KR20220001875A - Method for providing cognitive training contents using assignment allocation for prevention of dementia - Google Patents

Abstract

Provided is a dementia prevention cognitive function training service providing method capable of remote home training which comprises the following steps of: receiving registration of at least one user terminal from a manager terminal; generating a task template including at least one cognitive function training content of a selected type and difficulty when the type and difficulty of the at least one cognitive function training content pre-stored in the manager terminal are selected; and allocating the cognitive function training content included in the task template to at least one designated user terminal when designating the at least one user terminal registered in the manager terminal and task exposure time.

Description

A method of providing cognitive function training service for dementia prevention that allows remote home training

The present invention relates to a method of providing a dementia prevention cognitive function training service capable of remote at-home training, and provides a platform capable of generating and assigning training contents optimized for individuals as tasks.

The recent increase in the elderly population due to an aging society leads to a relative increase in geriatric diseases such as vascular dementia and Alzheimer's. Accordingly, the importance of physical exercise, continuous care, daily life, and health management support for the elderly in order to prevent the disease at an early stage and to slow the progression is emerging. Dementia treatment is mainly aimed at maintaining brain function or alleviating or delaying disease progression by reactivating degenerative brain cells. It is very important to maintain and improve the memory function created through repetitive practice. Therefore, it is effective to use a cognitive therapy program that activates brain function and improves cognitive ability through stimulation.

At this time, a method for providing dementia cognitive treatment contents was researched and developed. In the announcement on August 09, 2009), cognitive training contents to prevent dementia are provided to users' terminals, and users' brain function decline is prevented through repetitive cognitive training, and cognitive training is provided for each course according to the user's dementia symptoms. It provides cognitive training content for users and provides content for cognitive training in the form of problem-solving games so that users can continue cognitive training for a long time without feeling bored. Based on the evaluation result information of the contents for cognitive training, the contents that can determine the user's dementia status and the contents that determine the degree of activation of human brain function are subdivided into multiple classification areas and levels and stored, and contents that improve brain function A configuration is disclosed in which the content stored in the storage unit is selected for each area according to a predetermined classification method, and the user's achievement is given according to the score obtained according to the feedback delivered by the user to the content selected by the content selection unit. .

However, even if the above-described configuration is used, it is difficult to induce steady training due to the nature of the cognitive therapy program provided as game content, and there is no compulsion and arbitrariness that does not necessarily need to be solved. Also, there are various game contents to maintain and improve frontal lobe execution, spatiotemporal cognitive ability, attention concentration, memory, language ability, and calculation ability, but people try to continue only in the fields they are good at, and the fields they are not good at are The complex game does not progress due to giving up or not even trying from the beginning. Even if a game in a complex field is played, the general public, who is not an expert, can proceed with the program in a medium recovery method because they do not have professional knowledge about which field training is required and do not have a system. Accordingly, there is a need for research and development of a platform that can systematically and complexly conduct cognitive therapy programs.

An embodiment of the present invention enables an expert administrator to register and manage each user, and use a task assignment function to select and allocate dementia prevention training content corresponding to cognitive treatment suitable for an individual, a cognitive treatment program It can impose coercion on the progress of the brain and induce the progress of various programs, and it can maintain and improve the overall cognitive function of prefrontal execution, spatiotemporal cognitive ability, attention concentration, memory, language ability, and maintenance of computational ability. It can be managed so that the ability of one field is not left behind, and when users in a group solve the same problem, competition can be induced by listing the ranking based on the evaluation score, and at least one parameter measured for solving the problem By calculating the evaluation score based on However, the technical task to be achieved by the present embodiment is not limited to the technical task as described above, and other technical tasks may exist.

As a technical means for achieving the above-described technical problem, an embodiment of the present invention includes the steps of registering at least one user terminal from the manager terminal, the type and difficulty of at least one cognitive function training content pre-stored in the manager terminal. When selected, generating a task template including at least one cognitive function training content of the selected type and difficulty, and when designating at least one user terminal and task exposure time registered in the manager terminal, included in the task template and allocating cognitive function training content to at least one designated user terminal.

