Information identification method, device and equipment

Abstract

The invention provides an information identification method, device and equipment. The method comprises the following steps: acquiring first motion data acquired by an inertial sensor of single electronic equipment for a first object; converting the first motion data into a first time-frequency spectrogram; and according to the first time-frequency spectrogram and a pre-trained neural network model, identifying the first object to determine characteristics of the first object.

Description

Technical Field

[0001] The present application relates to electronic technology, and relates to but is not limited to information identification methods, devices and equipment. Background Art

[0002] The field of physiological identity recognition and emotional computing based on inertial sensors is mainly concentrated in the medical field, and generally requires multiple sensors for combined recognition. Sensors are usually placed in multiple locations such as arms, wrists, thighs, waists and ankles. Based on this, the user's identity, physiological characteristics, and emotions are sensed and recognized through a combination of multiple sensors. Summary of the invention

[0003] In a first aspect, an embodiment of the present application provides an information identification method, the method comprising: obtaining first motion data collected by an inertial sensor of a single electronic device for a first object; converting the first motion data into a first time-frequency spectrum graph; and identifying the first object based on the first time-frequency spectrum graph and a pre-trained neural network model to determine characteristics of the first object.

[0004] In a second aspect, an embodiment of the present application provides an information identification device, comprising: an acquisition module, configured to acquire first motion data collected by an inertial sensor of a single electronic device for a first object; a conversion module, configured to convert the first motion data into a first time-frequency spectrum graph; and an identification module, configured to identify the first object based on the first time-frequency spectrum graph and a pre-trained neural network model to determine the characteristics of the first object.

[0005] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processing unit and an inertial sensor; wherein the inertial sensor is used to collect first motion data for a first object; the processing unit is used to obtain the first motion data; and convert the first motion data into a first time-frequency spectrum; and identify the first object based on the first time-frequency spectrum and a pre-trained neural network model to determine the characteristics of the first object.

[0006] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, and when the processor executes the program, the method described in the first aspect is implemented.

[0007] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the method described in the first aspect when executed by a processor.

[0008] In a sixth aspect, an embodiment of the present application provides a computer program product, including a computer program or instructions, which, when executed by a processor, implements the method described in the first aspect of the present application.

[0009] In a seventh aspect, an embodiment of the present application provides a computer program, which enables an electronic device to execute the method described in the first aspect.

[0010] It can be understood that in the embodiment of the present application, the identification of the characteristics of the first object can be achieved by using the inertial sensor of a single electronic device, without the need to link the inertial sensors of multiple electronic devices; therefore, the identification of the characteristics of the first object can be achieved through a single ordinary consumer-grade electronic device (such as a smart phone, a smart watch or a smart bracelet, etc.).

[0011] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and are used together with the specification to illustrate the technical solution of the present application. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0014] Figure 1 A schematic diagram of the implementation flow of the information processing method provided in the embodiment of the present application;

[0015] Figure 2 A schematic diagram of converting motion data provided by an embodiment of the present application into a time-frequency spectrum diagram;

[0016] Figure 3 Schematic diagram of the structure of the pre-trained neural network model provided in the embodiment of the present application Figure 1 ;

[0017] Figure 4 A schematic diagram of a flow chart of a method for determining a characteristic of a first object provided in an embodiment of the present application;

[0018] Figure 5A schematic diagram of the implementation process of the physiological attribute recognition based on walking gait provided in the embodiment of the present application;

[0019] Figure 6 A schematic diagram of a method for determining a final classification prediction result provided in an embodiment of the present application;

[0020] Figure 7 Schematic diagram of the structure of the pre-trained neural network model provided in the embodiment of the present application Figure 2 ;

[0021] Figure 8 A schematic diagram of the structure of an information identification device provided in an embodiment of the present application;

[0022] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the specific technical solution of the present application will be further described in detail below in conjunction with the drawings in the embodiments of the present application. The following embodiments are used to illustrate the present application, but are not used to limit the scope of the present application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0025] In the following description, reference is made to “some embodiments”, “this embodiment”, “embodiments of the present application” and examples, etc., which describe a subset of all possible embodiments, but it can be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0026] The descriptions of “first, second, third” etc. that appear in the embodiments of the present application are only for illustration and distinction of the description objects. There is no distinction of order, nor do they indicate any special limitation on the number of devices in the embodiments of the present application, and cannot constitute any limitation on the embodiments of the present application.

[0027] The field of physiological identity recognition and emotional computing based on inertial sensors is mainly concentrated in the medical field, and generally requires multiple sensors for combined recognition. Sensors are usually placed in multiple locations such as arms, wrists, thighs, waists and ankles. Based on this, the user's identity, physiological characteristics, and emotions are sensed and recognized through a combination of multiple sensors.

[0028] However, the inventor of this application found in research and analysis that it is difficult for ordinary consumers to wear multiple sensors in daily life, and they will not necessarily buy multiple devices of the same brand to form multi-device sensor interconnection. In most cases, consumers will only buy one or two devices of the same brand. Therefore, using a single-point sensor to identify the user's identity, physiological characteristics, and emotional state is a problem that needs to be solved at present.

[0029] Figure 1 A schematic diagram of the implementation flow of the information processing method provided in the embodiment of the present application; Figure 1 As shown, the method may include the following steps 101 to 103:

[0030] Step 101, obtaining first motion data collected by an inertial sensor of a single electronic device for a first object;

[0031] Step 102, converting the first motion data into a first time-frequency spectrum;

[0032] Step 103: Identify the first object according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

[0033] It can be understood that in the embodiment of the present application, the identification of the characteristics of the first object can be achieved by using the inertial sensor of a single electronic device, without the need to link the inertial sensors of multiple electronic devices; therefore, the identification of the characteristics of the first object can be achieved through a single ordinary consumer-grade electronic device (such as a smart phone, a smart watch or a smart bracelet, etc.).

