current period is the period from the previous moment of the current moment to the current moment; the total calorie determination module is used to determine the total calorie consumption information of the target user during exercise according to the calorie consumption increment, Wherein, the movement process includes the process from the movement start time to the current time.

Optionally, the increment determination module includes: a first increment determination sub-module, configured to determine if the target user is in a steady state stage of exercise at the current moment according to the heart rate at the current moment, the current The respiratory frequency and steady-state stage caloric estimation model at each moment are used to obtain the caloric consumption increment of the target user in the current period; the second increment determination sub-module is used to determine if the target user is in exercise recovery at the current moment. stage, then based on the heart rate at the current moment, the respiratory frequency at the current moment and the recovery stage caloric estimation model, the caloric consumption increment of the target user in the current period is obtained.

Optionally, the device further includes: an information acquisition module, used to obtain the physiological characteristic information of the target user; and the first increment determination sub-module is used to: combine the heart rate at the current moment, the heart rate at the current moment, The respiratory frequency and the physiological characteristic information of the target user are input into the steady-state stage heat estimation model, and the heat consumption increment output by the steady-state stage heat estimation model is obtained; the second increment determiner The module is configured to: input the heart rate at the current moment, the respiratory rate at the current moment, and the physiological characteristic information of the target user into the recovery phase heat estimation model, and obtain all the values output by the recovery phase heat estimation model. The increase in caloric consumption.

Optionally, the steady-state stage caloric estimation model is trained through the following modules: a first acquisition module, used to obtain the heart rate, respiratory rate, actual caloric consumption and physiological characteristic information of the user in the steady-state stage of exercise; A training module for inputting the heart rate, respiratory rate, actual calorie consumption and physiological characteristic information of the user in the steady state stage of exercise into the deep learning network model, and training the deep learning network model to obtain the stable state. State stage heat estimation model.

Optionally, the recovery phase caloric estimation model is obtained by training through the following modules: a second acquisition module, used to acquire the heart rate, breathing frequency, actual caloric consumption and physiological characteristic information of the user in the exercise recovery phase; second training Module, used to input the heart rate, respiratory rate, actual calorie consumption and physiological characteristic information of the user in the exercise recovery stage into the deep learning network model, and train the deep learning network model to obtain the calorie estimate in the recovery stage. Model.

Optionally, the state determination module is configured to input the heart rate at the current moment into a motion state recognition model to obtain the motion state feature information output by the motion state recognition model.

Optionally, the movement state recognition model is trained through the following modules: a third training module, used to combine the user's heart rate when in the steady state stage of movement and a label used to represent that the heart rate corresponds to the steady state stage of movement; And the heart rate of the user when he is in the exercise recovery stage and the label used to characterize the correspondence between the heart rate and the exercise recovery stage are used as inputs of the classifier, and the classifier is trained to obtain the movement state recognition model.

Optionally, the device further includes: a basal metabolism determination module, configured to determine the current basal metabolism information of the target user based on the calorie consumption of the target user's current exercise after the target user finishes exercising, and The current basal metabolic information is displayed, wherein the current basal metabolic information is greater than the target user's basal metabolic information before this exercise.

According to a third aspect of an embodiment of the present disclosure, a device for estimating heat consumption is provided, including: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to: execute the first embodiment of the present disclosure. One aspect provides the steps of the method.

According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored. When the program instructions are executed by a processor, the steps of the caloric consumption estimation method provided by the first aspect of the present disclosure are implemented. .

The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects:

Through the above technical solution, the heart rate of the target user at the current moment is first obtained, and then based on the heart rate, the motion state characteristic information of the target user at the current moment is determined. The motion state characteristic information is used to characterize whether the target user is currently in the steady state stage of exercise or recovery. stage. According to the heat estimation model corresponding to the heart rate at the current moment, the respiratory rate at the current moment, and the motion state feature information at the current moment, the calorie consumption increment of the target user in the current period is obtained. Among them, the heat estimation model corresponding to different motion state feature information is Different, considering that the characteristics of calorie consumption are different in the steady-state stage of exercise and the recovery stage of exercise, therefore considering what exercise stage the target user is in, using different models for calorie estimation in different exercise stages can make the estimation of calorie consumption more accurate.

It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and do not limit the present disclosure.

Description of the drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the principles of the disclosure.

FIG. 1 is a flow chart of a method for estimating heat consumption according to an exemplary embodiment.

FIG. 2 is a flow chart of a method for estimating heat consumption according to an exemplary embodiment.

FIG. 3 is a schematic flowchart of a heat consumption estimation process according to an exemplary embodiment.

Figure 4 is a schematic diagram of model training according to an exemplary embodiment.

