KR20200107356A - User-customized motion recognition method using a mobile device - Google Patents

Abstract

A method of recognizing user motion using a mobile device including a sensor unit, the method comprising: acquiring motion sensor data according to movement of the mobile device using the sensor unit; Extracting a motion trajectory image of the mobile device from the obtained motion sensor data; Determining whether or not the predetermined model matches the extracted motion trajectory image; Downloading a software development tool (SDK) when the predetermined model and the extracted motion trajectory data match; And if the predetermined model and the extracted motion trajectory data do not match, collect motion data corresponding to the motion trajectory data, apply a machine learning algorithm to the collected motion data to generate a learning result, and then download the SDK. Thus, a user-customized motion recognition method comprising the step of reflecting the learning result in the SDK is disclosed.

Description

User-customized motion recognition method using a mobile device {USER-CUSTOMIZED MOTION RECOGNITION METHOD USING A MOBILE DEVICE}

The present invention relates to a user-customized motion recognition method and a mobile device for user-customized motion recognition.

With the recent rapid development of mobile terminal technology, mobile terminals are increasingly being used not only for the purpose of making phone calls, but also for internet games, watching satellite broadcasts, remote control using short-range communication, and image capture by equipped digital cameras. .

For example, a user may perform various functions of a mobile terminal (eg, memo, drawing, application execution, etc.) by using gesture recognition using a mobile terminal. Alternatively, an application (ex. game, SNS, education, music, memo, etc.) may be installed on a mobile terminal to utilize the application through gesture recognition using the mobile terminal.

Regarding the motion recognition technology using such a mobile terminal, Korean Patent Laid-Open No. 2015-0118639 detects motions made within an application through a sensor of the terminal, and recognizes motion sensor values related to motion as a specific motion by dividing it into time series. Thus, by providing a dynamic interface, a sensor value processing method and apparatus for motion recognition capable of processing motion sensor values generated in a plurality of sensing devices are disclosed.

However, according to the prior art as described above, a method of completing the same motion may be different for each user using the motion recognition function of a mobile device, but there is a disadvantage in that an accurate result reflecting this point cannot be derived. For example, when a user performs a circle motion using a mobile device, some users start from the left and perform a circle motion clockwise, while some users start from the right and perform a circle motion counterclockwise. can do. In this case, if the motion of the user is predicted using time series data according to the prior art, it is difficult to extract the same motion data result, and therefore, there is a problem in that the motion recognition rate of the mobile device user is lowered. In addition, if the user's motion is predicted using time series data, the motion data may be contaminated by various user variables (e.g., too fast input or too slow input, etc.) that may occur until the motion is completed. There is also a problem.

Korean Patent Publication No. 2015-0118639

An object of the present invention is to provide a motion recognition method capable of improving the accuracy of motion recognition by solving the shortcomings of user motion recognition technology using a conventional mobile device.

In order to solve this problem, the motion recognition method provided by the present invention is a method of recognizing user motion using a mobile device including a sensor unit, and obtaining motion sensor data according to the movement of the mobile device using the sensor unit. Step to do; Extracting a motion trajectory image of the mobile device from the obtained motion sensor data; Determining whether a predetermined model matches the extracted motion trajectory image; Downloading a software development tool (SDK) when the predetermined model matches the extracted motion trajectory data; And when the predetermined model and the extracted motion trajectory data do not match, collecting motion data corresponding to the motion trajectory data and applying a machine learning algorithm to the collected motion data to generate a learning result, And downloading the SDK) and reflecting the learning result in the SDK.

In one embodiment, the sensor unit, an acceleration sensor, a magnetic sensor, a gravity sensor (G-sensor), a group including a gyroscope sensor (gyroscope sensor) and a motion sensor (motion sensor) It may be characterized in that at least one sensor selected from.