[0034] The following describes further optional implementations and related terms of each of the above steps.

[0035] Step 101: Acquire first motion data collected by an inertial sensor of a single electronic device for a first object.

[0036] In the embodiment of the present application, there is no limitation on the type of the first object, which may be any type of organism (eg, a human or an animal).

[0037] In the embodiment of the present application, the inertial sensor of the single electronic device described in step 101 may be one inertial sensor in the single electronic device, or may be multiple inertial sensors in the single electronic device.

[0038] Exemplarily, in some embodiments, step 101 may include: obtaining first motion data collected by a single inertial sensor in a single electronic device for a first object. In this way, it is beneficial to expand the scope of application of the information processing method provided in the embodiment of the present application, and can be applied to electronic devices with a single inertial sensor or to electronic devices with multiple inertial sensors. Moreover, even if it is applied to an electronic device with multiple inertial sensors, the characteristic recognition of the first object can be realized through the motion data collected by only one inertial sensor. In this way, in the scenario where the characteristic recognition function of the first object is turned on, only one inertial sensor can be turned on, which is beneficial to saving the power consumption of the electronic device while being able to realize the characteristic recognition of the first object.

[0039] It can be understood that inertial sensors are an important sensor technology, which are mainly used to measure the motion changes of electronic devices (i.e., motion data), such as measuring one or more motion data of electronic devices such as acceleration, tilt, impact, vibration, rotation, multi-degree-of-freedom (DoF) motion, etc.

[0040] According to the measurement principle, inertial sensors are divided into accelerometers and gyroscopes. Accelerometers can measure the linear acceleration of electronic devices in three-dimensional space, including gravity acceleration (i.e., an example of first motion data). Gyroscopes can measure the angular velocity of electronic devices around their axes (i.e., an example of first motion data), providing real-time information about the rotation of the device.

[0041] According to the degree of integration, inertial sensors are divided into single-component inertial sensors and inertial measurement units (IMUs); among them, single-component inertial sensors: such as a separate accelerometer or gyroscope, are used to measure a single physical quantity (that is, the change in motion). Inertial measurement unit (IMU): is a device that integrates multiple inertial sensors, usually including accelerometers and gyroscopes, or other sensors such as magnetometers. IMU can simultaneously measure the first motion data such as acceleration, angular velocity and angle of the device, and fuse the data of the two to provide more comprehensive and accurate motion data.

[0042] According to the classification of technology type, inertial sensors include MEMS inertial sensors, fiber optic gyroscopes and laser gyroscopes; among them, MEMS inertial sensors: miniature inertial sensors manufactured using micro-nano technology, with the characteristics of miniaturization, low power consumption, high precision, high reliability and low cost. Fiber optic gyroscope: obtains the angular velocity information of electronic devices by measuring the changes in the light path, with high accuracy and stability. Laser gyroscope: uses laser technology to measure the angular velocity of electronic devices, also with high accuracy and stability.

[0043] In the embodiment of the present application, there is no limitation on the type of the inertial sensor, and the inertial sensor includes one or more types mentioned above.

[0044] In the embodiments of the present application, there is no restriction on the placement position of the single electronic device on the first object, as long as the motion condition / motion data of the first object can be measured. For example, in some embodiments, the electronic device is placed at a specific position of the first object; wherein the specific position includes: a wrist, a thigh, or an ankle.

[0045] It can be understood that the first motion data is the motion data collected by the inertial sensor of the single electronic device at the wrist of the first object, mainly considering that the wrist is the position where the single electronic device is often placed, for example, the user often holds a mobile phone, or the single electronic device is a smart watch, smart bracelet or smart ring worn by the user.

[0046] It can be understood that the first motion data is the motion data collected by the inertial sensor of the single electronic device at the thigh of the first object, mainly considering that the thigh is where the single electronic device is often placed, for example, users often put their mobile phones in their pockets (such as in their trouser pockets or in their shirt pockets near the thighs).

[0047] It can be understood that the first motion data is the motion data collected by the inertial sensor of the single electronic device at the ankle of the first object, mainly considering that the ankle is where the single electronic device is often placed, for example, sometimes the user will stick an inertial sensor on the foot, or the user often wears a smart anklet, or the user often wears smart sports shoes, etc. Among them, one or more inertial sensors are set in the smart anklet and the smart sports shoes.

[0048] As mentioned above, step 101 is to obtain first motion data collected by an inertial sensor of a single electronic device for a first object; and step 101 may further include: obtaining first motion data collected by a single inertial sensor in a single electronic device for the first object.

[0049] In a possible implementation, step 101 may further include: when the single electronic device is placed at a specific position of the first object, obtaining first motion data collected by the inertial sensor of the single electronic device for the first object; wherein the specific position includes: a wrist, a thigh, or an ankle. In this way, more accurate motion data (that is, data that is closer to the actual motion state of the first object) can be collected, which is beneficial to improving the accuracy of identifying the characteristics of the first object.

[0050] Step 102: convert the first motion data into a first time-frequency spectrum diagram.

[0051] It can be understood that the first motion data is data collected in one or more time periods, that is, a data sequence. In an embodiment of the present application, the first motion data may include motion data on a single axis, or may include motion data on multiple different axes. The motion data on multiple different axes may be converted into corresponding first time-frequency spectrograms respectively.

[0052] That is, in some embodiments, the first motion data includes motion data of the first object in N axes, N being greater than 1; step 102 may further include: converting the motion data in the N axes respectively to obtain N first time-frequency spectrograms. That is, converting the motion data in the j-th axis to obtain the j-th first time-frequency spectrogram; wherein j is greater than 0 and less than or equal to N. In this way, the motion data in each axis is converted using the same conversion method, thereby obtaining N first time-frequency spectrograms.

