Human-machine recognition method and device

Abstract

The present application provides a human-machine recognition method and device, which relates to the field of artificial intelligence technology and can also be used in the field of financial technology. The method includes: obtaining a sensor information group of a target terminal device; obtaining a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model; the preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm with a plurality of historical sensor information groups and their respective corresponding actual human-machine recognition results. The present application can improve the accuracy and efficiency of human-machine recognition.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular, to a human-machine recognition method and device. Background Art

[0002] The human-machine recognition technology is one of the most widely used technologies in the field of Machine Learning (ML). Its essence is to identify whether the operation is initiated by a human or a machine based on the collected environmental information and specific technical means. The usage scenarios almost cover all aspects such as registration, login, and information modification of information systems. For example, in online marketing activities, coupons are distributed according to the mobile phone numbers entered by users. To prevent black production from using a large number of mobile phone numbers and robots to quickly grab coupons in batches, it is necessary to determine whether the current submission behavior is a human behavior or a machine behavior when receiving user submissions. Machine behavior will cause losses to enterprises and represents illegal operations.

[0003] A common human-machine recognition method is to generate a verification code on the server side and require the user to correctly enter the verification code when submitting. The correct entry of the verification code by the user represents a human operation. If the verification code cannot be correctly entered, it is determined as a non-human operation; it is determined whether the current operation is performed manually based on the entry of the verification code; however, with the application of deep learning technology to the verification code recognition scenario, the verification code technology has been difficult to accurately identify human-machine operations; a new type of verification code will be recognized by deep learning in a short time; moreover, in some specific scenarios, overly complex verification codes will also affect the user experience and increase the burden on normal users. Summary of the Invention

[0004] Aiming at the problems in the prior art, this application proposes a human-machine recognition method and device, which can improve the accuracy and efficiency of human-machine recognition.

[0005] To solve the above technical problems, this application provides the following technical solutions:

[0006] In a first aspect, this application provides a human-machine recognition method, including:

[0007] Obtain a sensor information group of a target terminal device;

[0008] Obtain a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model;

[0009] The preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm with multiple historical sensor information groups and their respective corresponding actual human-machine recognition results.

[0010] Further, before obtaining the sensor information group of the target terminal device, it further includes:

[0011] Obtain a sample data set, where each piece of sample data in the sample data set includes: a unique historical sensor information group and its corresponding actual human-machine recognition result, and the actual human-machine recognition result is: manual operation or machine operation;

[0012] Use the sample data set to train a k-nearest neighbor classification model based on the dynamic time warping algorithm to obtain the human-machine classification model.

[0013] Further, each group of historical sensor information groups includes: historical sensor information collected multiple times corresponding to one historical operation;

[0014] Correspondingly, the obtaining of the sample data set includes:

[0015] Obtain a sample data set;

[0016] Delete the sample data in the sample data set whose number of acquisitions does not meet the preset acquisition number range;

[0017] Normalize the historical sensor information groups in the sample data set.

[0018] Further, the obtaining of the sensor information group of the target terminal device includes:

[0019] Obtain the sensor information collected multiple times within the acquisition time range, and all the sensor information constitutes the sensor information group, and the acquisition time range includes: the start time to the end time of the target operation.

[0020] Further, each sensor information includes: direction sensor, linear acceleration sensor, gravity sensor, acceleration sensor, gyroscope, and magnetic sensor information.

[0021] Further, the obtaining of the sensor information collected multiple times within the acquisition time range includes:

[0022] Within the acquisition time range, whenever the sensor information changes, collect the current sensor information.

[0023] In a second aspect, the present application provides a human-machine recognition device, including:

[0024] An acquisition module, configured to acquire a sensor information group of a target terminal device;

[0025] A recognition module, configured to obtain a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model;

[0026] The preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm using multiple historical sensor information groups and their respective corresponding actual human-machine recognition results.

[0027] Further, the human-machine recognition device further includes:

[0028] A sample acquisition module, configured to acquire a sample data set, where each sample data in the sample data set includes: a unique historical sensor information group and its corresponding actual human-machine recognition result, and the actual human-machine recognition result is: manual operation or machine operation;

[0029] A training module, configured to train a k-nearest neighbor classification model based on the dynamic time warping algorithm using the sample data set to obtain the human-machine classification model.