Figure 5 is a schematic diagram of the internal structure of the gated cycle unit model.

FIG. 6 is a block diagram of a heat consumption estimating device according to an exemplary embodiment.

FIG. 7 is a block diagram of a device for estimating heat consumption according to an exemplary embodiment.

Detailed ways

Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with aspects of the disclosure as detailed in the appended claims.

It should be noted that all actions to obtain signals, information or data in this application are performed under the premise of complying with the corresponding data protection regulations and policies of the country where the location is located, and with authorization from the owner of the corresponding device.

Currently, there are mainly the following methods for estimating a user's caloric consumption during exercise: heart rate-based linear regression method, heart rate-based machine learning model or deep learning model, and acceleration-based caloric consumption estimation. Among them, in the linear regression method based on heart rate, different models need to be regressed in different sports scenarios (such as walking, running, climbing, cycling, swimming, playing ball, etc.). For example, when a user is running in the morning After completing the mountain climbing exercise, you need to turn on the outdoor running option before the morning run. The wearable device uses the linear regression model corresponding to outdoor running to estimate the calorie consumption. After the morning run, the user needs to end the running exercise and reopen the mountain climbing option. , the wearable device uses a linear regression model corresponding to mountain climbing to estimate heat consumption. In this way, when switching sports scenes, different regression models need to be transformed to estimate heat consumption, and the scene adaptability is poor. Although heart rate-based machine learning models or deep learning models can solve the problem of scene dependence, due to the different calorie consumption characteristics in different exercise stages, it is usually not accurate enough to use a single deep learning model to estimate the calorie consumption in the entire stage.

In view of this, the present disclosure provides a caloric consumption estimation method, device and storage medium to improve the accuracy of estimating the user's caloric consumption during exercise. Embodiments of the present disclosure will be described in detail below.

It should be noted that, in one embodiment, the caloric consumption estimation method provided by the present disclosure can be applied to wearable devices, such as smart watches, smart bracelets and other devices. When the user wears the wearable device, the wearable device can calculate the energy consumption in real time. Collect the user's heart rate and calculate the user's calorie consumption during exercise. In another embodiment, the caloric consumption estimation method provided by the present disclosure can be applied to a terminal that establishes a connection relationship with a wearable device, such as a user's mobile phone, tablet computer, and other devices. The wearable device can transmit the collected user heart rate to the device in real time. The terminal can obtain the user's heart rate and calculate the user's calorie consumption during exercise.

FIG. 1 is a flow chart of a method for estimating heat consumption according to an exemplary embodiment. As shown in FIG. 1 , the method may include S101 to S104.

In S101, obtain the heart rate and respiratory rate of the target user at the current moment.

Among them, the target user is a user wearing a wearable device. The wearable device can collect the target user's heart rate according to the preset sampling frequency. For example, the preset sampling frequency is 1HZ, then the wearable device can collect the target user's heart rate every 1 second. Heart rate.

In addition to sports scenes such as running and mountain climbing that consume calories, low-intensity sports activities such as daily office work also consume calories. However, during low-intensity exercise, the pattern of calorie consumption is different from that during high-intensity exercise, showing a stronger nonlinear relationship. In order to more accurately estimate the calorie consumption under low-intensity exercise, the respiratory frequency of the target user is also obtained in this disclosure. The respiratory frequency represents the number of breaths per minute, which can reflect the user's exercise intensity to a certain extent. The respiratory frequency can be obtained by Detected by sensors in wearable devices.

In S102, based on the heart rate at the current moment, the motion state characteristic information of the target user at the current moment is determined. This motion state feature information is used to characterize whether the target user is in the motion steady state phase or the motion recovery phase at the current moment.

Among them, users usually go through an exercise startup phase, an exercise plateau phase, and an exercise recovery phase when exercising. The start-up stage of exercise is the stage when exercise is just beginning. This stage is short and consumes less calories. The stationary stage of exercise is the stage where breathing and heart rate fluctuations are small and relatively stable, such as the running stage at a constant speed. Since the start-up stage of exercise takes a long time, It is short and consumes less calories, so the exercise startup phase and the exercise plateau phase are regarded as the exercise steady state phase in this disclosure. In scenes such as outdoor long-distance running, cross-country running, mountain climbing, etc., athletes will frequently run at different speeds and take short breaks to recover their physical strength. This stage when the speed is significantly reduced or exercise is paused to restore physical strength, can be regarded as the exercise recovery stage.

Since the user's heart rate is different in different exercise stages, and the heart rate changes with the change of exercise stage, the characteristic information of the target user's movement status at the current moment can be determined based on the heart rate at the current moment.