According to any one of the problem solving means of the present invention described above, an expert administrator can register and manage each user, and a task assignment function for selecting and allocating dementia prevention training content corresponding to cognitive treatment suitable for an individual is provided. By using it, it is possible to impose coercion on the progress of the cognitive therapy program and induce the progress of a steady and diverse program, and the overall cognitive function for prefrontal execution, spatiotemporal cognitive ability, attention concentration, memory, language ability, and computational ability maintenance. You can maintain and improve your skills while keeping your skills in any one area behind, and when users in a group solve the same problem, you can induce competition by listing the rankings based on the evaluation score and solve the problem. By calculating an evaluation score based on at least one measured parameter, it is possible to use the calculated evaluation score as feedback data and to reassign tasks for each stage.

1 is a view for explaining a system for providing a dementia prevention cognitive function training service capable of remote at-home training according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating a training service providing server included in the system of FIG. 1 .
3 and 4 are diagrams for explaining an embodiment in which a dementia prevention cognitive function training service capable of remote at-home training according to an embodiment of the present invention is implemented.
5 is an operation flowchart illustrating a method of providing a dementia prevention cognitive function training service capable of remote at-home training according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily implement them. However, the present invention may be embodied in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is "connected" with another part, this includes not only the case of being "directly connected" but also the case of being "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated, and one or more other features However, it is to be understood that the existence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded in advance.

The terms "about", "substantially", etc. to the extent used throughout the specification are used in or close to the numerical value when manufacturing and material tolerances inherent in the stated meaning are presented, and are intended to enhance the understanding of the present invention. To help, precise or absolute figures are used to prevent unfair use by unconscionable infringers of the stated disclosure. As used throughout the specification of the present invention, the term "step of (to)" or "step of" does not mean "step for".

In this specification, a "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be implemented using two or more hardware, and two or more units may be implemented by one hardware.

Some of the operations or functions described as being performed by the terminal, apparatus, or device in the present specification may be performed instead of by a server connected to the terminal, apparatus, or device. Similarly, some of the operations or functions described as being performed by the server may also be performed in a terminal, apparatus, or device connected to the server.

In this specification, some of the operations or functions described as mapping or matching with the terminal means mapping or matching the terminal's unique number or personal identification information, which is the identification data of the terminal. can be interpreted as

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a view for explaining a system for providing a dementia prevention cognitive function training service capable of remote at-home training according to an embodiment of the present invention. Referring to FIG. 1 , the dementia prevention cognitive function training service providing system 1 capable of remote home training includes at least one user terminal 100 , a training service providing server 300 , and at least one manager terminal 400 . may include However, since the dementia prevention cognitive function training service providing system 1 capable of remote home training of FIG. 1 is only an embodiment of the present invention, the present invention is not limitedly interpreted through FIG. 1 .

At this time, each component of FIG. 1 is generally connected through a network 200 . For example, as shown in FIG. 1 , at least one user terminal 100 may be connected to the training service providing server 300 through the network 200 . In addition, the training service providing server 300 may be connected to at least one user terminal 100 and at least one manager terminal 400 through the network 200 . In addition, at least one manager terminal 400 may be connected to the training service providing server 300 through the network 200 .

Here, the network refers to a connection structure in which information exchange is possible between each node, such as a plurality of terminals and servers, and examples of such networks include RF, 3rd Generation Partnership Project (3GPP) network, and Long Term (LTE). Evolution) network, 5th Generation Partnership Project (5GPP) network, WIMAX (World Interoperability for Microwave Access) network, Internet, LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network) , PAN (Personal Area Network), Bluetooth (Bluetooth) network, NFC network, satellite broadcasting network, analog broadcasting network, DMB (Digital Multimedia Broadcasting) network, etc. include, but are not limited thereto.

In the following, the term at least one is defined as a term including the singular and the plural, and even if the at least one term does not exist, each element may exist in the singular or plural, and may mean the singular or plural. it will be self-evident In addition, that each component is provided in singular or plural may be changed according to embodiments.