[0053] In the embodiment of the present application, there is no limitation on the method of converting the first motion data into the first time-frequency spectrum diagram.

[0054] In a possible implementation, a short-time Fourier transform (STFT) may be performed on the motion data of the j-th axis to obtain a plurality of eigenvectors; and a corresponding first time-frequency spectrum diagram may be obtained according to the plurality of eigenvectors.

[0055] It can be understood that the basic idea of ​​short-time Fourier transform is to divide the original signal into several sub-signals (or signals in a time window), and then perform Fourier transform on each sub-signal to obtain its corresponding spectrum (i.e., eigenvector). In a possible implementation, these spectra can be spliced ​​in different time sequences to obtain a time-frequency spectrum diagram, thereby revealing the change of signal frequency over time.

[0056] In the embodiment of the present application, there is no restriction on the window function (ie, time window) used for the short-time Fourier transform and the sliding step size of the time window.

[0057] In the embodiment of the present application, the method for converting the first motion data into the first time-frequency spectrum is not limited to the above-mentioned short-time Fourier transform conversion method. In a wider range of time domain and frequency domain analysis, a feature extraction method based on, but not limited to, wavelet transform, fast Fourier transform, signal energy or information entropy, etc. can also be used to perform time-frequency analysis on the motion data of the j-th axis, thereby obtaining the j-th first time-frequency spectrum.

[0058] That is to say, in some other embodiments, step 102 can be implemented as follows: the motion data of the j-th axis is divided into a number of sub-signals (or signals within a time window), and then each sub-signal is subjected to wavelet transform, fast Fourier transform, signal energy calculation or information entropy calculation to obtain its corresponding eigenvector), and these eigenvectors are spliced ​​in different time sequences to obtain the j-th first time-frequency spectrum.

[0059] Figure 2 A schematic diagram of converting motion data provided in an embodiment of the present application into a time-frequency spectrum diagram; Figure 2 As shown, curves 201-203 represent data sequences collected by the inertial sensor on the x, y and z axes, respectively, wherein the horizontal direction of the curves 201-203 represents time, and the vertical direction of the curves 201-203 represents motion parameter values ​​(such as acceleration or angular velocity, etc.). Time-frequency analysis is performed on the curves 201-203 to obtain corresponding time-frequency spectrograms.

[0060] Step 103: Identify the first object according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

[0061] In some embodiments, the characteristics of the first object include one or more of the following:

[0062] (1) emotional characteristics of the first object;

[0063] (2) disease characteristics of the first subject;

[0064] (3) basic physiological characteristics of the first subject; the basic physiological characteristics include gender information and / or age information;

[0065] (4) Identity characteristics of the first object.

[0066] The inventor of the present application found in research and analysis that the corresponding motion states of users in different emotional states are different. For example, when a user is in an angry or happy emotional state, the user's gait and joint amplitude increase. When a user is in a sad or fearful emotional state, the user's gait and joint amplitude decrease.

[0067] In view of this, in an embodiment of the present application, the first motion data collected by the inertial sensor of a single electronic device for the first object is obtained; the first motion data is converted into a first time-frequency spectrum; according to the first time-frequency spectrum and the pre-trained neural network model, the first object is identified to determine the emotional characteristics of the first object; in this way, the emotional characteristics of the first object can be identified only by using the motion data collected by the inertial sensor of a single / single electronic device, without the need to link the identification through the inertial sensors of multiple electronic devices; therefore, the emotional characteristics of the first object can be identified by a single electronic device of ordinary consumer grade (such as a smart phone, a smart watch or a smart bracelet, etc.). This is obviously meaningful for improving the user experience of ordinary consumer grade electronic devices. For example, based on this kind of senseless identification for the user, the current emotional characteristics (i.e., emotional state) of the user are determined, and based on this, at least one of suitable music, videos, articles, etc. is recommended to the user, so as to meet the current emotional needs of the user. It can be seen that the method for identifying emotional characteristics provided by the embodiment of the present application can protect the user's face privacy and can recognize the user's emotional characteristics without the user having to shoot images or audio. This is a better senseless identification method.

[0068] The inventor of the present application found in the research and analysis that the corresponding motion state of the user in different physical health states is different. In view of this, in the embodiment of the present application, the first motion data collected by the inertial sensor of a single electronic device for the first object is obtained; the first motion data is converted into a first time-frequency spectrum; the first object is identified according to the first time-frequency spectrum and the pre-trained neural network model to determine the disease characteristics of the first object; in this way, the disease characteristics of the first object can be identified only by using the motion data collected by the inertial sensor of a single / single electronic device, without the need to link the identification through the inertial sensors of multiple electronic devices; therefore, the disease characteristics of the first object can be identified by a single electronic device of ordinary consumer grade (such as a smart phone, a smart watch or a smart bracelet, etc.). This is obviously meaningful for improving the user experience of ordinary consumer-grade electronic devices. For example, based on this senseless identification for the user, the current disease characteristics of the user are determined, and at least one of the appropriate methods of conditioning the body, diet structure, exercise tutorial, etc. is recommended to the user based on this. It can be seen that the disease characteristic identification method provided by the embodiment of the present application can protect the user's facial privacy and can identify the user's disease characteristics without the user having to take images or audio. This is a better non-invasive identification method.

[0069] In the embodiment of the present application, there is no limitation on the type of the disease characteristics. For example, the type of the disease characteristics includes but is not limited to: chronic diseases and / or falls, etc.