In S103, the calorie consumption increment of the target user in the current period is obtained based on the calorie estimation model corresponding to the current heart rate, the current respiratory rate, and the current motion state feature information.

Among them, the heat estimation models corresponding to different motion state characteristic information are different, and the current period is the period from the previous moment to the current moment. For example, take thirty moments t1, t2, t3,..., t28, t29, t30 from front to back as an example. For example, the current moment is t30, and the current period is from t29 to t30. period, the increment of caloric consumption in the current period is the caloric consumption of the target user from time t29 to time t30.

In this disclosure, different motion stages are modeled separately. The heat estimation models corresponding to different motion state feature information are different, and the heat estimation models corresponding to different motion state feature information can all be deep learning network models. This disclosure takes into account that the characteristics of caloric consumption are different in the steady-state phase of exercise and the recovery phase of exercise. Therefore, considering what phase of exercise the target user is in, different models are used for caloric estimation in different exercise phases, which can make the estimation of caloric consumption more accurate. .

Moreover, inputting the target user's breathing frequency into the corresponding heat estimation model at the same time can increase the model's recognition of sports scenes and exercise intensity, and achieve accurate calorie consumption in both sports scenes and non-sports scenes (such as daily office work). Estimate.

In S104, the total calorie consumption information of the target user during exercise is determined based on the calorie consumption increment.

The movement process includes the process from the start time of the movement to the current time. For example, when preparing to exercise, the target user can click the start exercise button on the wearable device or terminal, and can also select the type of exercise he or she is doing, such as jogging, mountain climbing, cross-country running, etc., and the target user triggers the start exercise button. The moment of pressing the button can be used as the start moment of exercise. In addition, the moment when the user wears the wearable device can also be used as the start time of exercise. Even if the user does not click the start exercise button, the user's caloric consumption can be estimated regardless of whether the user is in exercise or daily office status. For example, the calorie consumption increment and total calorie consumption information may be represented by calories.

In this disclosure, S103 may include:

If the target user is in the steady-state stage of exercise at the current moment, the target user's calorie consumption increment in the current period is obtained based on the heart rate at the current moment, the respiratory rate at the current moment and the steady-state stage calorie estimation model;

If the target user is in the exercise recovery stage at the current moment, the target user's calorie consumption increment in the current period is obtained based on the heart rate at the current moment, the respiratory rate at the current moment, and the recovery stage calorie estimation model.

Among them, in this disclosure, the steady-state stage of exercise and the recovery stage of exercise are modeled separately. The steady-state stage caloric estimation model is used to estimate the user's caloric consumption increment when the user is in the steady-state stage of exercise. The recovery stage caloric estimation model is used to estimate the user's caloric consumption increment. The increase in caloric consumption during the exercise recovery phase, where the caloric estimation model in the steady-state phase is different from the caloric estimation model in the recovery phase, and both the caloric estimation model in the steady-state phase and the caloric estimation model in the recovery phase can be deep learning network models. This disclosure takes into account that the characteristics of caloric consumption are different in the steady-state phase of exercise and the recovery phase of exercise. Therefore, considering what phase of exercise the target user is in, different models are used for caloric estimation in different exercise phases, which can make the estimation of caloric consumption more accurate. .

Figure 2 is a flow chart of a method for estimating heat consumption according to an exemplary embodiment. As shown in Figure 2, the method includes S201 to S206.

In S201, obtain the heart rate and respiratory rate of the target user at the current moment. For the implementation of step S201, reference may be made to S101.

In S202, based on the heart rate at the current moment, the motion state characteristic information of the target user at the current moment is determined. For the implementation of step S202, reference may be made to S102.

In S203, the physiological characteristic information of the target user is obtained.

For example, the target user's physiological characteristic information includes, for example, the target user's weight, age, gender and other information. It should be noted that the target user can fill in the physiological characteristic information such as weight and age on the wearable device or terminal. With the user's authorization, the wearable device or terminal can obtain the physiological characteristic information of the target user.

It should be noted that there is no restriction on the execution order of S203. S203 may also be executed before S201 or S202. The execution order shown in Figure 2 is only an example.

In S204, if the target user is in the steady-state stage of exercise at the current moment, the current heart rate, the current respiratory rate, and the target user's physiological characteristic information are input into the steady-state stage heat estimation model to obtain the steady-state stage heat Estimate the incremental caloric consumption output from the model.

In S205, if the target user is in the exercise recovery stage at the current moment, the heart rate at the current moment, the respiratory rate at the current moment, and the target user's physiological characteristic information are input into the recovery stage heat estimation model to obtain the recovery stage heat estimation model output. increase in caloric consumption.