[0030] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the human-machine recognition method when executing the program.

[0031] In a fourth aspect, the present application provides a computer-readable storage medium, on which computer instructions are stored, and the human-machine recognition method is implemented when the instructions are executed.

[0032] As can be seen from the above technical solutions, the present application provides a human-machine recognition method and device. Among them, the method includes: acquiring a sensor information group of a target terminal device; obtaining a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model; the preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm using multiple historical sensor information groups and their respective corresponding actual human-machine recognition results, which can improve the accuracy and efficiency of human-machine recognition; without the subjective participation of the user, the user experience is improved, and the defect that the verification code technology fails can be solved; it can effectively prevent fraud and avoid losses such as enterprise property. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0034] Figure 1 It is a flowchart of the human-machine recognition method in the embodiment of the present application;

[0035] Figure 2It is a schematic flowchart of step 001 and step 002 of the human-machine recognition method in the embodiments of the present application;

[0036] Figure 3 It is a schematic flowchart of the model training process in the application example of the present application;

[0037] Figure 4 It is a schematic flowchart of the model application process in the application example of the present application;

[0038] Figure 5 It is a schematic structural diagram of the human-machine recognition device in the embodiments of the present application;

[0039] Figure 6 It is a schematic structural diagram of the human-machine recognition device in the application example of the present application;

[0040] Figure 7 It is a schematic block diagram of the system composition of the electronic device in the embodiments of the present application. Specific embodiments

[0041] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0042] Based on this, in order to improve the accuracy and efficiency of human-machine recognition, the embodiments of the present application provide a human-machine recognition device, which can be a server or a client device. The client device may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, and a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, smart watches, and smart bracelets, etc.

[0043] In actual applications, the part for performing human-machine recognition can be executed on the server side as described above, or all operations can be completed in the client device. Specifically, it can be selected according to the processing ability of the client device and the limitations of the user usage scenario, etc. The present application does not make any limitations in this regard. If all operations are completed in the client device, the client device may further include a processor.

[0044] The above-mentioned client device may have a communication module (i.e., communication unit), which can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side, and in other implementation scenarios, it may also include a server of an intermediate platform, such as a server of a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, or a server cluster composed of multiple servers, or a server structure of a distributed device.

[0045] Any suitable network protocol can be used for communication between the server and the client device, including network protocols that have not been developed as of the filing date of this application. The network protocol may, for example, include TCP / IP protocol, UDP / IP protocol, HTTP protocol, HTTPS protocol, etc. Of course, the network protocol may also, for example, include RPC protocol (Remote Procedure Call Protocol) and REST protocol (Representational State Transfer) used on top of the above-mentioned protocols.

[0046] It should be noted that the human-machine recognition method and device disclosed in this application can be used in the fields of biometric recognition and financial technology, and can also be used in any field other than the fields of biometric recognition and financial technology. The application fields of the human-machine recognition method and device disclosed in this application are not limited.

[0047] Specifically, it will be described through the following various embodiments.

[0048] To improve the accuracy and efficiency of human-machine recognition, this embodiment provides a human-machine recognition method whose execution entity is a human-machine recognition device. The human-machine recognition device includes, but is not limited to, a server, such as Figure 1 As shown, the method specifically includes the following content:

[0049] Step 100: Obtain a sensor information group of a target terminal device.

[0050] Specifically, the target terminal device may be a smart phone, which includes various sensors such as a direction sensor, a linear acceleration sensor, a gravity sensor, an acceleration sensor, a gyroscope, and a magnetic sensor. The sensor information group includes: sensor information collected multiple times corresponding to a target operation at the target terminal device, and the sensor information collected each time forms the sensor information group; each sensor information may include: information collected by the direction sensor, linear acceleration sensor, gravity sensor, acceleration sensor, gyroscope, and magnetic sensor of the target terminal device.

[0051] Step 200: Obtain the human-machine recognition result of the target operation at the target terminal device according to the sensor information group and the preset human-machine classification model; the preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm using multiple historical sensor information groups and their respective corresponding actual human-machine recognition results.

