wherein, Yi is neural network's output, and Yi is neural network's expectation output, and the expectation output is the standard value of rhythm of the heart promptly, and the rhythm of the heart of user at corresponding time point is detected to the more accurate rhythm of the heart detection device of accessible, obtains this standard value, and more accurate rhythm of the heart detection device can include the rhythm of the heart area. The variables such as the set threshold, the number of cycles N, the learning rate, etc. can be determined by debugging after accumulating a large amount of test data.

Optionally, when training the neural network model corresponding to each motion state, a plurality of historical motion data acquired by the motion sensor and heart rate data acquired by the heart rate sensor may be preprocessed, and the neural network model may be trained according to the preprocessed data, where the preprocessing includes at least one of filtering, segmentation, and normalization.

In the technical scheme disclosed in this embodiment, the heart rate value under the different motion states of user is obtained through different neural network models, reduces the influence of the different motion states of user to heart rate detection, improves the accuracy of heart rate detection, and compare in traditional mode of revising the heart rate value according to certain error, the neural network model can be according to the corresponding heart rate value of arbitrary heart rate data output, the heart rate value that obtains is more accurate.

In another embodiment, as shown in fig. 5, on the basis of the embodiment shown in any one of fig. 2 to 4, before step S30, the method further includes:

step S50, preprocessing the heart rate data, where the preset processing includes at least one of filtering processing, segmentation processing, and normalization processing;

step S30 includes:

and step S33, processing the preprocessed heart rate data according to the motion state to obtain the heart rate value.

In this embodiment, before being handled heart rate data according to the motion state, still can carry out preliminary treatment to heart rate data to optimize heart rate data, the heart rate value that obtains according to the heart rate data after the optimization is more accurate.

Optionally, in the preprocessing, the heart rate data is first subjected to a filtering process, and the filtering process may be implemented by a band-pass filter. Because the heart rate data is the continuous signal of the collection in a period of time usually, therefore, after the filtering process, the heart rate data after the filtering process is segmented and cut into the heart rate data of a plurality of segments of different time periods, so as to process the heart rate data in different time periods separately. After segmentation, normalization processing is carried out on the segmented and cut heart rate data so as to simplify the processing process. And after the normalization processing, processing the heart rate data after the normalization processing according to the motion state to obtain a heart rate value.

Optionally, the motion state corresponding to the motion data is obtained, the motion data may be preprocessed, and the corresponding motion state is obtained according to the preprocessed motion data.

In the technical scheme disclosed in this embodiment, the heart rate data is optimized, and the heart rate value is calculated according to the optimized heart rate data, so that the calculation process is simplified, and the accuracy of the calculated heart rate value is improved.

In addition, an embodiment of the present invention further provides a wearable device, where the wearable device includes: the heart rate detection method comprises a memory, a processor and a heart rate detection program stored on the memory and capable of running on the processor, wherein the steps of the heart rate detection method are realized when the heart rate detection program is executed by the processor.

In addition, an embodiment of the present invention further provides a computer storage medium, where a heart rate detection program is stored on the computer storage medium, and when the heart rate detection program is executed by a processor, the steps of the heart rate detection method according to the above embodiments are implemented.

In another embodiment, referring to fig. 6 and 7, based on the embodiment shown in any one of fig. 2 to 5, fig. 6 is a schematic diagram of a component module of a heart rate detection device, and fig. 7 is a schematic flow chart of a heart rate detection method. The specific technical scheme of the embodiment is as follows:

a training stage: 1. the heart rate sensor, the acceleration sensor and the geomagnetic sensor are used for respectively acquiring human body signals of different motion states for a period of time, and the electrocardiogram recorder is used for recording the heart rates of the different motion states as labels. The signal is filtered using a band pass filter, segmented cut and normalized.