In the related art, individual differences are not considered when estimating calorie consumption. Different users may consume different calories even if they perform the same exercise. There are many factors that affect calorie consumption when the human body is exercising. For example, to complete the same exercise, a heavier user needs to do more external work to complete it. Weight has a great impact on calorie consumption. In addition, age will affect The body's metabolic rate will also affect the caloric consumption during exercise. Therefore, in this disclosure, when estimating caloric consumption, the physiological characteristic information of the target user is input into the corresponding caloric estimation model as one of the features, that is, individual differences are taken into account, which can reduce the individual variability of caloric consumption estimation, so that the caloric Consumption estimates are more accurate.

In S206, the total calorie consumption information of the target user during exercise is determined based on the calorie consumption increment. For the implementation of step S206, reference may be made to S104.

Through the above technical solution, when estimating calorie consumption, the physiological characteristic information and respiratory frequency of the target user are used as the input of the corresponding calorie estimation model, which can not only reduce the individual differences in calorie consumption estimation, but also increase the model's accuracy in sports scenes, The identification of exercise intensity makes the estimation of calorie consumption more accurate.

In an embodiment, the implementation of S102 may be:

Input the heart rate at the current moment into the motion state recognition model to obtain the motion state feature information output by the motion state recognition model.

The motion state recognition model can be a classifier, such as a Bayesian classifier. FIG. 3 is a schematic flowchart of a heat consumption estimation process according to an exemplary embodiment. As shown in Figure 3, the heart rate is first input into the exercise state recognition model. If the exercise state recognition model identifies that the target user is currently in the steady state stage of exercise, the heart rate is input into the steady state stage caloric estimation model to obtain the steady state stage. The calorie consumption increment output by the calorie estimation model. If the exercise state recognition model identifies that the target user is currently in the exercise recovery stage, the heart rate is input into the recovery stage calorie estimation model to obtain the calorie consumption increment output by the recovery stage heat estimation model.

In this way, by pre-training a motion state recognition model, the motion state recognition model can accurately identify the current motion stage of the target user.

Figure 4 is a schematic diagram of model training according to an exemplary embodiment. The training of the heat estimation model in the steady state phase, the heat estimation model in the recovery phase, and the exercise state recognition model in the present disclosure will be introduced below with reference to Figure 4 .

In one embodiment, the steady-state stage heat estimation model can be trained in the following manner:

Obtain the heart rate, respiratory rate, actual caloric consumption and physiological characteristics information of the user in the steady state stage of exercise;

The user's heart rate, respiratory rate, actual calorie consumption and physiological characteristic information in the steady-state stage of exercise are input into the deep learning network model, and the deep learning network model is trained to obtain a heat estimation model in the steady-state stage.

In one embodiment, the recovery phase caloric estimation model can be trained in the following manner:

Obtain the heart rate, respiratory rate, actual calorie consumption and physiological characteristics information of the user in the exercise recovery stage;

The heart rate, respiratory rate, actual calorie consumption and physiological characteristic information of the user in the exercise recovery stage are input into the deep learning network model, and the deep learning network model is trained to obtain a calorie estimation model in the recovery stage.

As shown in Figure 4, exercise data may include steady-state phase exercise data and recovery phase exercise data. The steady-state phase exercise data includes the heart rate, respiratory rate, actual caloric consumption and physiological characteristic information of the user in the steady-state phase of exercise. The recovery phase Exercise data includes heart rate, respiratory rate, actual calorie consumption and physiological characteristic information of users in the exercise recovery stage. The actual caloric consumption can be calculated through gas exchange through professional cardiopulmonary testing systems. The steady-state phase exercise data is used to train the steady-state phase heat estimation model, and the recovery phase exercise data is used to train the recovery phase heat estimation model.

In an embodiment, the motion state recognition model can be trained in the following manner:

The heart rate when the user is in the steady-state phase of exercise and the label used to characterize the heart rate corresponding to the steady-state phase of exercise, and the heart rate when the user is in the exercise recovery phase and the label used to characterize the heart rate corresponding to the exercise recovery phase are used as a classifier As input, the classifier is trained to obtain the motion state recognition model.

As shown in Figure 4, the heart rate of the user when he is in the exercise recovery stage is obtained from the recovery stage exercise data, and the label of the heart rate is labeled label 1. Label 1 is used to indicate that the heart rate corresponds to the exercise recovery stage. Obtain the heart rate of the user when he is in the steady-state stage of exercise from the steady-state stage exercise data, and label the heart rate label as label 2. Label 2 is used to indicate that the heart rate corresponds to the steady-state stage of exercise. The classifier can be a Bayesian classifier, for example. By training the classifier, a motion state recognition model can be obtained.