The at least one user terminal 100 may be a terminal of a user that executes cognitive function training content using a web page, an app page, a program or an application related to dementia prevention cognitive function training service capable of remote home training. In addition, the at least one user terminal 100 may be a terminal that receives cognitive function training content as a task from the manager terminal 400 when it is registered with an institution. And, the at least one user terminal 100 outputs the assigned cognitive function training content during the task exposure period of the assigned task, and when the deadline has passed or the task exposure period expires, the assigned cognitive function training content It may be a terminal whose exposure is terminated. In addition, when the at least one user terminal 100 outputs cognitive function training content and receives an input from the user, the data collected through at least one sensor or camera of the user terminal 100 is transmitted to the training service providing server 300 . ) may be a terminal that transmits And, when the at least one user terminal 100 is grouped or clustered into a group by the manager terminal 400, it may be a terminal that lists the ranking based on the evaluation score of other users in the group and the evaluation score of the user. have. In addition, at least one user terminal 100 may be a terminal that requests completion of the task as a reminder when the progress rate of the cognitive function training content does not satisfy the preset progress rate before the task exposure period expires.

Here, the at least one user terminal 100 may be implemented as a computer that can access a remote server or terminal through a network. Here, the computer may include, for example, navigation, a laptop equipped with a web browser, a desktop, and a laptop. In this case, the at least one user terminal 100 may be implemented as a terminal capable of accessing a remote server or terminal through a network. At least one user terminal 100 is, for example, as a wireless communication device that guarantees portability and mobility, navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS(Personal Handyphone System), PDA(Personal Digital Assistant), IMT(International Mobile Telecommunication)-2000, CDMA(Code Division Multiple Access)-2000, W-CDMA(W-Code Division Multiple Access), Wibro(Wireless Broadband Internet) ) terminal, a smart phone, a smart pad, a tablet PC, etc. may include all types of handheld-based wireless communication devices.

The training service providing server 300 may be a server that provides a dementia prevention cognitive function training service web page, an app page, a program or an application capable of remote home training. In addition, the training service providing server 300 may be a server that receives the registration of the user terminal 100 of at least one user managed by the manager from the manager terminal 400 . In addition, the training service providing server 300 may be a server that converts cognitive function training contents into a database by difficulty and type, and generates a task template including cognitive function training contents of the difficulty and type selected by the manager terminal 400 . And, when the task is assigned by designating the task exposure period and at least one user terminal 100, the assigned task is a server that outputs cognitive function training content during the task exposure period from the at least one user terminal 100 can In addition, the training service providing server 300 may be a server that collects information such as a progress rate of the cognitive function training content in at least one user terminal 100 and outputs it in the manager terminal 400 . In addition, the training service providing server 300 is a server that calculates an evaluation score by collecting at least one event collected from at least one user terminal 100 or an IoT-based terminal interworking with at least one user terminal 100 . can be In addition, the training service providing server 300 may be a server that allows at least one user terminal 100 to share a ranking within a group based on the calculated evaluation score.

Here, the training service providing server 300 may be implemented as a computer capable of accessing a remote server or terminal through a network. Here, the computer may include, for example, navigation, a laptop equipped with a web browser, a desktop, and a laptop.

The at least one manager terminal 400 may be a terminal of a manager using a web page, an app page, a program or an application related to dementia prevention cognitive function training service capable of remote home training. At this time, the at least one manager terminal 400 registers at least one user terminal 100 of at least one user managed by the manager, and creates and assigns a task template to the at least one user terminal 100, It may be a terminal that receives and outputs the result as feedback.

Here, the at least one manager terminal 400 may be implemented as a computer capable of accessing a remote server or terminal through a network. Here, the computer may include, for example, navigation, a laptop equipped with a web browser, a desktop, and a laptop. In this case, the at least one manager terminal 400 may be implemented as a terminal capable of accessing a remote server or terminal through a network. At least one manager terminal 400, for example, as a wireless communication device that guarantees portability and mobility, navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS(Personal Handyphone System), PDA(Personal Digital Assistant), IMT(International Mobile Telecommunication)-2000, CDMA(Code Division Multiple Access)-2000, W-CDMA(W-Code Division Multiple Access), Wibro(Wireless Broadband Internet) ) terminal, a smart phone, a smart pad, a tablet PC, etc. may include all types of handheld-based wireless communication devices.

2 is a block diagram for explaining a training service providing server included in the system of FIG. 1, and FIGS. 3 and 4 are dementia prevention cognitive function training services capable of remote home training according to an embodiment of the present invention. It is a diagram for explaining an embodiment.