[0070] In an embodiment of the present application, the first motion data collected by the inertial sensor of a single electronic device for the first object is obtained; the first motion data is converted into a first time-frequency spectrum; the first object is identified according to the first time-frequency spectrum and the pre-trained neural network model to determine the basic physiological characteristics of the first object; in this way, the basic physiological characteristics of the first object can be identified only by using the motion data collected by the inertial sensor of a single / single electronic device, without the need for linkage identification by the inertial sensors of multiple electronic devices; therefore, the basic physiological characteristics of the first object can be identified by a single electronic device of ordinary consumer grade (such as a smart phone, smart watch or smart bracelet, etc.). This is obviously meaningful for improving the user experience of ordinary consumer-grade electronic devices. For example, based on this kind of senseless identification for the user, the basic physiological characteristics of the user of the electronic device (such as the age group to which he belongs) are determined, and based on this, it is determined whether the user is the owner of the electronic device. For example, based on the identified age group, it is determined that the user is a child, and the screen is locked. For another example, based on the identified age group, it is determined that the user is an elderly person, and information that suits the preferences of the elderly is recommended.

[0071] In the embodiment of the present application, there is no limitation on the type of the basic physiological characteristics. For example, the type of the basic physiological characteristics includes but is not limited to: gender and / or age group, etc.

[0072] In the embodiment of the present application, there is no restriction on the type of the identity characteristics. For example, the type of the identity characteristics includes but is not limited to: target users or non-target users, etc. The information identification method provided in the embodiment of the present application can be used to identify the identity characteristics of the first object; in this way, the identity characteristics of the first object can be identified through a single ordinary consumer-grade electronic device (such as a smart phone, a smart watch or a smart bracelet, etc.). This is obviously meaningful for improving the user experience of ordinary consumer-grade electronic devices. For example, based on this kind of imperceptible recognition for the user, it is determined whether the current user of the electronic device is the owner of the electronic device; if not, the screen is locked.

[0073] For step 103, the first object is identified according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

[0074] In the embodiment of the present application, there is no limitation on the structure of the pre-trained neural network model, and the pre-trained neural network model may be any of a variety of neural network models, as long as the first time-frequency spectrum generated based on the motion data of the first object can be used to identify the characteristics of the first object. The pre-trained neural network model may be a deep neural network model or a convolutional neural network model, etc., and there is no specific limitation on this.

[0075] Figure 3 Schematic diagram of the structure of the pre-trained neural network model provided in the embodiment of the present application Figure 1 ;like Figure 3 As shown in the figure, Conv2d is a two-dimensional convolution, BN stands for batch normalization, ReLU is the activation function, maxpool is the maximum pooling, and adaptive maxpool means adaptive maximum pooling to adapt to input dimensions of different scales.

[0076] For step 103, the first object is identified according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

[0077] Furthermore, in some embodiments, Figure 4 As shown, step 103 may include the following steps 401 and 402:

[0078] Step 401: input the N first time-frequency spectrograms into the pre-trained neural network model to obtain the probabilities that the first object belongs to different candidate categories; different candidate categories correspond to the same characteristic type.

[0079] It should be noted that different candidate categories correspond to the same characteristic type. For example, the characteristic type is gender information in the basic physiological characteristics, and the candidate categories corresponding to the characteristic include but are not limited to male and / or female. For another example, the characteristic type is age information in the basic physiological characteristics, and the candidate categories corresponding to the characteristic include multiple different age groups. For another example, the characteristic type is emotional characteristics, and the candidate categories corresponding to the characteristic include but are not limited to: one or more of anger, happiness, sadness, fear, etc.

[0080] In some embodiments, the pre-trained neural network model is obtained by training the neural network model using the first sample data as input of the neural network model and taking the value of the loss function less than a first threshold as a goal;

[0081] The first sample data includes second motion data collected by the inertial sensor of a single electronic device for multiple different second objects respectively; exemplarily, in some embodiments, the first sample data includes second motion data collected by the inertial sensor of a single electronic device for one or more specific positions (such as wrist, thigh, ankle, etc.) of multiple different second objects.

[0082] The output of the neural network model includes but is not limited to one or more of the following:

[0083] scores of the second subject on multiple emotion dimensions;

[0084] the disease characteristics of the second subject;

[0085] The basic physiological characteristics of the second object; the basic physiological characteristics include gender information and / or age information;

[0086] An identity characteristic of the second object.

[0087] The loss function is used to indicate the difference between the result output by the neural network model and the true value. For example, if the output result of the neural network model includes the scores of the second subject in multiple emotional dimensions, the corresponding true value is the score given by the interviewer himself in each emotional dimension; for another example, if the output result of the neural network model includes basic physiological characteristics such as gender and / or age group, the true value is the real gender and / or real age group of the second subject.

[0088] In this way, the trained neural network model has the ability to identify the characteristics of the first object based on the motion data collected by the inertial sensor of a single electronic device.

[0089] Step 402: Determine the characteristics of the first object according to the probabilities of the different candidate categories.

[0090] Furthermore, in some embodiments, step 401 may include: inputting the N first time-frequency spectrograms corresponding to the motion data of the i-th time period into the pre-trained neural network model to obtain the probabilities that the first object belongs to different candidate categories; wherein i is greater than 0 and less than or equal to the number of time periods. Based on this, in some embodiments, the different candidate categories are voted using a soft voting algorithm and the probabilities of the different candidate categories corresponding to the multiple time periods, so that the winning candidate category is used as the characteristic of the first object.

[0091] Of course, in the embodiment of the present application, the further implementation of step 402 is not limited to the soft voting method, but may also be a hard voting method. Alternatively, the candidate category corresponding to the maximum probability among the probabilities of the different candidate categories corresponding to a time period may be determined as the characteristic of the first object.

[0092] For step 103, the first object is identified according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

[0093] Furthermore, in some other embodiments, step 103 may include: inputting the N first time-frequency spectrograms into the pre-trained neural network model to obtain scores of the first object in multiple different emotional dimensions; wherein the scores of the first object in multiple different emotional dimensions are used to measure the emotional characteristics of the first object, that is, the emotional characteristics of the first object are represented by the scores in the multiple different emotional dimensions.