In this disclosure, the steady-state stage heat estimation model and the recovery stage heat estimation model may both be gated recurrent unit models (GRU, gated recurrent unit). The GRU model has great advantages in dealing with time series problems. It introduces cell state to save long-term memory and improves the defect of RNN (Recurrent Neural Network) that can only save short-term memory. At the same time, the GRU model improves the problem of high computational complexity of the LSTM (Long Short-Term Memory, Long Short-Term Memory Network) deep learning algorithm, making the prediction delay as small as possible. Using the GRU model to build a mapping model between heart rate and caloric consumption can improve the estimation accuracy compared to other machine learning, and is better able to handle the problem of switching between different sports scenes.

It is worth mentioning that the methods of training the heat estimation model in the steady state phase and the training heat estimation model in the recovery phase are similar. The training parameters used to train different models, such as the number of iterations, loss functions and other parameters, can be the same or different, according to actual needs. Settings are made. The following training parameters are only examples and do not constitute a limitation on the implementation of the present disclosure. First initialize the GRU model, and then train the GRU model. For example, the number of iterations is M=256, the loss function loss=mae (mean absolute error), the optimizer selects the Adam optimizer, and the number of samples selected for one training is batch size =16.

Figure 5 is a schematic diagram of the internal structure of the gated cyclic unit model. As shown in Figure 5, two gates are used in the GRU model: update gate and reset gate. The input variables only need to undergo vector calculation twice. Among them, zt represents the update gate, which is used to control the degree to which the state information from the previous moment enters the current state. The larger the value of the update gate, the more state information from the previous moment will be introduced. rt represents the reset gate, and the control will come from the previous moment. The information of one state is written into the current temporary cell state. The smaller the reset gate is, the less information is written from the previous state. The GRU model can also realize the function of storing long and short-term memory. Its prediction process is as shown in the following formulas (1) to (5):

r _t =σ(W _r ·[h _t-1 ,x _t ]) (1)

z _t =σ(W _z ·[h _t-1 ,x _t ]) (2)

y _t =σ(W _o ·h _t ) (5)

Among them, σ is the sigmoid layer, including the sigmoid function, and its expression is as shown in formula (6):

The tanh layer contains the tanh function, whose expression is shown in formula (7):

e is the Hadamard Product, which means the corresponding elements in the operation matrix are multiplied, so the two multiplication matrices are required to be of the same type. Wr is the weight matrix of the reset gate, Wz is the weight matrix of the update gate,

is the weight matrix of the candidate state, and Wo is the weight matrix of the output. ht and ht-1 respectively represent the output of the cell state storing long-term memory at time t and t-1.

Represents the output of the candidate state at time t.

Initialize the parameters of the LSTM deep learning network: reset the weight matrix Wr of the gate, update the weight matrix Wz of the gate, and the weight matrix of the candidate state.

The output weight matrix Wo is initialized to a random number between 0 and 1. Set the number of neurons in the input layer of the GRU deep learning network to M, the number of layers to i, and the output of each layer is used as the input of the next layer. After the GRU model is initialized, the GRU model is trained. Taking the training of the steady-state stage heat estimation model as an example, the user's heart rate, respiratory rate, actual heat consumption and physiological characteristic information in the steady-state stage of exercise are input into the deep learning network model. , the back propagation algorithm can be used to optimize the training of the matrix in the deep learning network model, and when the training is completed, the heat estimation model in the steady state phase is obtained. Among them, the number of training iterations is set to epoch, the loss function is loss, the optimizer is optimizer, and the number of samples selected for one training is batch size. In this embodiment, M=256, i=2, epoch=50, loss=mae (mean absolute error), optimizer=Adam (Adam optimizer), and batch size=16. Taking the caloric estimation model in the training recovery phase as an example, the heart rate, respiratory rate, actual caloric consumption and physiological characteristic information of the user in the exercise recovery phase are input into the deep learning network model. The back propagation algorithm can be used to The matrix is optimized and trained, and the recovery phase heat estimation model is obtained when the training is completed. Among them, the number of training iterations is set to epoch, the loss function is loss, the optimizer is optimizer, and the number of samples selected for one training is batch size. In this embodiment, M=256, i=2, epoch=50, loss=mae (mean absolute error), optimizer=Adam (Adam optimizer), and batch size=16.

It can be seen from the above formula that the GRU model needs to learn the reset gate rt, update gate zt and temporary state during training.

There are three sets of parameters. The formula for the total number of parameters of the GRU model is: total parameters = 3 * hidden layer parameters * (input layer parameters + bias parameters + output layer parameters). During training, LSTM needs to learn the input gate it, the output gate ot, the forget gate ft and the temporary state.