[0052] Among them, the target operation is the operation for which human-machine recognition is performed this time; for example, the operation at the target terminal device can be the operation of entering a mobile phone number in the mobile phone number input box on the interface of a mobile phone software of the target terminal device, and the mobile phone software can be a financial trading system.

[0053] It can be understood that the historical sensor information group is a pre-obtained sensor information group for model training; the actual human-machine recognition result corresponding to the historical sensor information group can be a manual operation or a machine operation, and the machine operation can be a robotic arm operation.

[0054] The human-machine recognition method provided in the embodiments of the present application can be applicable to various application scenarios. For example, in the application scenario of marketing and distributing coupons, coupons are distributed according to the mobile phone number entered by the user. The acquisition start time of the sensor information group occurs when the input box in the APP obtains focus, and the acquisition ends when the input box loses focus. This can prevent black production from using robots to quickly grab coupons, ensure the marketing effect, safeguard the interests of the enterprise, and at the same time, avoid overly complex verification codes from affecting the user experience.

[0055] To further improve the reliability of the human-machine classification model, and then use the reliable human-machine classification model to improve the reliability of human-machine classification, refer to Figure 2 In an embodiment of the present application, before step 100, it further includes:

[0056] Step 001: Obtain a sample data set, and each piece of sample data in the sample data set includes: a unique historical sensor information group and its corresponding actual human-machine recognition result, and the actual human-machine recognition result is: a manual operation or a machine operation.

[0057] Step 002: Use the sample data set to train a k-nearest neighbor classification model based on the dynamic time warping algorithm to obtain the human-machine classification model.

[0058] In an example, in the scenario where volunteers and robotic arms enter mobile phone numbers on a mobile phone, the collected sensor data is used for testing. When the sample size is 400, the test results are shown in Table 1:

[0059] Table 1

[0060]

[0061] From the perspective of the test results, for scenarios with a small amount of input such as when a user enters a mobile phone number (the accuracy will be even higher if there is more input content), this application can accurately identify human-machine operations with high accuracy; it has strong versatility, simplifies the human-machine recognition process, and can improve the accuracy and efficiency of human-machine recognition.

[0062] In order to further improve the reliability of the sample data set and the reliability of applying the dynamic time warping algorithm, in an embodiment of the present application, each group of historical sensor information groups includes: historical sensor information collected multiple times corresponding to one historical operation; correspondingly, step 001 includes:

[0063] Step 010: Obtain a sample data set.

[0064] Step 020: Delete the sample data in the sample data set whose number of acquisitions does not meet the preset acquisition number range.

[0065] Specifically, the preset acquisition number range can be set according to actual needs, and the present application does not limit this; preferably, the preset acquisition number range can be set to the number of acquisitions being greater than or equal to 30 and less than or equal to 100, and sample data with fewer than 30 or more than 100 acquisitions is excluded.

[0066] Specifically, one historical operation corresponds to a unique group of historical sensor information groups, and the historical operation can be an operation that occurred before the target operation.

[0067] Step 030: Perform normalization processing on the historical sensor information groups in the sample data set.

[0068] Specifically, the deviation normalization method can be used to perform normalization processing on each dimension of the historical sensor information.

[0069] In order to further improve the accuracy of human-machine recognition in scenarios with a small amount of input such as when a user enters a mobile phone number, in an embodiment of the present application, step 100 includes:

[0070] Step 1001: Obtain sensor information collected multiple times within the acquisition time range, and all the sensor information forms the sensor information group. The acquisition time range includes: the start time to the end time of the target operation.

[0071] Among them, each sensor information may include: direction sensor, linear acceleration sensor, gravity sensor, acceleration sensor, gyroscope, and magnetic sensor information.

[0072] In order to further improve the reliability of determining sensor information, in an embodiment of the present application, step 101 includes:

[0073] Within the described acquisition time range, whenever the sensor information changes, the current sensor information is acquired.

