KR102433828B1 - Pain Behavior Monitoring System and Method for Managing Pain using the same - Google Patents

Abstract

The present invention relates to a pain behavior monitoring system capable of collecting and providing objective information on pain behavior using a wearable communication terminal and a pain management method using the same, and is worn on the user's wrist using a 3-axis motion sensor. A wrist worn terminal for recognizing two types of hand gestures and a position of a hand and synthesizing the recognized two pieces of information to recognize pain behavior; a user terminal for communicating with the wrist worn terminal and interworking with a monitoring server; a user and a monitoring server that checks the user's current situation through mutual interworking with the terminal and the medical staff terminal, and identifies and guides objective statistical values for the user's pain pattern.

Description

Pain behavior monitoring system and pain management method using same

The present invention relates to pain management, and more particularly, to a pain behavior monitoring system capable of collecting and providing objective information on pain behavior using a wearable communication terminal, and a pain management method using the same.

Pain caused by cancer, joint disease, degenerative disease, surgery, etc. or various accidents accompanying the extension of human lifespan has become a major problem in the medical community.

Recognition and management of pain, such as a patient's pain level, pain duration, and frequency of pain, have become important issues in the medical community, and it is a very important issue in terms of patient treatment and well-being.

Pain can be seen as a kind of warning to treat when the body is injured or ill. Therefore, the process of objectively and accurately judging the cause of pain is a very important process in treating a diseased body.

Pain in the body is caused by a number of causes. For example, it may be due to a patient-specific physical factor such as a disease possessed by the patient or the patient's constitution, or it may be due to an environmental factor caused by the influence of the surrounding environment.

In the case of patient-specific physical factors, the physician can determine the cause of the pain when treating the patient. However, in the case of environmental factors, since the doctor depends on the patient's memory statement, there is a problem in that the cause of the pain cannot be accurately determined according to the distortion or error of the memorized information.

In particular, there is a problem of misuse and abuse of drugs according to drug prescriptions based on such physical and environmental factors.

In addition, in order to improve the accuracy in estimating the cause of pain according to environmental factors, when a patient's sudden pain occurs, the body part where pain is caused, the degree of pain, and the environmental factors at the time of the pain are recorded separately through means of recording the condition. You can't help but expect the patient to record it.

In the prior art, most of the methods for monitoring pain are also methods using a contact sensor, so it is difficult to use in the physical/rehabilitation treatment process of a patient.

Although a method of calculating a pain index based on skin conductivity has been disclosed, this also requires a patient to wear a contact sensor, which is unnatural and difficult to guarantee accuracy.

On the other hand, in the face recognition field, technologies capable of predicting individual emotions through classification of facial expressions have been developed, but they have a problem in that they do not satisfy performance conditions depending on the environment to be applied to pain monitoring.

As one of the prior art techniques for pain behavior recognition, Rui Qin et al.'s paper "Continuous Pain Related Behavior Recognition from muscle Activity and Body Movements" uses an Animazaoo IGS-190 suit including 18 inertial sensors and an EMG sensor called BTS FREEEMG300. This is a study that senses the movement of the user's body, extracts the sensed data, and uses it for the random forest algorithm.

Finally, this study is a study to recognize pain motions that support the lower back in various postures, and there are limitations in recognizing various pain motions.

In addition, there is a limit when considering the use environment using the EMG sensor with poor wearability.

Therefore, there is a demand for the development of a technology that enables the collection and provision of objective information on pain behavior.

Republic of Korea Patent Publication No. 10-2017-0089727 Republic of Korea Patent Publication No. 10-2013-0016708 Republic of Korea Patent Publication No. 10-2020-0056660

The present invention is to solve the problems of pain management technology of the prior art, and provides a pain behavior monitoring system capable of collecting and providing objective information on pain behavior using a wearable communication terminal, and a pain management method using the same but it has a purpose.

An object of the present invention is to provide a pain behavior monitoring system and a pain management method using the same, which enable even a medical staff to monitor various pain patterns by collecting and providing objective information about pain behavior.