2. And selecting data signals of the acceleration sensor and the geomagnetic sensor, performing data dimensionality reduction and feature selection by using an LDA linear discriminant analysis method, and selecting the front k features as the input of the nearest neighbor KNN classifier. And training a nearest neighbor KNN classification model by using the k characteristics and the corresponding different motion states, and judging whether the input signal corresponds to the output state by adopting a classification rule of majority voting.

3. Selecting a heart rate sensor to collect signals of different motion states, respectively training different neural network models according to the different motion states, wherein an activation function is a sigmoid function, and 30% Drop out is added to a full connection layer in order to prevent overfitting and adding regularization operation to a convolutional layer. Selecting a mean square error loss function and an Adam optimization algorithm, setting the cycle number as N epochs, setting the learning rate as l, inputting PPG signals into the network, and outputting the PPG signals as heart rate values at corresponding moments. And stopping training when the training error is smaller than a set threshold value, and obtaining parameters of the neural network model for predicting the exercise heart rate.

Heart rate detection stage: the method comprises the steps of collecting human body signals of different motion states for a period of time by using a heart rate sensor, an acceleration sensor and a geomagnetic sensor respectively, filtering the signals by using a band-pass filter, and carrying out segmentation cutting and normalization. And (3) performing data dimensionality reduction and feature selection by using an LDA (linear discriminant analysis) method, inputting the features subjected to dimensionality reduction into a nearest neighbor KNN (K nearest neighbor) classifier, and classifying signals of the acceleration sensor and signals of the geomagnetic sensor by using the trained nearest neighbor KNN classifier to obtain different motion states. And selecting a corresponding heart rate prediction neural network model according to the classified motion state, and outputting a heart rate value.

In the technical scheme disclosed in the embodiment, different motion states are accurately and quickly classified by fusing multi-sensor data, a neural network model is trained aiming at the different motion states, and the heart rate is predicted by using the different neural network models. Therefore, heart rate prediction errors caused by the change of the motion state are effectively reduced, and the accuracy and the real-time performance of heart rate detection are improved. The algorithm related to the technical scheme has accuracy, self-adaptability and real-time performance, and the heart rate prediction method can obtain accurate heart rate values in both static and dynamic scenes, and is particularly suitable for heart rate prediction in physical exercise.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.

The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.

Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) as described above and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.

The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.

Claims

1. A heart rate detection method is characterized by comprising the following steps:

acquiring a motion state corresponding to the motion data;

outputting the heart rate value.

2. The method for detecting heart rate according to claim 1, wherein the step of obtaining the motion state corresponding to the motion data includes:

3. The method for detecting heart rate according to claim 1, wherein the step of obtaining the motion state corresponding to the motion data includes:

and acquiring a motion state according to the target characteristic data.

4. The method for detecting heart rate according to claim 1, wherein the step of processing the heart rate data collected by the heart rate sensor in the wearable device according to the motion state to obtain the heart rate value comprises:

and acquiring the heart rate value output by the neural network model.

5. The method for detecting heart rate according to claim 4, wherein the method for detecting heart rate further comprises:

acquiring a plurality of historical motion data acquired by the motion sensor;

determining a motion state corresponding to each historical motion data;

6. The method for detecting heart rate according to claim 1, wherein the step of processing the heart rate data collected by the heart rate sensor in the wearable device according to the motion state to obtain the heart rate value is preceded by the step of:

7. The method for detecting heart rate according to claim 1, wherein the step of processing the heart rate data collected by the heart rate sensor in the wearable device according to the motion state to obtain the heart rate value is preceded by the step of:

acquiring a user type corresponding to the wearable device;

8. A wearable device, characterized in that the wearable device comprises: memory, processor and a heart rate detection program stored on the memory and executable on the processor, the heart rate detection program when executed by the processor implementing the steps of the heart rate detection method according to any one of claims 1 to 7.

9. A computer storage medium, characterized in that the computer storage medium has stored thereon a heart rate detection program, which when executed by a processor implements the steps of the heart rate detection method according to any one of claims 1 to 7.