In one embodiment, the sensor unit is a gyro sensor unit, the motion sensor data obtained is a rotation angle shift value in the X-axis, Y-axis, and Z-axis directions, and the motion of the mobile device from the obtained motion sensor data In the step of extracting the trajectory image of, it may be characterized in that a two-dimensional motion trajectory image is extracted based on the rotation angle shift values in the X-axis, Y-axis, and Z-axis directions.

In one embodiment, between the step of obtaining the motion sensor data and the step of extracting the motion trajectory image, further comprising pre-processing the obtained motion sensor data, wherein the predetermined model and the extracted motion If the trajectory data of do not match, prior to applying the machine learning algorithm, giving a label for each motion data, and before generating the learning result, the motion data collected for the machine learning It may be characterized in that it further comprises a step of amplifying.

A mobile device for user-customized motion recognition according to another embodiment of the present invention includes: a processor including a motion recognition unit and a motion data collection/learning unit; A sensor unit acquiring motion data of the mobile device; And a memory for storing a program for operation and control of the processor and user-specified motion model data, wherein the motion data obtained by the sensor unit includes at least one of coordinate change data and rotation angle data for each axis, , The motion recognition unit extracts the motion data acquired by the sensor unit as a trajectory image and compares it with a user-designated motion model previously stored in the memory to determine whether the data are matched, and when it is determined that the data are not matched, the motion data collection/learning When it is determined that the data are matched, the control unit controls to download the software development tool, and the motion data collection/learning unit receives the motion data collection request from the motion recognition unit, and A learning result is generated by collecting motion data corresponding to a trajectory image, applying a machine learning algorithm to the collected motion data, and reflecting the learning result on the downloaded software development tool.

According to the motion recognition method of the present invention, it is possible to improve the accuracy of a user's motion recognition rate using a mobile device.

In addition, in the user motion recognition technology using a mobile device, the pollution degree of motion data can be reduced, and motion recognition results can be accurately extracted for various users.

In addition, according to the present invention, not only the motion corresponding to the predefined motion model is recognized, but a user's own unique motion can be created and used, and various user groups can also create a group-specific motion to be unique. Since one motion can be created and used, there is an advantage in that the utilization of the motion recognition service can be increased.

1 is a block diagram of a mobile device according to an embodiment of the present invention.
2 is a configuration diagram of a sensor unit according to an embodiment of the present invention.
3 is a flow chart of a user-customized motion recognition method according to an embodiment of the present invention.
4 is a diagram illustrating a gyro sensor according to an embodiment of the present invention.
5 and 6 are diagrams for explaining a method of converting acquired motion data into an image according to an embodiment of the present invention.
7 is a graph of motion data acquired according to an embodiment of the present invention.
8 is an extracted image corresponding to the data of FIG. 7.

Hereinafter, embodiments disclosed in the present specification will be described with reference to the accompanying drawings.

In particular, even if there is no reference numeral, elements that can be inferred from the description in the specification may be interpreted according to the content described literally.

Even a word used at the level of a person skilled in the art may not accurately reflect an actual function and role, and thus may be used incorrectly. In this case, it is in principle to be interpreted according to the intention of the inventor by analogy based on the unified contents in the specification, and it is not necessary to be interpreted only limited to the limitation of the text itself, and according to the interpretation of the name or function, etc. Should be judged more flexibly by referring to the overall contents.

The embodiments disclosed in the present specification will be described in detail with reference to the accompanying drawings, but identical or similar components are denoted by the same reference numerals regardless of the reference numerals, and redundant descriptions thereof will be omitted.

The suffixes "unit" and "module" for the constituent elements used in the following description are given or used interchangeably in consideration of only the ease of preparation of the specification, and do not themselves have distinct meanings or roles. In addition, in describing the embodiments disclosed in the present specification, when it is determined that a detailed description of related known technologies may obscure the subject matter of the embodiments disclosed in the present specification, the detailed description thereof will be omitted. In addition, the accompanying drawings are for easy understanding of the embodiments disclosed in the present specification, and the technical idea disclosed in the present specification is not limited by the accompanying drawings, and all modifications included in the spirit and scope of the present invention It should be understood to include equivalents or substitutes.