Referring to FIG. 2 , the training service providing server 300 includes a registration unit 310 , a generation unit 320 , an assignment unit 330 , a grouping unit 340 , a calculation unit 350 , a task management unit 360 , It may include a sharing unit 370 and an utterance management unit 380 .

The training service providing server 300 according to an embodiment of the present invention or another server (not shown) operating in conjunction with at least one user terminal 100 and at least one manager terminal 400 enables remote home training. When the dementia prevention cognitive function training service application, program, app page, web page, etc. are transmitted, at least one user terminal 100 and at least one manager terminal 400 are provided with dementia prevention cognitive function training capable of remote home training You can install or open service applications, programs, app pages, web pages, etc. In addition, the service program may be driven in at least one user terminal 100 and at least one manager terminal 400 by using a script executed in a web browser. Here, the web browser is a program that enables the use of a web (WWW: world wide web) service. It means a program that receives and displays hypertext written in HTML (hyper text mark-up language), for example, Netscape. , Explorer, Chrome, etc. In addition, the application means an application on the terminal, for example, includes an app (app) executed in a mobile terminal (smartphone).

Referring to FIG. 2 , the registration unit 310 may receive registration of at least one user terminal 100 from the manager terminal 400 . For example, assuming that organization A manages users C to Y in zone B, and it is the responsibility of administrator Z, administrator Z can register users C to Y by searching based on identifiers such as user IDs of C to Y. can Alternatively, after the administrator creates and registers the IDs and passwords of users C to Y, the generated IDs and passwords are delivered to users C to Y so that they can log in.

The generator 320 generates a task template including at least one cognitive function training content of the selected type and difficulty when the type and difficulty of at least one pre-stored cognitive function training content are selected in the manager terminal 400 . can do. In this case, the task template may include a task template name and cognitive function training content, and the task template name may include a type and difficulty level of the cognitive function training content. In addition, one task template may include at least one set of information that bundles at least one cognitive function training content. These task templates can be output as a list of templates, and even if the administrator does not select each cognitive function training content, even if the administrator selects only the type and difficulty level, the list of task templates is automatically output, and the administrator selects the list of task templates Selecting any one of them can be configured to create an assignment.

The allocator 330, when designating the at least one user terminal 100 registered in the manager terminal 400 and the task exposure time, assigns the cognitive function training content included in the task template to the designated at least one user terminal 100 ) can be assigned. In this case, when the allocator 330 creates a group in the manager terminal 400 and assigns the same task to the group, the assignment unit 330 designates the group itself so as to substitute for individually designating users included in the group. do. In addition, the task exposure time may be set from the time the task is assigned to the time the task is finished, but it does not exclude arbitrarily designating it in the manager terminal 400 .

The grouping unit 340 receives at least one user terminal 100 from the manager terminal 400 in the registration unit 310 and then groups the at least one user terminal 100 registered in the manager terminal 400 into a group. In the case of clustering, at least one user terminal 100 included in the group may be identified by tagging the group name. In addition, when the same task is assigned to at least one user terminal 100 included in the group, the performance result of the task may be shared among the at least one user terminal 100 included in the group. The performance result of the task may be, for example, a ranking according to an evaluation score to be described later.

The calculation unit 350, when the assignment unit 330 specifies at least one user terminal 100 registered in the manager terminal 400 and the task exposure time, the cognitive function training content included in the task template is designated at least When the cognitive function training content included in the task template is outputted from the at least one user terminal 100 after allocating to one user terminal 100 , at least one event occurring in the at least one user terminal 100 . can be collected. And, the calculation unit 350, the difficulty (wL), correct answer (R), hint usage (H), task execution time (T), opportunity usage (C), concentration of the cognitive function training content included in the task template (F), it is possible to calculate the evaluation score based on the free or non-assisted (A) and the positive (Positive) addition point (E). Here, the evaluation score may be calculated by Equation 1 below, but does not exclude variations according to the embodiment.