In the present invention, the model learning process that requires relatively many calculations is performed externally, and only the model used for inference is loaded on the device to perform on-device deep learning inference, enabling efficient pain behavior monitoring regardless of storage performance and computational performance An object of the present invention is to provide a pain behavior monitoring system and a pain management method using the same.

The present invention provides a pain behavior monitoring system and a pain management method using the same to check the user's current situation through interworking with a user terminal and a medical staff terminal, and to identify and guide objective statistical values for the user's pain pattern. Its purpose is to provide

The present invention recognizes two types of hand motions and hand positions and recognizes and classifies pain behaviors by synthesizing the recognized two pieces of information, thereby combining and recognizing various pain behaviors, such as chest pain, abdominal pain, back pain, joint pain, muscle pain, etc. An object of the present invention is to provide a pain behavior monitoring system capable of judging symptoms caused by various pain behaviors and a pain management method using the same.

The present invention relates to a pain behavior monitoring system and a pain management method using the same, in which a medical staff terminal can be used directly by medical staff, accessing a monitoring server, performing a search operation, acquiring a treatment record for a user, and inputting a report therefor Its purpose is to provide

Other objects of the present invention are not limited to the above-mentioned objects, and other objects not mentioned will be clearly understood by those skilled in the art from the following description.

The pain behavior monitoring system according to the present invention for achieving the above object is worn on the user's wrist using a three-axis motion sensor, recognizes two types of hand motion and hand position, and synthesizes the recognized two pieces of information for pain. A wrist worn terminal for recognizing a behavior; A user terminal that communicates with the wrist worn terminal and interworks with a monitoring server; Checks the user's current situation through interaction with the user terminal and the medical staff terminal, and the user's pain pattern It is characterized in that it includes; a monitoring server that identifies and guides objective statistical values for

Here, the wrist wearable terminal uses a three-axis motion sensor including a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, a smart watch or a smart band worn on the user's wrist. It is characterized in that it is implemented as

In addition, the medical staff terminal is a terminal that medical staff can directly use, and it is characterized in that it can access a monitoring server to search for a pain behavior pattern, obtain a treatment record for the user, and input a report therefor.

And only the model used for inference is mounted on the wrist worn terminal so that on-device deep learning can be performed, and it is characterized by further comprising a deep learning learning server that performs a model learning process that requires computation from the outside.

In addition, the wrist wearable terminal includes a sensing data acquisition unit that collects motion sensing data of a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer, and when sensing data is collected as much as a specific window size, feature extraction for extracting the feature value of the window It is characterized in that it includes an on-device deep learning pain behavior recognition unit that recognizes pain behavior by recognizing the type of hand motion and the position of the hand through on-device deep learning when the feature value is extracted from the part and the feature extraction unit.

In addition, the monitoring server includes a user information storage unit for receiving and registering user information input through a user terminal, and a pain behavior change trend calculation for storing the user's pain behavior data transmitted through the user terminal and calculating the pain behavior change trend A pain behavior pattern inquiry unit that performs pattern inquiry processing when a pain behavior pattern inquiry input is received through the medical staff terminal, and a pain behavior report that stores the pain behavior report input through the medical staff terminal and provides the pain behavior report to the user through the user terminal It is characterized in that it includes a provision unit.

A pain management method using a pain behavior monitoring system according to the present invention for achieving another object includes collecting motion sensing data from sensors including a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer in a wrist worn terminal. ;When sensing data is collected as much as a specific window size, extracting the feature value of the window; On-device deep learning pain recognition to recognize the pain behavior by recognizing the type of hand motion and the position of the hand through on-device deep learning when the feature value is extracted It characterized in that it comprises; behavior recognition step.

Here, the step of transmitting the pain behavior information recognized in the on-device deep learning pain behavior recognition step to the user terminal, and transmitting the pain behavior information to the monitoring server by the user terminal is characterized in that it further comprises.