[0094] Furthermore, in some other embodiments, in a possible implementation method, step 103 may include: inputting the N first time-frequency spectrograms into the pre-trained neural network model to obtain scores of the first object in multiple different emotional dimensions; determining the emotional characteristics of the first object based on the scores in the multiple different emotional dimensions.

[0095] In the embodiment of the present application, there is no limitation on which specific emotional dimensions the multiple different emotional dimensions are. For example, in one possible implementation, the multiple different emotional dimensions may include but are not limited to: at least two of valence, arousal, dominance and liking. Among them, Valence (emotional positivity or negativity) indicates the degree of positivity or negativity of the emotional experience. Arousal (emotional activation level) indicates the activation level or intensity of the emotional experience. Dominance (emotional control) indicates the sense of control or dominance of the situation by an individual in the emotional state. Liking (degree of liking) indicates the degree of liking or pleasure of the emotional experience by an individual.

[0096] In the embodiment of the present application, emotional characteristics can also be understood as affective characteristics.

[0097] The following examples describe possible implementations of the information identification method described in one or more of the above embodiments.

[0098] In the embodiment of the present application, a single-point sensor fixed on the thigh, wrist or other position where the mobile phone is usually placed is used to identify various attributes of the user, and an identification result similar to that of a multi-point sensor combination is obtained. The application of the walking gait sensor for identity authentication (i.e., identity characteristics, such as target user or non-target user), basic physiological attributes (i.e., basic physiological characteristics, such as gender and / or age, etc.), disease attributes (i.e., disease characteristics, such as chronic diseases and / or falls, etc.), and emotional attributes (i.e., emotional characteristics, such as anger, happiness, sadness, fear) on ordinary consumer-grade devices has been confirmed. Among them, ordinary consumer-grade devices, for example, can be user devices such as smart phones, smart bracelets, smart watches or smart rings.

[0099] The following describes the technical solutions for walking gait behavior in three aspects: identity authentication, physiological attribute recognition, and emotion recognition.

[0100] 1. Identity authentication based on walking gait

[0101] In some embodiments, the three-dimensional data sequence of x, y, and z axes is converted into a time-frequency spectrum, which is then converted into a three-channel combination of w*h*d as a feature input into a single-class classification network or a multi-classification network for identity recognition. Wherein, w represents the time-frequency spectrum corresponding to the data sequence of the x-axis; h represents the time-frequency spectrum corresponding to the data sequence of the y-axis; d represents the time-frequency spectrum corresponding to the data sequence of the z-axis, for example Figure 2 shown.

[0102] For example, if the inertial sensor is a three-axis accelerometer, the three-dimensional data sequence of x, y, and z axes is the acceleration in three directions; or if the inertial sensor is a three-axis gyroscope, the three-dimensional data sequence of x, y, and z axes is the angular velocity in three directions.

[0103] In some embodiments, the short-time Fourier transform (STFT) can be used to frame and window the original signal in the time domain, and perform feature extraction, thereby solving the feature vector of the short-time Fourier transform. Such an STFT feature extraction process is performed for the three-dimensional data sequences of the x, y, and z axes respectively. Then, the feature vector solved by the STFT is converted into a visual image according to different colors under the same color system, for example Figure 2 The time-frequency spectrum is shown.

[0104] It should be noted that in this application, there is no restriction on the feature extraction method of the three-dimensional data sequence of the x, y, and z axes. STFT is only a method for time domain, frequency domain analysis and feature extraction. In a wider range of time domain and frequency domain analysis, the following methods may also be used but are not limited to: wavelet transform, fast Fourier transform, signal energy, information entropy and other feature extraction methods to perform time-frequency analysis on the three-dimensional data sequence of the x, y, and z axes.

[0105] Figure 3 A schematic diagram of the structure of a pre-trained neural network model provided in an embodiment of the present application; Figure 3 As shown in the figure, Conv2d is a two-dimensional convolution, BN stands for batch normalization, ReLU is an activation function, maxpool is maximum pooling, and adaptive maxpool represents adaptive maximum pooling to adapt to input dimensions of different scales. During the training phase of the neural network model, the sample data used to train the neural network model includes time-frequency spectrograms corresponding to multiple users, which are obtained by converting the gait data of the multiple users.

[0106] In some embodiments, the classifier can select a single-class classifier, set the gait characteristics of the target user as the positive class, and the non-target user as the negative class. Due to the imbalance of positive and negative samples, the single-class classifier based on the neural network can adjust the sample weight through sample_weight (that is, the sample weight is used as an input information of the neural network model), or classify by building a single output node (probability output) in the output layer. This is an idea of ​​​​outlier detection.

[0107] In some embodiments, the neural network model may also support multi-user recognition, for example, multiple classifiers may be connected to recognize different users.

[0108] The inventor of the present application collected motion data at the wrist, thigh and ankle based on a single / single-point inertial sensor, and verified the above information recognition method based on this. The verification results are shown in Table 1 below. Based on the verification results, it can be seen that the accuracy of the above method is still high with only a single inertial sensor. The verification at these three positions, rather than other positions, is mainly due to the fact that these three positions are where single-point sensors such as mobile phones are often placed.