When a component is referred to as being "connected" or "connected" to another component, it is understood that it may be directly connected or connected to the other component, but other components may exist in the middle. Should be. On the other hand, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there is no other component in the middle.

Singular expressions include plural expressions unless the context clearly indicates otherwise. In addition, throughout the specification of the present application, when a certain part "includes" a certain component, it means that other components may be further included rather than excluding other components unless otherwise stated.

In the present invention, the mobile device can be implemented in various forms. For example, mobile devices described herein include mobile phones, smart phones, notebook computers, terminals for digital broadcasting, personal digital assistants (PDAs), portable multimedia players (PMPs), navigation, etc. It includes a terminal, a smart TV remote control, a game controller, and a motion stick (stick type).

It is apparent to those skilled in the art that the present invention may be embodied in other specific forms without departing from the essential features of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the drawings. In the drawings, the same reference numerals refer to the same elements.

1 is a block diagram of a mobile device 100 for recognizing user motion according to an embodiment of the present invention. Referring to FIG. 1, the mobile device 100 may include a processor 110, a sensor unit 120, a memory unit 130, a display unit 140, an input unit 150, and a communication unit 160. However, FIG. 1 is one of various embodiments of the present invention, and a mobile device may be implemented by more components than those shown in FIG. 1, and a mobile device may be implemented by fewer components. Hereinafter, the above components will be described.

The mobile device 100 may include at least one processor 110, and the processor 110 may include a central processing electronic unit (CPU), an application processor (AP), or a communication processor (CP). The processor 110 may include a motion recognition unit 112 and a motion data collection/learning unit 114. For example, the processor 110 detects the movement of the mobile device 100 through the sensor unit 120, or when a motion recognition event predefined by the user occurs, the motion recognition unit 112 is set to the active mode. Can be controlled to switch. The motion recognition unit 112 receives the motion data acquired by the sensor unit 120 in the active mode and compares the user-designated motion model previously stored in the memory 130 with the motion data acquired from the sensor unit 120 Can recognize the motion of For example, the motion recognition unit 112 may recognize a user's motion by comparing the motion data obtained from the sensor unit 120 with a trajectory or image of a user-designated motion model stored in the memory 130. Here, the motion data acquired through the sensor unit 120 may include data based on at least one of a coordinate change (translational motion) and a rotation angle change (rotary motion for each axis) of the mobile device 100. That is, the motion data may include coordinate change data and/or rotation angle data for each axis. The motion recognition unit 112 compares the motion data detected through the sensor unit 120 with the motion model previously stored in the memory 130, so that a data error (for example, a coordinate change or a rotation angle error for each axis) is When it is less than a predetermined threshold value determined in advance, it may be determined that both motion data match. Conversely, when a data error as a result of the comparison exceeds a predetermined threshold value, it may be determined that both motion data do not match. When it is determined by the motion recognition unit 112 that both motion data match, the user motion recognition is determined to be successful, and a software development kit (SDK) can be downloaded. In another embodiment, whether or not the motion recognition is successful may be output to the display unit 140 or a result of the successful motion recognition may be transmitted to an external device or a network through the communication unit 160. On the other hand, if it is determined by the motion recognition unit 112 that both motion data do not match, the processor 110 may collect the data and control the machine learning algorithm to be applied. That is, if it is determined that the data do not match, the motion recognition unit 112 transmits a motion data collection request to the motion data collection/learning unit 114. The motion data collection/learning unit 114 receives a motion data collection request from the motion recognition unit 112, collects motion data corresponding to the motion trajectory data, and learns by applying a machine learning algorithm to the collected motion data. Generate the result. A detailed operation of the motion data collection/learning unit 114 will be described later.