And when the user terminal transmits the pain behavior and recognition result to the monitoring server, the monitoring server is characterized in that the pain behavior change trend is calculated.

And when the medical staff terminal accesses the monitoring server and inputs information to obtain the desired data and searches the pain behavior pattern, the monitoring server provides the desired user's pain behavior pattern information to the medical staff terminal, and the pain provided through the medical staff terminal When a medical team prepares a pain behavior report based on the behavior pattern information and inputs it to the monitoring server, the monitoring server provides a report on pain behavior to the user terminal.

As described above, the pain behavior monitoring system and the pain management method using the same according to the present invention have the following effects.

First, it is possible to collect and provide objective information on pain behavior using a wearable communication terminal.

Second, by collecting and providing objective information on pain behavior, medical staff can even monitor various pain patterns.

Third, the model learning process, which requires a relatively large amount of computation, is performed externally, and only the model used for inference is loaded on the device to perform on-device deep learning inference to enable efficient pain behavior monitoring regardless of storage performance and computational performance. do.

Fourth, it is possible to check the current situation of the user through interworking with the user terminal and the medical staff terminal, and to identify and guide objective statistical values for the user's pain pattern.

Fifth, by recognizing two types of hand motion and hand position, and synthesizing the recognized two pieces of information to recognize and classify pain behavior, various pain behaviors are combined and recognized to provide a variety of pain behaviors such as chest pain, abdominal pain, back pain, arthralgia, and muscle pain. It allows us to judge the symptoms of pain behavior.

Sixth, a medical staff terminal that can be directly used by medical personnel is connected to the monitoring server to perform a search operation, to obtain a treatment record for the user, and to input a report therefor.

1 is a configuration diagram comparing the characteristics of sensors used in sensor-based monitoring technology;
2 is a configuration diagram showing various pain behavioral characteristics;
3 is a block diagram of a pain behavior monitoring system according to the present invention;
4 is a detailed configuration diagram of a wrist wearable terminal according to the present invention;
5 is a detailed configuration diagram of a monitoring server according to the present invention;
6 is a flowchart illustrating a pain behavior recognition process according to the present invention;
7 is a configuration diagram showing a user registration process;
8 is a block diagram showing the process of recognizing and storing pain behavior
9 is a block diagram showing a process of writing a report on pain behavior in a medical staff terminal

Hereinafter, preferred embodiments of the pain behavior monitoring system and the pain management method using the same according to the present invention will be described in detail as follows.

The characteristics and advantages of the pain behavior monitoring system and the pain management method using the same according to the present invention will become apparent through detailed description of each embodiment below.

1 is a configuration diagram comparing characteristics of sensors used in sensor-based monitoring technology, and FIG. 2 is a configuration diagram showing various pain behavior characteristics.

The pain behavior monitoring system and the pain management method using the same according to the present invention enable the collection and provision of objective information on pain behavior using a wearable communication terminal.

To this end, the present invention may include a configuration in which the medical staff monitors various pain patterns by collecting and providing objective information on pain behavior.

In the present invention, the model learning process that requires relatively many calculations is performed externally, and only the model used for inference is loaded on the device to perform on-device deep learning inference, enabling efficient pain behavior monitoring regardless of storage performance and computational performance It may include a configuration to do so.

The present invention may include a configuration for checking the current situation of the user through mutual interworking with the user terminal, the medical staff terminal, and the like, and identifying and guiding the user's objective statistical values for the user's pain pattern.

The present invention recognizes two types of hand motions and hand positions and recognizes and classifies pain behaviors by synthesizing the recognized two pieces of information, thereby combining and recognizing various pain behaviors, such as chest pain, abdominal pain, back pain, joint pain, muscle pain, etc. It may include a configuration that allows to determine the symptoms caused by various pain behaviors.

The present invention may include a configuration in which a medical staff can directly use a medical terminal that can be used to access a monitoring server, perform a search operation, obtain a treatment record for a user, and input a report therefor.