At this time, R, which is the correct answer, means whether the task was performed correctly, and it is a value that measures whether or not the correct answer matches the correct answer or the degree of similarity to the correct answer depending on the task. H, the hint usage amount, which can be provided to a hint while performing a task, means whether or not the provided hint is used, and is measured by the amount of hint used. T, which is the task execution time, means the time taken while performing the task, and is a value measured from the start time of the task to the time it is completed. Opportunity usage C means the amount of consumed opportunity while performing a task. In this case, C is measured as the amount of opportunity used, and since repeated learning is possible, it can indicate whether the corresponding task is a repetitive learning task. F, which is concentration, means a degree of concentration while performing a task, and may be measured as the time spent looking at the screen after analyzing images and audio using a camera or a microphone, but is not limited thereto. In this case, the time and concentration of looking at the screen may be calculated through gaze analysis in the image. A without assistance means whether or not the person received help from others while performing the task. This is also the result of measuring whether there was an assistant while performing the task using a camera or microphone.

E (Extra Point), which is a positive extra point, means a score that is measured differently depending on the task. Extra points can be given when the task is in a positive emotional state. For example, in the case of social activity training, it may be a score based on the amount of conversations shared while performing a task or cooperation calculated according to the degree of distribution of utterances measured by the user. In the case of physical activity training, it may be a score according to the degree or accuracy of induced physical activity measured by a sensor while performing a task. In the case of emotional activity training, it may be an emotional distribution according to a user's expression calculated by a camera input. In summary, when the cognitive function training content included in the task template is social activity training content, the cognitive function training content included in the task template is based on the amount of speech recognized by the user terminal 100, and the cognitive function training content included in the task template is physical activity training content. In this case, when the cognitive function training content included in the task template is emotional activity training content based on the amount of physical activity measured in the user terminal 100 or the IoT-based terminal interlocked with the user terminal 100, the user terminal 100 It can be based on the user's emotional distribution, which is an object in the image frame collected from the camera of

In this case, emotion classification can use P.Ekman's human emotion classification model. The reason for using this is that the method of using multiple sensors is because there are many instruments that need to be mounted on the user's body and it is the simplest way to classify emotions using facial expressions. it will be self-evident

emotional state change in facial expression Surprise -Eyebrows are raised, raised high and curved
- The skin under the eyebrows is pulled
-Horizontal Wrinkles on the forehead
- eyelids open wide; The upper eyelid goes up and the lower eyelid goes down. The sclera of the eye is visible above the iris
-Chin goes down and mouth and teeth open, but there is no change in the muscles of the mouth Fear -Eyebrows rise and fall together
-The wrinkles on the forehead are gathered in the center, and the entire forehead is not wrinkled.
- the upper eyelid is raised, the sclera is exposed and the lower eyelid is pulled down
-Mouth is opened and lips are slightly pulled down and down or drooping down Disgust -Upper lip rises
-The lower lip also rises and supports the rise of the upper lip or protrudes slightly as it descends.
- wrinkled nose
- Cheeks rise
- Lines appear in the lower eyelid, the eyelids are raised but not tense
-Eyebrows go down and upper eyelids go down Anger - eyebrows go down
-Vertical wrinkles appear between the eyebrows
-The lower eyelid is pulled and sometimes lowered by the eyebrows
-The lower eyelid may be pulled and raised
-Both eyes stare for a long time and protrude slightly
-Close your lips firmly so that both mouths are straight or down. If you shout, tense in a square shape Happiness - Mouth curls backwards and rises
-Mouth may be open and teeth may be visible
-Wrinkles are created from the nose to the edge of the mouth
- Cheeks rise
-Wrinkles under the lower eyelids, rising but not straining
-Crow's feet are created at the outer edges of both eyes Sadness -The inner part of the eyebrows goes down
-The skin under the eyebrows is triangularized and the inner eyebrows are raised
- the inner part of the upper eyelid rises
-Dropped lips or trembling lips

In this case, the positive emotional state may be set to surprise and happiness among the emotional state classifications described above, but is not limited to the listed ones. Here, when the above-described facial features are learned by an artificial intelligence algorithm, when a user's face is input as a query, it can be determined whether or not to add a score. For example, the facial images of happy people are accumulated enough to build big data and stored, and the above-described emotional state characteristics are set as ROI (Region of Interest), that is, a feature within the region of interest to facilitate learning and training. In this case, when a certain face image is input as a query later, it is possible to check whether the face input as a query is a happy face or the like. To this end, when a sufficient amount of data to model the artificial intelligence algorithm is accumulated, the calculator 350 stores the accumulated raw data in parallel and distributed, and stores the raw data included in the raw data. Unstructured data, structured data, and semi-structured data may be refined, and preprocessing including classification may be performed as meta data. Then, the calculator 350 may construct big data by analyzing the preprocessed data including data mining. The calculator 350 performs analysis including data mining on the preprocessed data to construct big data, and then divides the data set of the built big data into training data and experimental data. Intelligent learning can proceed. Here, the learning may be performed by any one or a combination of at least one of supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. not limited