In addition, the mobile device 100 may include at least one sensor unit 120. The sensor unit 120 is at least selected from a group including an acceleration sensor, a magnetic sensor, a gravity sensor, a gyroscope sensor, and a motion sensor. It can be one sensor. Of course, the mobile device disclosed in this specification may combine and utilize information sensed by at least two or more of these sensors. According to an embodiment, the sensor unit 120 is preferably a motion sensor unit capable of sensing motion of the mobile device 100. Referring to FIG. 2, the sensor unit 120 includes a gyro sensor 122 that senses a rotation angle of an object. According to another embodiment, the sensor unit 120 multiplies the step length of a user holding the mobile device and/or the geomagnetic sensor 124 that calculates the movement direction when the mobile device moves to multiply the movement distance by a step (one step). It may further include an acceleration sensor 126 to calculate. As is well known, the acceleration sensor 126 is a sensor that measures acceleration around a ground surface. By using the gyro sensor 122 and the acceleration sensor 126 together, the motion of an object can be more accurately measured. For example, if the object to be measured moves in two dimensions, one gyro sensor and two acceleration sensors can be used in combination, and if the object to be measured moves in three dimensions, three gyro sensors and three acceleration sensors are used. Can be used in combination. According to an embodiment, a three-dimensional motion of an object may be measured by combining six sensors having 6 degrees of freedom. In addition, according to a characteristic of a gyro sensor whose characteristic changes according to a temperature change, the sensor unit 120 may further include a temperature sensor (not shown) for compensating for an error according to temperature.

In addition, the mobile device 100 may include at least one memory unit 130. The memory unit 130 may store programs for operation and control of the processor 110, and input/output data. It can also perform a function for storing (eg, messages, images, videos, etc.). In addition, a motion model preset by a user may be stored in the memory 130.

Further, the mobile device 100 may include at least one display unit 140. The display unit 140 outputs information processed by the mobile device 100. For example, when the mobile device is in a call mode, a user interface (UI) or a graphical user interface (GUI) related to a call is displayed. In addition, when the mobile device 100 is in a video call mode or a photographing mode, a photographed or/and received image, UI, or GUI is displayed. Meanwhile, the display unit 140 and the touch pad may form a mutual layer structure to form a touch screen.

Further, the mobile device 100 may include at least one input unit 150. The input unit 150 may include a camera unit for inputting a video signal, a microphone unit for inputting an audio signal, and/or a user input unit for a user inputting a motion control of a terminal.

In addition, the mobile device 100 may include at least one communication unit 160. The communication unit 160 may include one or more components for performing wireless communication between the mobile device 100 and a communication system or wireless communication between the mobile device 100 and a network in which the mobile device 100 is located. For example, the communication unit 160 may include a mobile communication unit, a wireless Internet unit, a short-range communication unit, a location information unit, and the like.

With motion recognition Motion data collection/ Learning Department

As described above, the motion recognition unit 112 may determine whether to match the motion data obtained from the sensor unit 120 with a user-specified motion model stored in the memory 130. In addition, the motion recognition unit 112 may control the motion data collection/learning unit 114 to perform a machine learning algorithm when the sensor data and the motion model stored in the memory are not matched as a result of comparison. In addition, according to an embodiment, the motion recognition unit 412 may control the memory 130 to store motion data corresponding to the motion model. Here, the motion model for the user's motion may consist of at least one motion data. For example, the motion data may include at least one of rotation angle, acceleration, or rotation intensity data for one period for a specific motion of the user. In addition, according to an embodiment, when the motion recognition unit 412 recognizes the user's motion, the motion recognition unit 412 may control the display unit 140 to display user motion information.