Behavioral recognition is a technology for recognizing a user's movement. There are computer vision-based technology that utilizes images, and sensor-based technology that utilizes information obtained by attaching a sensor to a user.

Computer vision-based technology has various environmental restrictions, such as camera installation and the presence of obstacles.

The sensor-based technology recognizes the user's movement by using a wearable device that can be attached to the user's body or can be easily put on and taken off.

Sensor-based technology is a technology that extracts features using motion sensors such as accelerometers and gyroscopes and uses the extracted features as input values for algorithms, machine learning, and deep learning to perform behavior recognition.

This technology can be considered as the most commercialized technology because it is easy to mount on small devices such as smart bands, smart watches, and smartphones. For example, by using motion sensors built-in in smart bands such as Xiaomi Mi Band and Galaxy Fit and various smart devices such as Galaxy S, exercise is analyzed and life pattern improvement service is suggested to users according to the analyzed amount of exercise.

The present invention relates to a system and method for monitoring pain behavior related to chronic pain, in particular, to collect and provide objective information about pain behavior using a wearable communication terminal, and to enable medical staff to monitor various pain patterns. It relates to a pain behavior monitoring system and method.

As shown in FIG. 2, various chronic pains such as chronic back pain, chest pain, and abdominal pain generally last 3 or 6 months or more after the onset of the disease, and it is necessary to manage the prognosis well.

In order to solve such chronic pain, it is necessary to understand the frequency and location of the pain well.

However, it is difficult for patients and medical staff to accurately and concretely judge their own pain during treatment hours, and the expression of pain is often subjective.

In addition, there are many cases in which the patient easily passes over the pain behavior for pain that occurs unconsciously.

There is a technique for recognizing pain behavior for chronic low back pain using an EMG (electromyography, electromyography) sensor, but it is difficult to use in practice because wearability is not considered. In addition, it has a disadvantage of very poor versatility by recognizing an operation for one chronic pain.

Accordingly, the present invention is to collect and provide objective information on pain behavior caused by chronic pain using a wrist-worn communication terminal, and to provide medical staff monitoring for pain patterns.

3 is a block diagram of a pain behavior monitoring system according to the present invention.

The pain behavior monitoring system according to the present invention includes a deep learning learning server 10 , a wrist wearable terminal 20 , a user terminal 30 , a monitoring server 40 , and a medical staff terminal 50 .

The wrist wearable terminal 20 is implemented as a smart watch or smart band worn on the user's wrist using a 3-axis motion sensor such as a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer. can be

This generates sensing data by sensing wrist motion caused by pain behavior, recognizes pain behavior, and communicates with the user terminal 30 using short-range wireless communication such as Bluetooth.

The user terminal 30 communicates with the wrist wearable terminal 20 through short-range wireless communication such as Bluetooth, and is implemented as a user's smartphone capable of interworking with the monitoring server 40 .

This notifies the pain behavior recognized by the wearable communication terminal 20 to the monitoring server 40 , and allows various information provided by the monitoring server 40 to be displayed to the user.

The monitoring server 40 checks the user's current situation through interworking with the user terminal 30 and the medical staff terminal 50, and identifies and guides objective statistical values for the user's pain pattern.

The medical staff terminal 50 is a terminal that medical personnel can use directly, and may access the monitoring server 40 to perform a search operation, obtain a treatment record for the user, and input a report therefor.

In the present invention, the external deep learning learning server 10 performs a model learning process that requires relatively many calculations, and only the model used for inference is mounted on the device to perform on-device deep learning inference.

Recently, deep learning has been used in various devices and applications as it shows higher accuracy than existing techniques in various fields. However, there is a problem that requires a lot of storage space and high computational performance to use deep learning.

Due to these limitations, devices with insufficient storage space and computational performance can utilize deep learning by utilizing a separate server. However, there is a problem in that various costs such as power consumption and communication costs occur in devices that use batteries and operate wirelessly, such as wearable devices and automobiles.