Four sets of parameters, the formula for the total number of parameters of LSTM is: total parameters = 4 * hidden layer parameters * (input layer parameters + bias parameters + output layer parameters). GRU model training requires fewer iteration parameters than LSTM. Therefore, GRU model training is faster than LSTM and the number of convergence iterations is fewer than LSTM. The electromyographic features input by LSTM during prediction require seven complex vector calculations to obtain the final estimated angle, while the GRU model only requires five vector calculations to obtain the final estimated angle during online prediction. The GRU model predicts online The timeliness is also better than the LSTM model. Therefore, the timeliness of the GRU model is better than that of the LSTM model in both the training phase and the prediction phase. Therefore, the heat estimation model in the steady state phase and the heat estimation model in the recovery phase in this disclosure can both be the gated cycle unit model, that is, the GRU model.

Through the above technical solution, for complex sports states such as mountain climbing and cross-country running where the steady stage of movement and the movement recovery stage (multiple rests) are frequently switched, the present disclosure can use different models for different movement stages to calculate calorie consumption. Estimation is more effective than using a single model to estimate calories for the entire stage, and can effectively improve the accuracy of energy consumption estimation.

The caloric consumption estimation method provided by this disclosure may also include:

After the target user finishes exercising, the current basal metabolic information of the target user is determined based on the target user's calorie consumption during this exercise, and the current basal metabolic information is displayed. The current basal metabolic information is greater than the target user's basal metabolism before this exercise. information.

Among them, basal metabolism refers to the energy metabolism under the basic state of the human body. Exercise can improve the basal metabolism of the human body. The greater the intensity of exercise, the higher the calories consumed, and the higher the level of improvement in basal metabolism. After the target user finishes exercising, Determine the target user's current basal metabolic information based on the target user's calorie consumption during this exercise, and display the current basal metabolic information on a wearable device or terminal. The method of calculating the basal metabolic information can refer to related technologies, for example, based on the target user's current basal metabolic information. Determine a coefficient for the caloric consumption of each exercise, and multiply the original basal metabolic information by this coefficient to obtain the current basal metabolic information. Since exercise can improve the basal metabolic level of the human body, the current basal metabolic information of the target user after the end of this exercise Greater than the target user’s basal metabolic information before this exercise.

Based on the same inventive concept, the present disclosure also provides a device for estimating heat consumption. Figure 6 is a block diagram of a device for estimating heat consumption according to an exemplary embodiment. As shown in Figure 6, the device 600 may include:

The frequency acquisition module 601 is used to obtain the heart rate and respiratory frequency of the target user at the current moment; the state determination module 602 is used to determine the motion state characteristic information of the target user at the current moment based on the heart rate at the current moment, Wherein, the motion state characteristic information is used to characterize whether the target user is in a steady state phase or a recovery phase of motion at the current moment; the increment determination module 603 is used to determine whether the target user is in a steady state or recovery phase based on the heart rate at the current moment, the current The respiratory frequency at the moment and the caloric estimation model corresponding to the motion state characteristic information at the current moment are used to obtain the caloric consumption increment of the target user in the current period. The caloric estimation models corresponding to different motion state characteristic information are different, so The current period is the period from the previous moment to the current moment; the total calorie determination module 604 is used to determine the total calorie consumption of the target user during exercise according to the calorie consumption increment. Calorie consumption information, wherein the exercise process includes the process from the start time of the exercise to the current time.

Optionally, the increment determination module 603 includes: a first increment determination sub-module, configured to determine if the target user is in the steady state stage of exercise at the current moment according to the heart rate at the current moment, the The respiratory frequency and steady-state stage heat estimation model at the current moment are used to obtain the heat consumption increment of the target user in the current period; the second increment determination sub-module is used to determine if the target user is in motion at the current moment. In the recovery phase, the calorie consumption increment of the target user in the current period is obtained based on the heart rate at the current moment, the respiratory rate at the current moment and the recovery phase calorie estimation model.

Optionally, the device 600 further includes: an information acquisition module, configured to acquire the physiological characteristic information of the target user; and the first increment determination sub-module is configured to: obtain the heart rate at the current moment, the current The respiratory frequency at the moment and the physiological characteristic information of the target user are input into the steady-state stage heat estimation model, and the heat consumption increment output by the steady-state stage heat estimation model is obtained; the second increment is determined The sub-module is configured to: input the heart rate at the current moment, the respiratory rate at the current moment, and the physiological characteristic information of the target user into the recovery stage heat estimation model, and obtain the heat estimation model output by the recovery stage. The heat consumption increases.