In addition, as described above, the motion data collection/learning unit 114 collects motion data corresponding to the motion data detected by the sensor unit 120, and learns by applying a machine learning algorithm to the collected motion data. By doing so, the learning outcome is created. The motion data collection/learning unit 114 may start motion data collection/learning when receiving a motion data collection request from the motion recognition unit 112. To this end, the motion data collection/learning unit 114 includes a learning memory unit (not shown) configured to receive, classify, store, and/or output information used for data mining, data analysis, and machine learning algorithm technology. can do. The learning memory unit may be configured to store reception, detection, generation, and predefined information through the mobile device. The learning memory unit may be configured to store the data in one or more databases in order to identify, colorize, classify, manipulate, store, and/or retrieve data for use in learning, data mining, predictive analysis, or other machine learning techniques. have. The information stored in the learning memory unit may be used by a controller using at least one of different types of data analysis, machine learning algorithms, and machine learning technologies. Examples of such algorithms are K-Nearest neighbor systems, fuzzy logic (e.g., possibility theory), neural networks, and Boltzmann machines. , Vector quantization, pulsed neural nets, support vector machines, maximum margin classifiers, hill-climbing, inductive logic systems, Bayesian networks (baysian networks), petri nets (e.g. finite state machines, mealy machines, moore finite state machines), classifier trees (E.g. perceptron trees, support vector trees, markov trees, decision tree forests, random forests), horse warfare model And systems (pandemonium models and systems), clustering, artificially intelligent planning, artificially intelligent forecasting, data fusion, sensor fusion, image fusion fusion, reinforcement learning, augmented reality, pattern recognition, and automated planning.

According to another embodiment, the machine learning algorithm may include an ensemble structure algorithm. For example, as a plurality of machine learning algorithms constituting the ensemble structure, for example, a decision tree (DT) classification algorithm, a random forest (RF) classification algorithm, and a support vector machine (SVM) classification algorithm It may include. The decision tree classification algorithm is a method of learning in a tree structure to derive results, so that result interpretation and understanding are easy, data processing speed is fast, and rules can be derived based on a search tree. In addition to DT, RF can be applied as a way to improve the low classification accuracy of the decision tree algorithm. RF is a method of deriving the result of learning a plurality of DTs in an ensemble, and has the advantage that it is more difficult to understand the result than DT, but the result accuracy may be higher than DT. In addition, SVM can be applied as a way to improve overfitting that can occur through DT or RF learning. The SVM classification algorithm classifies data belonging to different classifications on a plane basis, and generally has high accuracy and may structurally have low sensitivity to overfitting. In the above example, an ensemble structure using three machine learning algorithms has been described, but the number or type of machine learning algorithms is not necessarily limited thereto, and the number and/or type of machine learning algorithms may be changed as necessary. Based on the ensemble structure configured in this way, by learning using the collected motion data sample information, a user-customized motion recognition method can be provided by reflecting the learning result.

Customized motion Recognition method

Referring to Figure 3, a description will be given of a user-customized motion recognition method using a mobile device according to an embodiment of the present invention. First, motion sensor data is acquired by sensing a user motion using a mobile device including a motion sensor unit (S100). As described above, it is preferable that the motion sensor unit includes a gyro sensor. In addition, the sensor unit may further include at least one sensor selected from a group including an acceleration sensor, a magnetic sensor, a gravity sensor, and a motion sensor. The motion sensor unit senses the user's motion and outputs the acquired sensor data to the motion recognition unit. The sensor data acquired by the motion sensor unit may include data based on at least one of a coordinate change (translational motion) and a rotation angle change (rotational motion for each axis) of the mobile device. The motion recognition unit receives sensor data from the motion sensor unit and performs pre-processing (S110), and then extracts the motion trajectory data as an image based on the sensor data (S120). The motion recognition unit compares the extracted motion trajectory data with a preset motion model stored in the memory to determine whether to match (S130). It may be determined whether or not the data match, and when the error of both data is less than or equal to a predetermined threshold value determined in advance, it may be determined that both data match. Conversely, when an error of both data exceeds a predetermined threshold as a result of the comparison, it may be determined that both data are not matched. If the comparison result (S130) is determined to match the data (YES), the procedure can be terminated by downloading (S170) an SDK (Software Download Kit). On the other hand, if it is determined that the data are not matched in the comparison result (S130) (NO), motion data collection/learning using machine learning may be performed. As described above, if it is determined that the data are not matched (NO) as a result of the comparison (S130), the motion recognition unit transmits a motion data collection request to the motion data collection/learning unit, and the motion data collection/learning unit, the motion trajectory data. It collects motion data corresponding to and applies a machine learning algorithm to the collected motion data. In order to apply the machine learning algorithm, the motion data collection/learning unit first performs a step of labeling the motion data (S-140). Users can designate a label for each learning data. After labeling the motion data, motion data corresponding to the motion trajectory data is collected (S-150). When motion data is collected, a learning result is generated by learning by applying a machine learning algorithm to the collected data (S-160). Meanwhile, prior to the motion data learning step (S-160), in order to improve accuracy of machine learning learning, the step of amplifying the collected motion data may be further included. When the machine learning application result and the learning result are generated, various user-customized motion recognition can be performed by downloading the SDK and applying the learning result to it.