In order to solve this cost problem, the study of on-device deep learning inference that enables the use of deep learning within the device is needed. Using deep learning inference within the device without a communication connection with the outside, which is a variable cost source, solves the problem.

Deep learning technology trains a model that has a specific dataset and consists of weights that represent features of it. It is a technology that makes inferences by properly classifying features by inputting new data into the learned model.

Here, the model composed of many weights is too large to be computed. On-device deep learning solves the problems of deep learning by using various techniques such as quantization of weights and fixed-point arithmetic.

In addition, the model learning process, which requires a relatively large number of computations, is performed externally, and only the model used for inference is loaded on the device to perform on-device deep learning inference.

The detailed configuration of the wrist wearable terminal 20 is as follows.

4 is a detailed configuration diagram of a wrist wearable terminal according to the present invention.

The wrist wearable terminal according to the present invention includes a sensing data acquisition unit 21 that collects motion sensing data such as a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer, and when the sensing data is collected by a specific window size, the characteristics of the window The feature extractor 22 that extracts the value, and the on-device deep learning that recognizes the pain behavior by recognizing the type of hand motion and the position of the hand through on-device deep learning when the feature value is extracted from the feature extracting unit 22 and a behavior recognition unit 23 .

And the detailed configuration of the monitoring server 40 is as follows.

5 is a detailed configuration diagram of a monitoring server according to the present invention.

As shown in FIG. 5 , the monitoring server 40 includes a user information storage unit 41 that receives and registers user information input through the user terminal 30 , and the user's pain behavior data transmitted through the user terminal 30 . A pain behavior change trend calculation unit 42 that stores and calculates a pain behavior change trend, and a pain behavior pattern inquiry unit 43 that performs a pattern inquiry process when a pain behavior pattern inquiry input is received through the medical staff terminal 50; , and a pain behavior report providing unit 44 that stores the pain behavior report input through the medical staff terminal 50 and provides it to the user through the user terminal 30 .

The pain management method using the pain behavior monitoring system according to the present invention having such a configuration will be described in detail as follows.

6 is a flowchart illustrating a pain behavior recognition process according to the present invention.

The wrist wearable terminal 20 that recognizes the pain behavior generated by the pain behavior collects motion sensing data such as a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer. The collected sensing data recognizes pain behavior in real time using a recognition method as shown in FIG. 6 .

Pain behavior recognition collects sensing data as much as a specific window size (S601), extracts window feature values (S602), and recognizes the type of hand motion (S603) and the position of the hand through on-device deep learning (S604) pain Recognize action (S605)

The wrist-worn terminal calculates roll, pitch, and yaw using Equation 1 from the collected sensing data and uses them as feature values. Additionally, 3-axis accelerometer and 3-axis gyroscope data are used together as feature values.

As described above, the present invention recognizes two types of hand gestures and a hand position using the extracted feature values using two types of on-device deep learning inferences, and recognizes pain behaviors by synthesizing the recognized two pieces of information.

As the pain behaviors to be recognized are classified into two parts, the pain behavior recognition part is divided into two parts as shown in FIG. can be judged

In hand gesture recognition, a wide variety of movements such as supporting, tapping, scratching, grabbing, and rubbing can be recognized. Each of the above hand gestures appears as a periodic movement, a sudden large movement, etc., and can be recognized using on-device deep learning inference using these characteristics.

Among the characteristics, the accelerometer may indicate velocity variability in a specific direction, and a periodic change of direction may be expressed as the characteristic.

In addition, since all recognized features can represent strong peaks and weak peaks, these features can be well learned in the on-device deep learning learning process.

In hand position recognition, various hand positions such as back, waist, stomach, shoulder, arm, leg, calf, knee, and groin can be recognized. The position of the hand as above has different characteristics, such as the inclination of the hand and the direction of the wrist, and these characteristics can be extracted using roll, pitch, and yaw. Therefore, these characteristics can be well learned in the on-device deep learning learning process.