[0109] Table 1

[0110] Body Position wrist thigh ankle Accuracy 89.34％ 93.96％ 89.18％

[0111] For example, the motion data of the user's arm is collected through the mobile phone held by the user, the smart bracelet or smart watch worn on the user's wrist, or the smart ring worn on the user's finger, and the collected data is converted into a time-frequency spectrum diagram, and then input into a pre-trained neural network model to identify whether the user is a target user or a non-target user. For another example, the motion data of the user's leg is collected through a smart device such as a mobile phone in the user's pocket, and the collected data is converted into a time-frequency spectrum diagram, and then input into a pre-trained neural network model to identify whether the user is a target user or a non-target user. For another example, the motion data of the user's leg is collected through a sensor attached to the user's foot or a smart anklet worn by the user, and the collected data is converted into a time-frequency spectrum diagram, and then input into a pre-trained neural network model to identify whether the user is a target user or a non-target user.

[0112] 2. Identify physiological attributes based on walking gait

[0113] In the embodiment of the present application, the direction of physiological attribute recognition based on walking gait mainly refers to the two basic attributes of gender and age, and can also be expanded to other types of chronic diseases, falls and other physiological attributes.

[0114] In a possible implementation, equally spaced sampling is used. For each person's gait signal, 150 points are sampled each time as input signal data to complete the construction of the time-frequency spectrum. Since the sampling rate of the inertial sensor (including but not limited to an accelerometer or a gyroscope) of the data is 100 Hz, the input signal of each sample of the time-frequency spectrum corresponds to a gait signal of 1.5 seconds (i.e., 150 sampling values).

[0115] Input the time spectrum to Figure 5 After the neural network shown, the probability of gender or age group is generated; then the final classification prediction result, that is, the classification result of gender or age, is obtained through the probabilistic soft voting method of the time-frequency spectrum diagram of multiple fragment time series.

[0116] Figure 6 A schematic diagram of a method for determining a final classification prediction result provided in an embodiment of the present application; Figure 6 As shown, the gait signal 601 of the same person is divided into data of different time segments, and the gender and age group prediction is completed according to the same user prediction network. It is understandable that the gender and age group predicted based on the gait data of different segments of the same person may be inconsistent. Therefore, voting can be performed based on the prediction results (i.e., the corresponding probabilities) corresponding to each segment, and the probabilities corresponding to the gender or age group predicted based on the gait data of each segment are summed up, and the prediction is performed based on the final summed value to obtain the final classification prediction result.

[0117] The following example illustrates the implementation logic of the probabilistic soft voting method:

[0118] Assumptions:

[0119] The gait data of time segment 1 is input into the pre-trained neural network model 1, and the results are: the probability of being a male is 68%, and the probability of being a female is 32%;

[0120] The gait data of time segment 2 is input into the pre-trained neural network model 1, and the results are: the probability of being a male is 24%, and the probability of being a female is 76%;

[0121] The gait data of time segment 3 is input into the pre-trained neural network model 1, and the results are: the probability of being a male is 52%, and the probability of being a female is 48%;

[0122] Then, the final probability of being a male is the average of the above three probabilities: (68% + 24% + 52%) / 3 = 48%; the final probability of being a female is the average of the above three probabilities: (32% + 76% + 48%) / 3 = 52%;

[0123] Therefore, the final classification prediction result is: female.

[0124] The logic for determining the final classification prediction results of the age groups is the same as above.

[0125] Table 2 is the test results provided by the embodiments of the present application. As shown in Table 2, the accuracy of predicting gender based on the angular velocity collected by the gyroscope (single inertial sensor) is 96.67%, and the accuracy of predicting gender based on the acceleration collected by the accelerometer (single inertial sensor) is 96.67%; in the two-age classification with 60 years old as the dividing line, the accuracy of predicting age based on the angular velocity collected by the gyroscope is 96.67%, and the accuracy of predicting age based on the acceleration collected by the accelerometer is 96.67%; in the three-age classification of 20-40 years old, 50-59 years old, and 60-69 years old, the accuracy of predicting age based on the angular velocity collected by the gyroscope is 93.33%, and the accuracy of predicting age based on the acceleration collected by the accelerometer is 93.33%.

[0126] Table 2

[0127]

[0128] In other embodiments, it is also possible to Figure 7 The neural network shown performs multi-task prediction on the age and gender classification results, where 701 outputs the gender classification result and 702 outputs the age classification result.

[0129] 3. Identify emotional state based on walking gait

[0130] Since the amplitude of the thigh raising angle of all emotions is different from the neutral gait, these effects can be the basis for completing emotion recognition through gait signals. For example, when the user is in an angry or happy emotional state, the user's gait and joint amplitude increase. When the user is in a sad or fearful emotional state, the user's gait and joint amplitude decrease.

[0131] In some embodiments, a neural network structure similar to the above scheme can be used, but according to the classification method of the Pluzik emotion wheel model, the scores of four more significant emotions / emotional dimensions (such as valence, arousal, dominance and liking) are selected as recognition targets, and the inertial sensor data of the bracelet during walking is selected as the recognition data source.

[0132] Different emotions are decomposed into indicators in four directions: valence (emotional dimension 1), arousal (emotional dimension 2), dominance (emotional dimension 3), and liking (emotional dimension 4); but not limited to these emotion evaluation dimensions; each direction / dimension has a continuous floating point data score between 1-10 to measure a person's emotions. During the online use stage, the AI ​​model outputs the scores of these four dimensions, based on which the user's current emotional characteristics can be calculated.

[0133] The inventor of the present application verified the above method by modeling the scoring of indicators in these different dimensions and then performing numerical regression prediction. The four-dimensional emotional state of the test user is identified through the regressed floating-point scoring. This embodiment uses MAE (Mean Absolute Error) as a measurement indicator. The experimental results (test results) are as low as 0.77, and the regression model results are very close to the actual score value of the user's emotional scale. The inventor of the present application found through the above research and analysis that motion data collected only based on a single inertial sensor can still accurately identify the user's emotional characteristics.