Gyro Using a sensor Motion How to convert trajectory data into an image

Hereinafter, a method of extracting motion data acquired using a gyro sensor as a trajectory image according to a preferred embodiment of the present invention will be described. Prior to this, referring to FIG. 4, each axis (X axis, Y axis, Z axis) of the gyro sensor is shown. The gyro sensor, also called an angular velocity meter, measures the inclination based on information obtained by measuring the angular velocity. The gyro sensor in the present invention includes a tuning fork type MEMS sensor. Referring to FIG. 4, the gyro sensor calculates roll, pitch, and yaw values corresponding to the X, Y, and Z axes, respectively, for the X, Y, and Z axis directions. Each detects the rotation angle (angular velocity).

Next, a method of converting motion data into a trajectory image will be described with reference to FIGS. 5 and 6. According to a preferred embodiment of the present invention, motion data detected through the sensor unit may be extracted as an image. According to an embodiment, an image may be extracted using dominant data among detected motion data. For example, if the sensor module is a gyro sensor, and the motion data detected through the sensor module is rotation angle data for each axis (X-axis, Y-axis, Z-axis), two dominant data among the three detection data are And extracts a 2D image using the determined data. The definition of'dominant data' can be variously defined by the user. For example, determining two dominant data out of three data means that one of the plurality of acquired data It may also be described in the same sense as excluding data. In relation to this, an example in which the user performs the Korean consonant'a' in motion using a mobile device will be described with reference to FIG. 5. When drawing the first stroke of'a' using the mobile device, the user can rotate the mobile device from left to right by rotating the wrist, and looking down from above, as shown in FIG. 5(a). In this case, the rotation angle of the X-axis of the gyro sensor increases. 5(b) shows a state in which the user has completed drawing the first stroke of'A' using the mobile device. Next, when the user draws the second stroke of'a' using the mobile device, the user can rotate the mobile device from the top to the bottom by lowering the mobile device or rotating the wrist. It is the same as Fig. 6(a). In this case, the Z-axis rotation angle of the gyro sensor decreases. 6(b) shows a state in which the user completes drawing the second stroke of'a' using the mobile device. 7 is a graph showing sensor data acquired when the'a' motion is recognized as illustrated in FIGS. 5 and 6. Referring to FIG. 7, a green graph indicates X-axis rotation angle data, a blue graph indicates Y-axis rotation angle data, and a red graph indicates Z-axis rotation angle data. Among the three acquired data, the dominant data is X It can be axis and Z axis rotation angle data. As in the above-described example, the dominant data may be determined by excluding data having the least amount of change among acquired data. Referring to FIG. 7, (a) is a section in which the user's wrist holding the mobile device naturally rotates from right to left before the user writes, and (b) is the point at which the user starts writing. , From left to right, the gyro sensor detection value continues to increase and then stops. According to the above example, it corresponds to the section where the first stroke of'a' is drawn. (c) The section is a section where the gyro sensor detection value decreases from top to bottom and then stops. According to the above example, it corresponds to the section where the second stroke of'a' is drawn. (d) The section corresponds to the state in which the user has stopped writing. Dominant data is determined among the sensor data thus obtained, and a two-dimensional image ('a') may be extracted using the determined data. The extracted image is shown in FIG. 8.