For example, when using artificial intelligence algorithms, regression and neural networks can be used, and generalized additive models, random forests, boosted MARS, Nonparametric models such as SVM can be used to predict emotional state, and ensemble learning can be applied to emotional state estimation. Alternatively, based on data crawled by web crawling, multiple regression models, neural network models, and decision trees can be applied to generate emotional state prediction models.

At this time, the neural network has a structure in which one or more hidden layers are arranged between an input layer composed of independent variables and an output layer composed of dependent variables, and each layer is independent Nodes (or neurons) are arranged, and the nodes of the hidden layer receive the values received from the previous layer and their respective weights as a weighted sum, and pass the activation function to the next layer. These operations are performed in the hidden layer, and finally, the result value can be derived after being performed in the output layer. Unlike linear regression, which assumes a linear relationship between the dependent variable and the independent variable, artificial neural networks explain the nonlinearity of data well and show high explanatory power even for complex and irregular patterns. control can be seen.

SVR (Support Vector Regression) is a regression model based on SVM (Support Vector Machine). SVM finds a hyperplane that maximizes the margin that classifies these data by expressing data in a high-dimensional vector space using a kernel function. SVR is a method that applies the idea of SVM to a regression problem. The existing regression model defines the difference between the actual value and the predicted value as a loss function, whereas the SVR defines an insensitive loss function that ignores the case where the difference is smaller than a specific value and calculates it as a loss only when the difference is large. SVR uses a kernel function to express data in a high-dimensional vector space, and is a form of estimating a linear regression model that minimizes the insensitive loss function in the vector space.

Random Forests extract multiple bootstrap samples from training data, learn an independent decision tree for each sample, and combine the outputs of multiple decision trees to make the final It is an ensemble method based on a decision tree that determines an output value. When growing each decision tree, it has the effect of reducing the correlation between the learned decision trees and consequently reducing the variance of predictions by considering only some randomly selected variables for each branch. When growing trees, a decision tree for the purpose of classification goes through a process of repeatedly separating data in the direction of reducing impurity by using data impurity measures such as Gini coefficient or entropy. In addition to this, various variable setting, big data collection, pre-processing, and learning may be performed, and the present embodiment is not limited to the above.

The concentration P is, in the image frame input through the camera of the user terminal 100, the size of the pupil, the position of the pupil, the eye fixation (Eye Fixation), the eye jump (Saccade), the gaze time (Fixation Duration) and the number of times of gaze ( Fixation Count) based on eye tracking (EyeTracking) algorithm can be calculated. At this time, since various eye tracking algorithms have been developed, any one may be used. However, the algorithm for concentration, ie, immersion, detects a face in an image generated by a camera, and detects the In order to record the continuous information of the face, the face is tracked until it disappears from the camera, and through eye tracking, it is possible to measure and use the gaze status and gaze time of the content by time period. First, YOLO-v3 can be used to recognize a face, and a part recognized as a face is displayed as a bounding box. In this case, a binarization operation may be performed using a previously published adaptive threshold estimation method for pupil detection. After that, the pupil is separated after binarization by using the characteristic that the gray level of the pupil region of the eye is lower than that of the surrounding region, and the center of gravity of the black pixel is obtained to remove noise other than the eye, and the center of the pupil is obtained. Finally, by calibrating the position of the light reflected on the pupil in the captured image, it is possible to determine whether the user is gazing at the screen.