[0134] It can be understood that in the embodiments of the present application, a single-point inertial sensor is used to identify the user's identity (target user or non-target user), basic physiological characteristics (gender, age group), emotional characteristics, and disease characteristics (chronic diseases, falls), and good recognition effects are achieved, solving the problem that the traditional bioinformatics field requires a combination of multiple inertial sensors to identify physiological characteristics.

[0135] It can be understood that in the embodiment of the present application, the user's physiological characteristics are identified by using an integrated learning soft voting method in the form of a multi-frame sequence, without the need for manually defined whole-frame recognition, which has a good application in the problem of gait recognition of variable-length walking. Moreover, compared with the hard voting method, it is more robust. For example, in the hard voting method, if the number of classifications and the number of sub-frames are multiples or the same, the target category cannot be identified.

[0136] It can be understood that in the embodiments of the present application, the user's walking gait is unconsciously recognized by physiological and emotional recognition through a bracelet or a handheld mobile phone, without the need to collect the user's facial image, audio, etc., which causes less interference to the user.

[0137] It can be understood that in the embodiment of the present application, inertial sensor data is collected and identified with low power consumption.

[0138] It should be noted that the execution entity of the information identification method provided in the embodiment of the present application and the electronic device used to collect the motion data of the first object can be the same device or different devices.

[0139] It should be noted that although the steps of the method in the present application are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.; or, steps in different embodiments may be combined into a new technical solution.

[0140] Based on the foregoing embodiments, the embodiments of the present application provide an information identification device, which includes the modules included and the units included in the modules, which can be implemented by a processor; of course, it can also be implemented by a specific logic circuit; in the implementation process, the processor can be an AI acceleration engine (such as NPU, etc.), GPU, central processing unit (CPU), microprocessor (MPU), digital signal processor (DSP) or field programmable gate array (FPGA), etc.

[0141] Figure 8 A schematic diagram of the structure of the information identification device provided in the embodiment of the present application; Figure 8 As shown, the information identification device 800 includes:

[0142] An acquisition module 801 is configured to acquire first motion data collected by an inertial sensor of a single electronic device for a first object;

[0143] A conversion module 802, configured to convert the first motion data into a first time-frequency spectrum diagram;

[0144] The identification module 803 is configured to identify the first object according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

[0145] In some embodiments, the characteristics of the first object include one or more of the following:

[0146] emotional characteristics of the first subject;

[0147] a disease characteristic of the first subject;

[0148] The basic physiological characteristics of the first object; the basic physiological characteristics include gender information and / or age information;

[0149] An identity characteristic of the first object.

[0150] In some embodiments, the first motion data includes motion data of the first object in N axes, where N is greater than 1; converting the first motion data into a first time-frequency spectrum diagram includes: converting the motion data in the N axes respectively to obtain N first time-frequency spectrum diagrams.

[0151] In some embodiments, the first object is identified based on the first time-frequency spectrum graph and a pre-trained neural network model to determine the characteristics of the first object, including: inputting the N first time-frequency spectrum graphs into the pre-trained neural network model to obtain the probabilities that the first object belongs to different candidate categories; different candidate categories correspond to the same characteristic type; and determining the characteristics of the first object based on the probabilities of the different candidate categories.

[0152] In some embodiments, the first motion data includes motion data of multiple time periods; the inputting the N first time-frequency spectrograms into the pre-trained neural network model to obtain the probabilities that the first object belongs to different candidate categories, includes: inputting the N first time-frequency spectrograms corresponding to the motion data of the i-th time period into the pre-trained neural network model to obtain the probabilities that the first object belongs to different candidate categories; wherein i is greater than 0 and less than or equal to the number of time periods; determining the characteristics of the first object based on the probabilities of the different candidate categories includes: voting for the different candidate categories using a soft voting algorithm and the probabilities of the different candidate categories corresponding to the multiple time periods, so as to use the winning candidate categories as the characteristics of the first object.

[0153] In some embodiments, the characteristics of the first object include emotional characteristics; the identifying the first object based on the first time-frequency spectrum graph and the pre-trained neural network model to determine the characteristics of the first object includes: inputting the N first time-frequency spectrum graphs into the pre-trained neural network model to obtain scores of the first object in multiple different emotional dimensions; wherein the scores of the first object in multiple different emotional dimensions are used to measure the emotional characteristics of the first object.

[0154] In some embodiments, the electronic device is placed at a specific position of the first object; wherein the specific position includes: a wrist, a thigh, or an ankle.

[0155] The description of the above device embodiment is similar to the description of the above method embodiment, and has similar beneficial effects as the method embodiment. For technical details not disclosed in the device embodiment of the present application, please refer to the description of the method embodiment of the present application for understanding.

[0156] It should be noted that the division of modules in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional unit in each embodiment of the present application may be integrated into a processing unit, or may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of a software functional unit. It may also be implemented in the form of a combination of software and hardware.

[0157] It should be noted that in the embodiment of the present application, if the above method is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiment of the present application can be essentially or partly embodied in the form of a software product that contributes to the relevant technology. The computer software product is stored in a storage medium, including several instructions to enable an electronic device to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM), a magnetic disk or an optical disk. In this way, the embodiment of the present application is not limited to any specific combination of hardware and software.

[0158] An embodiment of the present application provides an electronic device, Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application; Fig. 9As shown, the electronic device 900 includes a memory 901 and a processor 902, wherein the memory 901 stores a computer program that can be run on the processor 902, and the processor 902 implements the steps in the method provided in the above embodiment when executing the program.

[0159] It should be noted that the memory 901 is configured to store instructions and applications executable by the processor 902, and can also cache data to be processed or already processed by the processor 902 and various modules in the electronic device 900 (for example, image data, audio data, voice communication data, and video communication data), which can be implemented through flash memory (FLASH) or random access memory (Random Access Memory, RAM).