[0074] To improve and maintain stable recognition accuracy and enhance the user experience, this application provides an application example of a human-machine recognition method. Based on the operation of a mobile phone terminal using machine learning algorithms, by obtaining various sensor data on the mobile phone side and using an improved machine learning algorithm to classify the collected data, this method includes: model training and model application, which are specifically described as follows:

[0075] See Figure 3 , model training includes:

[0076] S101: Prepare sample data, that is, collect and organize the sample data; S102: Normalize the full amount of sample data; S103: Perform KNN training based on DTW and output the best model, that is, train the model and adjust the parameters to achieve the optimal effect.

[0077] See Figure 4 , model application includes:

[0078] S201: Prepare the data to be predicted, that is, collect and organize the data to be predicted; S202: Normalize the data to be predicted according to the sample data; S203: Based on the prediction results of the trained KNN model, input the data to be predicted into the trained model to calculate the prediction results.

[0079] To further illustrate this solution, this application provides another application example of a human-machine recognition method, including:

[0080] Step 1: Collect sample data.

[0081] Among them, the sample data is the data generated under manual operation and machine operation, and the content is the six types of sensor information of the mobile phone. The six types of sensors are the orientation sensor, linear acceleration sensor, gravity sensor, acceleration sensor, gyroscope, and magnetic sensor. Each type of sensor has information in 3 directions: the x-axis, y-axis, and z-axis, for a total of 18-dimensional information. The information collected at one time point is as follows:

[0082] {[Three-dimensional orientation information], [Three-dimensional linear acceleration information], [Three-dimensional gravity sensor information], [Three-dimensional acceleration information], [Three-dimensional gyroscope information], [Three-dimensional magnetic information]}

[0083] {[64.5999984741211, -40.290000915527344, 39.88999938964844],

[0084] [7.850800037384033,5.0976996421813965,-7.277699947357178],

[0085] [6.290299892425537,4.865699768066406,5.738100051879883],

[0086] [13.875222206115723,8.920432090759277,-1.7070082426071167],

[0087] [136.42800903320312,-692.9299926757812,-198.8300018310547],

[0088] [-39.625,3.0,1.6875]}。

[0089] The above information is for one data collection time point. Once the sensor information changes, the current sensor information will be collected immediately.

[0090] The start time of mobile terminal information collection occurs when the input box of the APP gains focus, and the collection ends when the input box loses focus. One sample data corresponds to multiple collection times (when the sensor changes). One sample data can contain information collected up to more than 100 times. The sample data format is as follows:

[0091] Label: [Collection Information 1, Collection Information 2,... Collection Information 100,...]

[0092] The label value of the data manually input is 0, and the label value of the robotic arm is 1.

[0093] Step 2: Normalization processing.

[0094] Eliminate the sample data with the number of collection information less than 30 or greater than 100. At the same time, since the size standards represented by the values of each sensor are different, in order to correctly calculate the distance between samples using the Dynamic Time Warping (DTW) method, it is necessary to normalize all dimensions of the sample data. Apply the min-max normalization method, and the formula is as follows:

[0095]

[0096] Among them, x represents the current sensor information, x min represents the data with the smallest value among the data of the same sensor type as x in the sample data where x is located, x maxRepresents the data with the largest median value among the data in the sample data where x is located and having the same sensor type as the x sensor.

[0097] For example: The data sample after normalization is as follows: [{[0.196, 0.023, 0.13], [0.62, 0.32, 0.014],..., [0.013, 3.0, 0.231]},..., {[0.23, 0.0341, 0.12],..., [0.013, 3.0, 0.231]}]. All the normalized sample data values are between [0 - 1].

[0098] Step 3: Model training.

[0099] All the converted data samples are input into the model for training according to the categorical variables. Applying the K - NN algorithm based on DTW, the essence of model training is to calculate the similarity based on DTW, and use cross - validation to select the appropriate K value to minimize the error rate. The similarity calculation formula between two samples is as follows:

[0100] r i,j = e(i, j)+min{r(i - 1, j - 1), r(i - 1, j), r(i, j - 1)}

[0101] Among them, r(i, j) represents the warping similarity between the first i values of the X sample and the first j values of the Y sample, and e(i, j) represents the Euclidean distance between the i - th sequence point of the sample and the j - th sequence point of the sample. The improved calculation formula is as follows:

[0102]

[0103] Train and select the appropriate K value through cross - validation as the final model. Split out part of the training data and validation data in a ratio of 6:4. The value of K starts from 3 and gradually increases. When the error rate no longer decreases, take the current K value as the final model.