The on-device deep learning learning process trains the on-device deep learning model by using a dataset related to the target to be recognized, just like general deep learning learning.

The difference is that in on-device deep learning, a model trained outside the device is loaded on the device, and the device performs only on-device inference. In order to construct the data set, various people wearing wrist-worn terminals are performed by combining hand gestures and hand positions that they want to recognize. An on-device deep learning model is trained using a dataset composed by manually inputting hand movements and hand positions for each time.

And the monitoring server that calculates and guides the change trend of the pain pattern has a characteristic of calculating and guiding the change trend of the pain pattern, such as information about continuous pain symptoms or sudden abnormal pain.

Such a change in pain pattern is classified and analyzed on a daily/weekly/monthly/year basis, and a report from a medical staff can be given as an example.

7 is a block diagram illustrating a user registration process, and FIG. 8 is a block diagram illustrating a pain behavior recognition and storage process.

And FIG. 9 is a block diagram showing a process of writing a report on pain behavior in a medical staff terminal.

When the user terminal 30 starts accessing the monitoring server 40 for the first time, the same procedure as in FIG. 7 is performed so that the user's personal information is received and stored in the user account to provide the monitoring service to the user. .

When a user account is created, the monitoring server 40 receives the pain behavior and time recognized from the user terminal 30 as shown in FIG. 8 and analyzes it on a daily/weekly/monthly/year basis. do.

And, as shown in FIG. 9, when accessing the monitoring server 40 through the medical staff terminal 50, the user's pain behavior analysis information can be obtained and a report can be input therefor.

The report exemplifies heart disease related to chest pain, joint disease related to arthralgia, and gastrointestinal disease related to abdominal pain.

Specifically, when the user terminal 30 transmits the pain behavior and recognition result to the monitoring server 40 , the monitoring server 40 calculates a change trend in the pain behavior.

In this state, when the medical staff terminal 50 accesses the monitoring server 40 and inputs information to obtain the desired data and searches the pain behavior pattern, the monitoring server 40 returns the desired user's pain behavior pattern information to the medical staff. It is provided through the terminal 50 .

When the medical staff prepares a pain behavior report based on the pain behavior pattern information provided through the medical staff terminal 50 and inputs it to the monitoring server 40 , the monitoring server 40 guides the pain behavior report to the user terminal 30 . do

The pain behavior monitoring system and the pain management method using the same according to the present invention described above enable medical staff to monitor various pain patterns by collecting and providing objective information about pain behavior using a wearable communication terminal. it was made to

In particular, the present invention recognizes two types of hand gestures and the position of the hand, and synthesizes the recognized two pieces of information to recognize and classify the pain behavior, thereby combining and recognizing various pain behaviors, such as chest pain, abdominal pain, back pain, arthralgia, It allows you to judge the symptoms caused by various pain behaviors, such as muscle pain.

As described above, it will be understood that the present invention is implemented in a modified form without departing from the essential characteristics of the present invention.

Therefore, the specified embodiments are to be considered in an illustrative rather than a restrictive view, the scope of the present invention is indicated in the claims rather than the foregoing description, and all differences within the equivalent scope are included in the present invention. will have to be interpreted.

10. Deep Learning Learning Server
20. Wrist Wearable Terminal
30. User terminal
40. Monitoring Server
50. Medical staff terminal

Claims (10)