The assistance free man A is a face other than a face having a corresponding facial feature point of the user in the image frame received from the camera of the user terminal 100, or a voice other than the user's voice among the voices recognized by the user terminal 100 Whether or not the assistant's voice is recognized can be determined by classifying it with a speaker identification algorithm. At this time, speaker identification is a technique for modeling and classifying phonetic characteristics of individual vocal tracts. Low-level features such as short-term frequency fluctuations and pitch heights, and high-level features such as short-term, words, syllables or long-term frequency fluctuations It is preferable to use a machine learning algorithm because it models rich phonetic features such as For example, a machine learning algorithm-based Gaussian mixture model, or a convolution-pooling operation with robustness to the limit sensitive to discrepancy between learning and test data caused by noise about environmental factors and STFT (Short-Time Fourier Transform) algorithm may be used. In the former case, it can be used simply when it is performed in a quiet room, but it is not suitable for a noisy environment, and in the latter case, it is robust against noise caused by other environmental factors such as living noise. It is also possible to selectively apply the former or the latter based on the execution environment of the cognitive function training content.

Taking the latter case as an example, in the data learning step, speech samples of a plurality of speakers may be input. The input speech sample undergoes a three-step feature extraction process. The first step is to generate a unique speech map (SoundMap), the second step is to normalize it, and the third step is to sample it in a certain section using a sliding window algorithm. The preprocessed window updates the internal weight of a multi-layered convolutional neural network to ensure noise robustness and models a general pattern for each speaker. In the recognition stage, the speaker of each window is classified by the output activation function value without updating the internal weight of the convolutional neural network. Since the speakers classified for each window may be different, unlike the learning stage, the classified results are arithmetic averaged to finally classify the speakers.

Since the above-described speech data has a time series characteristic rather than a static signal, a single characteristic for each speech tends to be meaningless. Accordingly, in order to extract features without losing the correlation between time, frequency, and volume, an STFT algorithm that performs a Fourier transform operation for each time section may be used. Since actual speech data is a discontinuous function, discontinuous signal, time measurement, frequency measurement, and window length are applied as variables when calculating STFT. A sound map is a three-dimensional vector having the horizontal axis as time, the vertical axis as frequency, and pixel intensity as the size. Appropriate for modeling with Local Connectivity properties. The generated sound map has a different size for each length of speech and its dimension is very high, so it can be sampled into uniform sections using a sliding window algorithm and divided into windows.

Also, in the field of speech recognition, where there is a large discrepancy between data during training and testing, noise robustness is the most important factor. Noise robustness can be secured by repeatedly applying a convolution-pooling operation to speaker identification according to an embodiment of the present invention. The generalization performance of the convolution operation is suitable for modeling the variance factors of individual phonetic features. In the convolution-pooling operation, the convolution-pooling module may include, but is not limited to, a dropout layer and three fully-connected layers. Learning of convolution operation is a process of finding the weight of a filter that extracts correlation while preserving spatial relationships between features. It refers to the dimensionality reduction process that reduces complexity. If this is repeated, strong noise robustness against living noise can be secured.

Weights other than the difficulty weight wL of Equation 1 may be classified into a + weight having a value of 1 or more and a - weight having a value less than 1. Alternatively, the + weight can be set in such a way that a value of + is literally given, and the - weight is given a - value. That is, when it has a value of 1 or more, the corresponding factor acts as a factor that increases the score, and when it has a value of less than 1, the corresponding factor acts as a factor that lowers the score. + weights are wL, wR, wP, and wE, and - weights are wH, wT, wC, and wA. That is, the higher the degree of difficulty, the closer or correct the answer is, the higher the concentration, the higher the positive score, the higher the final evaluation score. Conversely, the higher the hint usage, the longer the task execution time, the larger the opportunity usage, and the more help you receive, the lower your final evaluation score. Accordingly, those to which a + weight is added may be defined as parameters whose evaluation score increases as the size increases, and those to which a - weight is added may be defined as parameters whose evaluation score decreases as the size increases. Of course, it will be apparent that the + weight and the - weight are not determined and may be changed according to embodiments. Further, it is not limited to the listed weights and is not excluded for reasons not listed.

The task management unit 360 is based on the difficulty of the cognitive function training content included in the task template in the calculation unit 350, the correct answer, the hint usage, the task execution time, the opportunity usage, the concentration, the freedom of assistance and the positive addition points. After calculating the evaluation score, the pre-stored step-by-step task corresponding to the evaluation score may be reassigned to the at least one user terminal 100 .