Optionally, the state determination module 602 is configured to input the heart rate at the current moment into a motion state recognition model to obtain the motion state feature information output by the motion state recognition model.

Optionally, the device 600 further includes: a basal metabolism determination module, configured to determine the current basal metabolism information of the target user based on the calorie consumption of the target user's current exercise after the target user finishes exercising, And display the current basal metabolic information, wherein the current basal metabolic information is greater than the target user's basal metabolic information before this exercise.

Regarding the devices in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments related to the method, and will not be described in detail here.

The present disclosure also provides a computer-readable storage medium on which computer program instructions are stored. When the program instructions are executed by a processor, the steps of the heat consumption estimation method provided by the present disclosure are implemented.

FIG. 7 is a block diagram of a device 800 for caloric consumption estimation according to an exemplary embodiment. For example, the device 800 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, or the like.

7, the device 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814, and communications component 816.

Processing component 802 generally controls the overall operations of device 800, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above-mentioned caloric consumption estimation method. Additionally, processing component 802 may include one or more modules that facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802.

Memory 804 is configured to store various types of data to support operations at device 800 . Examples of such data include instructions for any application or method operating on device 800, contact data, phonebook data, messages, pictures, videos, etc. Memory 804 may be implemented by any type of volatile or non-volatile storage device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EEPROM), Programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

Power component 806 provides power to various components of device 800. Power components 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to device 800 .

Multimedia component 808 includes a screen that provides an output interface between the device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide action. In some embodiments, multimedia component 808 includes a front-facing camera and/or a rear-facing camera. When the device 800 is in an operating mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front-facing camera and rear-facing camera can be a fixed optical lens system or have a focal length and optical zoom capabilities.

Audio component 810 is configured to output and/or input audio signals. For example, audio component 810 includes a microphone (MIC) configured to receive external audio signals when device 800 is in operating modes, such as call mode, recording mode, and speech recognition mode. The received audio signal may be further stored in memory 804 or sent via communication component 816 . In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

The I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module, which may be a keyboard, a click wheel, a button, etc. These buttons may include, but are not limited to: Home button, Volume buttons, Start button, and Lock button.

Sensor component 814 includes one or more sensors that provide various aspects of status assessment for device 800 . For example, the sensor component 814 can detect the open/closed state of the device 800, the relative positioning of components, such as the display and keypad of the device 800, and the sensor component 814 can also detect a change in position of the device 800 or a component of the device 800. , the presence or absence of user contact with the device 800 , device 800 orientation or acceleration/deceleration and temperature changes of the device 800 . Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

Communication component 816 is configured to facilitate wired or wireless communication between apparatus 800 and other devices. Device 800 may access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In one exemplary embodiment, communication component 816 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communications component 816 also includes a near field communications (NFC) module to facilitate short-range communications. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

In an exemplary embodiment, apparatus 800 may be configured by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable Gate array (FPGA), controller, microcontroller, microprocessor or other electronic components are implemented for performing the above heat consumption estimation method.

In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions, such as a memory 804 including instructions, which can be executed by the processor 820 of the device 800 to complete the above-described heat consumption estimation method is also provided. For example, the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

In another exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program executable by a programmable device, the computer program having a function for performing the above when executed by the programmable device. The code part of the calorie consumption estimation method.

Other embodiments of the disclosure will be readily apparent to those skilled in the art from consideration of the specification and practice of the disclosure. This application is intended to cover any variations, uses, or adaptations of the disclosure that follow the general principles of the disclosure and include common knowledge or customary technical means in the technical field that are not disclosed in the disclosure. . It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It is to be understood that the present disclosure is not limited to the precise structures described above and illustrated in the accompanying drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the disclosure is limited only by the appended claims.

Claims

A method for estimating calorie consumption, which is characterized by including:

Obtain the heart rate and respiratory rate of the target user at the current moment;

According to the heart rate at the current moment, the motion state characteristic information of the target user at the current moment is determined, wherein the motion state characteristic information is used to represent that the target user is in a steady state stage of motion at the current moment. Still in the exercise recovery stage;

According to the heat estimation model corresponding to the heart rate at the current moment, the respiratory rate at the current moment, and the motion state feature information at the current moment, the calorie consumption increment of the target user in the current period is obtained, wherein different movements The heat estimation models corresponding to the state characteristic information are different, and the current period is the period from the previous moment of the current moment to the current moment;

According to the calorie consumption increment, the total calorie consumption information of the target user during exercise is determined, wherein the exercise process includes the process from the start time of exercise to the current time.
The method of claim 1, wherein the target is obtained based on the heat estimation model corresponding to the heart rate at the current moment, the respiratory frequency at the current moment, and the motion state feature information at the current moment. The user's incremental calorie consumption during the current period, including:

If the target user is in the steady-state stage of exercise at the current moment, then based on the heart rate at the current moment, the respiratory rate at the current moment and the steady-state stage caloric estimation model, the target user's energy in the current period is obtained. Increased caloric consumption;

If the target user is in the exercise recovery stage at the current moment, then based on the heart rate at the current moment, the respiratory rate at the current moment and the recovery stage caloric estimation model, the target user's fitness in the current period is obtained. Increasing caloric consumption.
The method of claim 2, further comprising:

Obtain physiological characteristic information of the target user;

Obtaining the incremental calorie consumption of the target user in the current period based on the heart rate at the current moment, the respiratory rate at the current moment and the steady-state stage calorie estimation model includes:

Input the heart rate at the current moment, the respiratory rate at the current moment, and the physiological characteristic information of the target user into the steady-state stage heat estimation model to obtain the heat output from the steady-state stage heat estimation model. consumption increase;

Obtaining the incremental calorie consumption of the target user in the current period based on the heart rate at the current moment, the respiratory rate at the current moment and the recovery phase calorie estimation model includes:

The heart rate at the current moment, the respiratory rate at the current moment, and the physiological characteristic information of the target user are input into the recovery phase heat estimation model, and the heat consumption increase output by the recovery phase heat estimation model is obtained. quantity.
The method according to claim 2, characterized in that the steady-state stage heat estimation model is trained in the following manner:

Obtain the heart rate, respiratory rate, actual caloric consumption and physiological characteristics information of the user in the steady state stage of exercise;

The user's heart rate, respiratory rate, actual calorie consumption and physiological characteristic information in the steady-state stage of exercise are input into the deep learning network model, and the deep learning network model is trained to obtain the steady-state stage heat estimation model.
The method according to claim 2, characterized in that the recovery phase heat estimation model is trained in the following manner:

Obtain the heart rate, respiratory rate, actual calorie consumption and physiological characteristics information of the user in the exercise recovery stage;

The user's heart rate, respiratory rate, actual calorie consumption and physiological characteristic information in the exercise recovery stage are input into the deep learning network model, and the deep learning network model is trained to obtain the recovery stage calorie estimation model.
The method according to claim 1, wherein determining the motion state characteristic information of the target user at the current moment based on the heart rate at the current moment includes:

Input the heart rate at the current moment into the motion state recognition model to obtain the motion state feature information output by the motion state recognition model.
The method according to claim 6, characterized in that the motion state recognition model is trained in the following manner:

The heart rate when the user is in the steady-state phase of exercise and the label used to characterize the heart rate corresponding to the steady-state phase of exercise, and the heart rate when the user is in the exercise recovery phase and the label used to characterize the heart rate corresponding to the exercise recovery phase are used as a classifier As input, the classifier is trained to obtain the motion state recognition model.
The method according to any one of claims 1 to 7, characterized in that both the steady-state stage heat estimation model and the recovery stage heat estimation model are gated cycle unit models.
The method of claim 1, further comprising:

After the target user finishes exercising, the current basal metabolic information of the target user is determined based on the calorie consumption of the current exercise, and the current basal metabolic information is displayed, wherein the current basal metabolic information is greater than The target user's basal metabolic information before this exercise.
A heat consumption estimating device, characterized by including:

The frequency acquisition module is used to obtain the heart rate and respiratory rate of the target user at the current moment;

A state determination module, configured to determine the motion state characteristic information of the target user at the current moment according to the heart rate at the current moment, wherein the motion state characteristic information is used to characterize the motion state characteristic information of the target user at the current moment. Is it in the steady-state phase of exercise or the recovery phase of exercise?

Increment determination module, configured to obtain the calorie consumption of the target user in the current period based on the heart rate at the current moment, the respiratory rate at the current moment, and the calorie estimation model corresponding to the motion state feature information at the current moment. Increment, wherein the heat estimation models corresponding to different motion state characteristic information are different, and the current period is the period from the previous moment of the current moment to the current moment;

A total calorie determination module, configured to determine the total calorie consumption information of the target user during exercise according to the calorie consumption increment, wherein the exercise process includes the process from the start time of exercise to the current time.
A heat consumption estimating device, characterized by including:

processor;

Memory used to store instructions executable by the processor;

Wherein, the processor is configured to perform the steps of the method according to any one of claims 1 to 9.
A computer-readable storage medium on which computer program instructions are stored, characterized in that when the program instructions are executed by a processor, the steps of the method described in any one of claims 1 to 9 are implemented.