a wrist wearable terminal that is worn on the user's wrist using a three-axis motion sensor to recognize two types of hand motions and hand positions, and recognize pain behaviors by synthesizing the recognized two pieces of information;
a user terminal communicating with the wrist-worn terminal and interworking with a monitoring server;
A monitoring server that checks the current situation of the user through interworking with the user terminal and the medical staff terminal, and identifies and guides objective statistical values for the user's pain pattern; includes;
The wrist wearable terminal recognizes hand gestures including supporting, tapping, scratching, grabbing, and rubbing according to the learning of periodic movement of hand gestures, movement magnitude, speed variability of moving in a specific direction, and peak characteristics of periodic direction change. ,
By extracting hand inclination and wrist direction features using roll, pitch, and yaw, the hand position of the back, waist, belly, shoulder, arm, leg, calf, knee, and groin is recognized,
A pain behavior monitoring system, characterized in that the pain behavior is recognized by synthesizing the type of hand motion and the position information of the hand. According to claim 1, wherein the wrist wearable terminal,
Pain behavior characterized by being implemented as a smart watch or smart band worn on the user's wrist using a 3-axis motion sensor including a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer monitoring system. According to claim 1, wherein the medical staff terminal,
A pain behavior monitoring system, characterized in that it can access a monitoring server with a terminal that can be directly used by medical staff, perform an operation of searching for a pain behavior pattern, obtain a treatment record for a user, and input a report therefor. According to claim 1, wherein only the model used for inference is mounted on the wrist worn terminal so that on-device deep learning can be performed, and further comprising a deep learning learning server that externally performs a model learning process that requires calculation Pain behavior monitoring system, characterized in that. According to claim 1, wherein the wrist wearable terminal,
A sensing data acquisition unit that collects motion sensing data of a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer;
a feature extraction unit that extracts a feature value of the window when sensing data is collected as much as a specific window size;
Pain behavior monitoring system, characterized in that it comprises an on-device deep learning pain behavior recognition unit that recognizes pain behavior by recognizing the type of hand motion and the position of the hand through on-device deep learning when the feature value is extracted from the feature extraction unit. According to claim 1, wherein the monitoring server,
a user information storage unit for receiving and registering user information input through a user terminal;
A pain behavior change trend calculator for storing the user's pain behavior data transmitted through the user terminal and calculating the pain behavior change trend;
A pain behavior pattern inquiry unit that performs pattern inquiry processing when a pain behavior pattern inquiry input is received through the medical staff terminal;
and a pain behavior report providing unit that stores the pain behavior report input through the medical staff terminal and provides the pain behavior report to the user through the user terminal. Collecting motion sensing data from sensors including a 3-axis accelerometer, a 3-axis gyroscope, and a 3-axis magnetometer in a wrist-worn terminal;
extracting a feature value of a window when sensing data is collected as much as a specific window size;
When the feature value is extracted, the on-device deep learning pain behavior recognition step of recognizing the pain behavior by recognizing the type of hand motion and the position of the hand through on-device deep learning;
In the on-device deep learning pain behavior recognition stage, it includes supporting, tapping, scratching, grabbing, and rubbing according to the learning of the periodic movements of hand gestures, the magnitude of the movements, the velocity variability of moving in a specific direction, and the peak characteristics of the periodic change of direction. hand gesture recognition,
By extracting hand inclination and wrist direction features using roll, pitch, and yaw, the hand position of the back, waist, belly, shoulder, arm, leg, calf, knee, and groin is recognized,
A pain management method using a pain behavior monitoring system, characterized in that the pain behavior is recognized by synthesizing the type of hand motion and the position information of the hand. The method of claim 7, wherein the pain behavior information recognized in the on-device deep learning pain behavior recognition step is transmitted to the user terminal,
The pain management method using the pain behavior monitoring system, characterized in that it further comprises the step of transmitting, by the user terminal, the pain behavior information to the monitoring server. The pain management method using the pain behavior monitoring system according to claim 8, wherein when the user terminal transmits the pain behavior and recognition result to the monitoring server, the monitoring server calculates the change trend of the pain behavior. 10. The method of claim 9, wherein when the medical staff terminal accesses the monitoring server and inputs information for obtaining the desired data and searches the pain behavior pattern, the monitoring server provides the desired user's pain behavior pattern information to the medical staff terminal,
Based on the pain behavior pattern information provided through the medical staff terminal, when the medical staff writes a pain behavior report and inputs it to the monitoring server, the monitoring server guides the pain behavior report to the user terminal using a pain behavior monitoring system. How to manage pain.

Images