[0104] Step 4: Model application.

[0105] Convert the data collected when the user inputs into a normalized data sequence, and input it into the model for classification and recognition.

[0106] From the software level, in order to improve the accuracy and efficiency of human - machine recognition, this application provides an embodiment of a human - machine recognition device for implementing all or part of the content in the human - machine recognition method. See Figure 5 , the human - machine recognition device specifically includes the following content:

[0107] Acquisition module 01, used to acquire the sensor information group of the target terminal device.

[0108] An identification module 02, configured to obtain a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model; the preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm with multiple historical sensor information groups and their respective corresponding actual human-machine recognition results.

[0109] In an embodiment of the present application, the human-machine recognition device further includes:

[0110] A sample acquisition module, configured to acquire a sample data set, and each piece of sample data in the sample data set includes: a unique historical sensor information group and its corresponding actual human-machine recognition result, and the actual human-machine recognition result is: manual operation or machine operation.

[0111] A training module, configured to train a k-nearest neighbor classification model based on the dynamic time warping algorithm with the sample data set to obtain the human-machine classification model.

[0112] The embodiments of the human-machine recognition device provided in this specification can specifically be used to execute the processing procedures of the embodiments of the above human-machine recognition method, and its functions will not be elaborated here, and reference can be made to the detailed description of the embodiments of the above human-machine recognition method.

[0113] To further illustrate the solution, see Figure 6 , the present application further provides an application example of a human-machine recognition device. A data sending unit 10 is configured to collect sensor data and convert the data into a time series representation; a data conversion unit 20 is configured to perform normalization processing on data in each dimension; a model calculation unit 30 is configured to calculate a prediction result; a result receiving unit 40 is an application end that calls the model service; wherein, the function implemented by the data sending unit 10 can be equivalent to the combined function of the above acquisition module and sample acquisition module; the function implemented by the model calculation unit 30 can be equivalent to the functions implemented by the above identification module and training module.

[0114] As can be seen from the above description, the human-machine recognition method and device provided by the present application can improve the accuracy and efficiency of human-machine recognition; without the subjective participation of users, the user experience is improved, and the defect that the verification code technology fails can be solved; it can effectively prevent fraud and avoid losses such as enterprise property.

[0115] From the hardware level, in order to improve the accuracy and efficiency of human-machine recognition, the present application provides an embodiment of an electronic device for implementing all or part of the content in the above human-machine recognition method. The electronic device specifically includes the following content:

[0116] A processor, a memory, a communications interface, and a bus; wherein, the processor, the memory, and the communications interface complete communication with each other through the bus; the communications interface is used to implement information transmission between the human-machine recognition device and related devices such as user terminals, etc. This electronic device can be a desktop computer, a tablet computer, a mobile terminal, etc., and this embodiment is not limited thereto. In this embodiment, this electronic device can be implemented with reference to the embodiments for implementing the human-machine recognition method and the embodiments for implementing the human-machine recognition device, and the content is incorporated herein, and repeated parts will not be elaborated.

[0117] Figure 7 It is a schematic block diagram of the system composition of the electronic device 9600 according to an embodiment of the present application. As Figure 7 shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It should be noted that this Figure 7 is exemplary; other types of structures can also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0118] In one or more embodiments of the present application, the human-machine recognition function can be integrated into the central processing unit 9100. Among them, the central processing unit 9100 can be configured to perform the following controls:

[0119] Step 100: Obtain a sensor information group of a target terminal device.

[0120] Step 200: Obtain a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model.

[0121] Step 300: The preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm with multiple historical sensor information groups and their respective corresponding actual human-machine recognition results.

[0122] From the above description, it can be seen that the electronic device provided by the embodiments of the present application can improve the accuracy and efficiency of human-machine recognition.

[0123] In another implementation manner, the human-machine recognition device can be separately configured from the central processing unit 9100. For example, the human-machine recognition device can be configured as a chip connected to the central processing unit 9100, and the human-machine recognition function is implemented through the control of the central processing unit.