The present invention described above can be implemented as a computer-readable code in a medium on which a program is recorded. The computer-readable medium includes all types of recording devices storing data that can be read by a computer system. Examples of computer-readable media include HDD (Hard Disk Drive), SSD (Solid State Disk), SDD (Silicon Disk Drive), ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. There is also a carrier wave (eg, transmission over the Internet). In addition, the computer may include a control unit of the terminal. Therefore, the detailed description above should not be construed as restrictive in all respects and should be considered as illustrative. The scope of the present invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention.

100: mobile device
110: processor
112: motion recognition unit
114: Motion collection/learning department
120: sensor unit
122: gyro sensor
124: geomagnetic sensor
126: acceleration sensor
130: memory unit
140: display unit
150: input
160: communication department

Claims (5)

A method of recognizing user motion using a mobile device including a sensor unit,
Acquiring motion sensor data according to the movement of the mobile device using the sensor unit;
Extracting a motion trajectory image of the mobile device from the obtained motion sensor data;
Determining whether a predetermined model matches the extracted motion trajectory image;
Downloading a software development tool (SDK) when the predetermined model matches the extracted motion trajectory data; And
When the predetermined model and the extracted motion trajectory data do not match, the motion data corresponding to the motion trajectory data is collected and a machine learning algorithm is applied to the collected motion data to generate a learning result, and the SDK) downloading and reflecting the learning result in the SDK.
The method of claim 1, wherein the sensor unit,
Characterized by at least one sensor selected from the group including an acceleration sensor, a magnetic sensor, a gravity sensor (G-sensor), a gyroscope sensor, and a motion sensor. User-customized motion recognition method.
The method of claim 1, wherein the sensor unit is a gyro sensor unit,
The obtained motion sensor data is a rotation angle shift value in the X-axis, Y-axis, and Z-axis directions,
In the step of extracting a motion trajectory image of the mobile device from the obtained motion sensor data, extracting a two-dimensional motion trajectory image based on rotation angle shift values in the X-axis, Y-axis, and Z-axis directions. User-customized motion recognition method characterized by.
The method of claim 1, further comprising pre-processing the obtained motion sensor data between the step of obtaining the motion sensor data and the step of extracting the motion trajectory image,
If the predetermined model and the extracted motion trajectory data do not match, prior to applying the machine learning algorithm, further comprising the step of assigning a label for each motion data,
Before generating the learning result, the method further comprising amplifying the motion data collected for the machine learning.
As a mobile device for user-customized motion recognition,
A processor including a motion recognition unit and a motion data collection/learning unit;
A sensor unit acquiring motion data of the mobile device; And
Including a memory for storing a program for operation and control of the processor and user-specified motion model data,
The motion data acquired by the sensor unit includes at least one of coordinate change data and rotation angle data for each axis,
The motion recognition unit extracts the motion data acquired by the sensor unit as a trajectory image and compares it with a user-designated motion model previously stored in the memory to determine whether the data matched. When it is determined that the motion data collection request is transmitted and the data are matched, the control unit controls to download the software development tool,
The motion data collection/learning unit receives a motion data collection request from the motion recognition unit, collects motion data corresponding to the motion trajectory image, and generates a learning result by applying a machine learning algorithm to the collected motion data. , A mobile device for user-customized motion recognition, characterized in that reflecting the learning result on the downloaded software development tool.

Images