[0160] In the embodiments of the present application, there is no limitation on the type of electronic device, and the electronic device may be any device with information processing capabilities. For example, the electronic device may be a smart phone, a laptop, a tablet computer, a smart bracelet, a smart watch, a smart anklet, a smart ring, a smart sports shoe, etc.

[0161] An embodiment of the present application also provides a computer-readable storage medium for storing a computer program.

[0162] Optionally, the computer-readable storage medium can be applied to the electronic device in the embodiments of the present application, and the computer program enables the electronic device to execute the various methods in the embodiments of the present application, which will not be described in detail here for the sake of brevity.

[0163] An embodiment of the present application also provides a computer program product, including computer program instructions.

[0164] Optionally, the computer program product may be applied to the electronic device in the embodiments of the present application, and the computer program instructions enable the electronic device to execute the various methods in the embodiments of the present application, which will not be described in detail here for the sake of brevity.

[0165] The embodiment of the present application also provides a computer program.

[0166] Optionally, the computer program may be applied to the electronic device in the embodiments of the present application. When the computer program runs on the electronic device, the electronic device executes the various methods in the embodiments of the present application. For the sake of brevity, they are not described here in detail.

[0167] It should be noted here that the description of the above electronic device, storage medium, computer program product and computer program embodiment is similar to the description of the above method embodiment, and has similar beneficial effects as the method embodiment. For technical details not disclosed in the electronic device, storage medium, computer program product and computer program embodiment of this application, please refer to the description of the method embodiment of this application for understanding.

[0168] It should be understood that "one embodiment" or "an embodiment" or "some embodiments" mentioned throughout the specification means that specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in one embodiment" or "in some embodiments" appearing throughout the specification may not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in various embodiments of the present application, the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application. The above-mentioned sequence numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments. The above description of each embodiment tends to emphasize the differences between the various embodiments, and the same or similar aspects can be referenced to each other. For the sake of brevity, this article will not repeat them.

[0169] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, object A and / or object B can represent three situations: object A exists alone, object A and object B exist at the same time, and object B exists alone.

[0170] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0171] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as: multiple modules or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or modules can be electrical, mechanical or other forms.

[0172] The modules described above as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules; they may be located in one place or distributed on multiple network units; some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0173] In addition, all functional modules in the embodiments of the present application may be integrated into one processing unit, or each module may be a separate unit, or two or more modules may be integrated into one unit; the above-mentioned integrated modules may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0174] A person skilled in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, etc., various media that can store program codes.

[0175] Alternatively, if the above-mentioned integrated unit of the present application is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application can essentially or in other words, the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling an electronic device to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0176] The methods disclosed in several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0177] The features disclosed in several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0178] The features disclosed in several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments or device embodiments.

[0179] The above is only an implementation method of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An information identification method, characterized in that: The method comprises: Acquire first motion data collected by an inertial sensor of a single electronic device for a first object; converting the first motion data into a first time-frequency spectrum; The first object is identified according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

2. The method according to claim 1, characterized in that The characteristics of the first object include one or more of the following: emotional characteristics of the first subject; a disease characteristic of the first subject; The basic physiological characteristics of the first object; the basic physiological characteristics include gender information and / or age information; An identity characteristic of the first object.

3. The method according to claim 2, characterized in that The first motion data includes motion data of the first object in N axes respectively, where N is greater than 1; The converting the first motion data into a first time-frequency spectrum diagram includes: converting the motion data in the N axes respectively to obtain N first time-frequency spectrum diagrams.

4. The method according to claim 2 or 3, characterized in that: The identifying the first object according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object includes: Inputting the N first time-frequency spectrograms into the pre-trained neural network model to obtain the probabilities that the first object belongs to different candidate categories respectively; different candidate categories correspond to the same characteristic type; A characteristic of the first object is determined based on the probabilities of the different candidate categories.

5. The method according to claim 4, characterized in that The first motion data includes motion data of multiple time periods; The step of inputting the N first time-frequency spectrograms into the pre-trained neural network model to obtain the probabilities that the first objects belong to different candidate categories respectively includes: Inputting N first time-frequency spectrograms corresponding to the motion data of the i-th time period into the pre-trained neural network model to obtain the probabilities that the first object belongs to different candidate categories respectively; wherein i is greater than 0 and less than or equal to the number of the time periods; The determining the characteristic of the first object according to the probabilities of the different candidate categories includes: The different candidate categories are voted by using a soft voting algorithm and the probabilities of the different candidate categories corresponding to the multiple time periods, so as to use the winning candidate category as the characteristic of the first object.

6. The method according to claim 2 or 3, characterized in that: The characteristics of the first object include emotional characteristics; The identifying the first object according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object includes: The N first time-frequency spectrograms are input into the pre-trained neural network model to obtain scores of the first object in multiple different emotional dimensions; wherein the scores of the first object in multiple different emotional dimensions are used to measure the emotional characteristics of the first object.

7. The method according to any one of claims 1 to 6, characterized in that The electronic device is placed at a specific position of the first object; wherein the specific position includes: a wrist, a thigh or an ankle.

8. An information recognition device, characterized in that: The device comprises: An acquisition module configured to acquire first motion data collected by an inertial sensor of a single electronic device for a first object; a conversion module, configured to convert the first motion data into a first time-frequency spectrum; The recognition module is configured to recognize the first object according to the first time-frequency spectrum diagram and the pre-trained neural network model to determine the characteristics of the first object.

9. An electronic device, characterized in that: The electronic device comprises: a processing unit and an inertial sensor; wherein, The inertial sensor is used to collect first motion data for the first object; The processing unit is used to obtain the first motion data; convert the first motion data into a first time-frequency spectrum; and identify the first object according to the first time-frequency spectrum and a pre-trained neural network model to determine the characteristics of the first object.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images