[0124] As Figure 7As shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It should be noted that the electronic device 9600 does not necessarily have to include Figure 7 all the components shown in Figure 7 ; in addition, the electronic device 9600 may further include

[0125] As Figure 7 shown, the central processing unit 9100, sometimes also referred to as a controller or an operation control, may include a microprocessor or other processor devices and / or logic devices. The central processing unit 9100 receives inputs and controls the operations of the various components of the electronic device 9600.

[0126] Among them, the memory 9140, for example, may be one or more of a buffer, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory, or other suitable devices. The above information related to failures can be stored, and in addition, programs for executing relevant information can also be stored. And the central processing unit 9100 can execute the programs stored in the memory 9140 to implement information storage or processing, etc.

[0127] The input unit 9120 provides inputs to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to supply power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0128] The memory 9140 may be a solid-state memory. For example, it may be a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be such a memory that stores information even when powered off, can be selectively erased, and has more data. Examples of such a memory are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or the processes for operating the electronic device 9600 through the central processing unit 9100.

[0129] The memory 9140 may further include a data storage unit 9143 for storing data such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers of the electronic device for communication functions and / or for performing other functions of the electronic device (such as a messaging application, an address book application, etc.).

[0130] The communication module 9110 is a transmitter / receiver 9110 that transmits and receives signals via the antenna 9111. The communication module (transmitter / receiver) 9110 is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as in the case of a conventional mobile communication terminal.

[0131] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module, and / or a wireless local area network module, etc. The communication module (transmitter / receiver) 9110 is also coupled to the speaker 9131 and the microphone 9132 via the audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, so as to implement normal telecommunication functions. The audio processor 9130 may include any suitable buffers, decoders, amplifiers, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0132] As can be seen from the above description, the electronic device provided by the embodiments of the present application can improve the accuracy and efficiency of human-machine recognition.

[0133] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all steps in the human-machine recognition method in the above embodiments. A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, all steps in the human-machine recognition method in the above embodiments are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0134] Step 100: Obtain a sensor information group of a target terminal device.

[0135] Step 200: Obtain a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model.

[0136] Step 300: The preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm using multiple historical sensor information groups and their respective corresponding actual human-machine recognition results.

[0137] As can be seen from the above description, the computer-readable storage medium provided by the embodiments of the present application can improve the accuracy and efficiency of human-machine recognition.

[0138] The various embodiments of the above methods in the present application are all described in a progressive manner. For the same or similar parts among the various embodiments, reference can be made to each other, and the key points of each embodiment are the differences from other embodiments. For the relevant parts, reference can be made to the partial description of the method embodiments.

[0139] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0140] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0141] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0142] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Therefore, the instructions executed on the computer or other programmable device provide means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1Steps of the functions specified in one or more boxes.

[0143] In this application, specific embodiments are used to elaborate on the principles and implementation manners of this application. The descriptions of the above embodiments are only for helping to understand the method of this application and its core idea. At the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A human - machine recognition method, characterized in that, it includes: Obtain a sensor information group of the target terminal device; According to the sensor information group and a preset human - machine classification model, obtain the human - machine recognition result of the target operation at the target terminal device, and the human - machine recognition result is: manual operation or machine operation, and the machine operation includes: robotic arm operation; The preset human - machine classification model is obtained by training a k - nearest neighbor classification model based on the dynamic time warping algorithm using multiple historical sensor information groups and their respective corresponding actual human - machine recognition results; The obtaining the sensor information group of the target terminal device includes: Obtain sensor information collected multiple times within the acquisition time range, and all the sensor information constitutes the sensor information group. The acquisition time range includes: the start time to the end time of the target operation. Each sensor information includes: direction sensor, linear acceleration sensor, gravity sensor, acceleration sensor, gyroscope, and magnetic sensor information; if the target operation is to distribute coupons according to the mobile phone number input by the user, then determine the time point when the focus is obtained in the APP input box as the start time, and determine the time point when the focus is lost in the APP input box as the end time; Before obtaining the sensor information group of the target terminal device, it further includes: Obtain a sample data set, and each piece of sample data in this sample data set includes: a unique historical sensor information group and its corresponding actual human - machine recognition result, and the actual human - machine recognition result is: manual operation or machine operation; Use the sample data set to train a k - nearest neighbor classification model based on the dynamic time warping algorithm to obtain the human - machine classification model; Each group of historical sensor information groups includes: historical sensor information collected multiple times corresponding to one historical operation; Correspondingly, the obtaining the sample data set includes: Obtain a sample data set; Delete the sample data in the sample data set whose number of acquisitions does not meet the preset acquisition number range; Perform normalization processing on the historical sensor information groups in the sample data set, including: obtaining the result after normalization processing using the following formula: Among them, x represents the current sensor information, x min represents the data with the smallest median value among the data of the same sensor type as x in the sample data where x is located, x max represents the data with the largest median value among the data of the same sensor type as x in the sample data where x is located; The obtaining the sensor information collected multiple times within the acquisition time range includes: Within the acquisition time range, whenever the sensor information changes, collect the current sensor information; The using the sample data set to train a k - nearest neighbor classification model based on the dynamic time warping algorithm to obtain the human - machine classification model includes: Calculate the similarity based on the dynamic time warping algorithm, use cross - validation to determine the value of K, split the training data and validation data in a ratio of 6:4, start taking the value of K from 3 and gradually increase it. When the error rate no longer decreases, take the current value of K as the final value of K.

2. A human - machine recognition device, characterized in that, it includes: An obtaining module, configured to obtain a sensor information group of the target terminal device; An identification module, configured to obtain a human-machine recognition result of a target operation at the target terminal device according to the sensor information group and a preset human-machine classification model, where the human-machine recognition result is: manual operation or machine operation, and the machine operation includes: robotic arm operation; The preset human-machine classification model is obtained by training a k-nearest neighbor classification model based on the dynamic time warping algorithm using multiple historical sensor information groups and their respective corresponding actual human-machine recognition results; The obtaining module is specifically configured to: obtain sensor information collected multiple times within a collection time range, and all the sensor information forms the sensor information group, where the collection time range includes: the start time to the end time of the target operation, and each sensor information includes: direction sensor, linear acceleration sensor, gravity sensor, acceleration sensor, gyroscope, and magnetic sensor information; if the target operation is to distribute coupons according to the mobile phone number input by the user, the time point when the focus is obtained in the APP input box is determined as the start time, and the time point when the focus is lost in the APP input box is determined as the end time; A sample obtaining module, configured to obtain a sample data set, where each piece of sample data in the sample data set includes: a unique historical sensor information group and its corresponding actual human-machine recognition result, and the actual human-machine recognition result is: manual operation or machine operation; A training module, configured to train a k-nearest neighbor classification model based on the dynamic time warping algorithm using the sample data set to obtain the human-machine classification model; Each group of historical sensor information groups includes: historical sensor information collected multiple times corresponding to one historical operation; Correspondingly, the obtaining of the sample data set includes: Obtain a sample data set; Delete the sample data in the sample data set whose number of collections does not meet the preset collection number range; Perform normalization processing on the historical sensor information groups in the sample data set, including: obtaining the result after normalization processing using the following formula: where x represents the current sensor information, x min represents the data with the smallest median among the data of the same sensor type as x in the sample data where x is located, x max represents the data with the largest median among the data of the same sensor type as x in the sample data where x is located; The obtaining of the sensor information collected multiple times within the collection time range includes: Within the collection time range, whenever the sensor information changes, collect the current sensor information; The training of the k-nearest neighbor classification model based on the dynamic time warping algorithm using the sample data set to obtain the human-machine classification model includes: Calculate the similarity based on the dynamic time warping algorithm, determine the value of K using the cross-validation method, split the training data and the validation data according to 6:4, start taking the value of K from 3 and gradually increase it. When the error rate no longer decreases, take the current value of K as the final value of K.

3. An electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that when the processor executes the program, it implements the human-machine recognition method according to claim 1.

4. A computer-readable storage medium, on which computer instructions are stored, characterized in that when the instructions are executed, they implement the human-machine recognition method according to claim 1.

Images