The task management unit 360, when designating at least one user terminal 100 registered in the manager terminal 400 and the task exposure time in the assignment unit 330, assigns the cognitive function training content included in the task template to the specified at least After allocating to one user terminal 100 , the progress rate of the cognitive function training content included in the task template in at least one user terminal 100 may be monitored and updated in real time. Then, the task management unit 360, when the remaining task exposure time of the task exposure time enters less than a preset time, and the progress rate does not satisfy the preset completion rate, a reminder call (Call) to at least one user terminal 100 ) or send a message. Through this, it is possible to perform the task without sagging while maintaining tension in the state set on time by giving compulsion and setting a deadline. In addition, the task performance rate can be analyzed and visualized so that the administrator can check the performance level of users by group and task at a glance.

Sharing unit 370, after receiving at least one user terminal 100 from the manager terminal 400 in the registration unit 310, at least one user terminal 100 from the cognitive screening test (Cognitive Screening Test) results In the case of registration, the cognitive screening test result may be shared with the manager terminal 400 . Of course, in addition to cognitive screening tests, Mini-Mental Status Examination, Hasegawa Dementia Scale, 7-Minute Screening, Clock Drawing Test, and Simple Blessed Test (Short Blessed Test, SBT), Mini-Memory Test, Korean Version of Community Dementia Screening Tool (CSID-K), Korean Version of Mini-Mental State Examination, Korean Version of MMSE ( Korean Mini-Mental State Examination), MMSE (MMSE-KC) of CERAD Assessment Book in Korean, Korean Version of Hasegawa Dementia Scale Revised, or IADL (Assessment for Daily Living Performance), etc., but are listed are not excluded for reasons not listed. In addition to this, various tests can be received from various institutions, can be integrated and shared with administrators.

The speech management unit 380 receives at least one user terminal 100 from the manager terminal 400 in the registration unit 310, and then, from the interactive artificial intelligence speaker interlocked with the at least one user terminal 100, a conversation It is possible to collect voice conversations between the type artificial intelligence speaker and the user, and map the collected voice conversation analysis result to at least one user terminal 100 and store it. In this case, at least one user terminal 100 may be linked with a natural language processing-based conversational artificial intelligence speaker. Among the cognitive abilities of the elderly living alone or the elderly who rarely engage in social activities, language ability is the first to deteriorate. There is a reason that language ability decreases due to aging, but what accelerates this further is that there is no one to talk to and the opportunity to speak decreases. Therefore, it is possible to determine what kind of cognitive function training content to provide in the future by making a routine conversation try through a chatting agent and collecting the results.

First of all, if there is no other person to talk to, you can make it play the role of a companion. For example, when the user does not use the user terminal 100 for a long time in the evening and uses the user terminal 100 in the morning, it may be determined that the user wakes up after going to bed. In this case, you can perform a preemptive speech providing a preset message such as "Good night? Today's weather is XX and the news is XX. What are you planning to do today?" In a study examining how an agent's interaction strategy according to the degree of autonomy could be accepted by the elderly, it was found that preemptive speech was acceptable, and conversations about practical safety and health aspects of the elderly were preferred. An utterance strategy with high autonomy is viewed as proactive, and an utterance strategy with low autonomy is viewed as on-demand. , utterances according to requests refer to utterances that respond only to the utterances of the elderly. As the human brain reflexively thinks the answer the moment it hears a question, it can induce brain rotation by external stimuli even if there is no answer.

To this end, the conversational agent (Agent) may use a character-based conversational artificial intelligence. Conversational artificial intelligence that recognizes situations based on context can be used. Through continuous conversation, it is possible to understand the situation of conversation and express subjective opinions and emotions based on this. In this case, the interactive agent includes a plurality of unit agents, and each agent may be further subdivided into a knowledge base and an inference engine. The conversational agent composed of each unit agent reads and understands the conversation of the other party, and recognizes the overall conversation situation in consideration of one's own emotional state. Through this, it can proceed in the order of devising a response strategy, establishing an action plan and executing it.

First, the target language analysis agent is an agent that analyzes the meaning of the target and the emotions of the target through the sentences received from the other party. This agent grasps the meaning and emotion that the other party wants to convey through fragmentary analysis and in-depth analysis from the decision-making agent's feedback and the other party's remarks, and delivers them to the conversational situation judgment agent. Fragment analysis refers to a semantic and emotional analysis through the sentences received from the current counterpart. In-depth analysis is an analysis module for inferring the partner's actual intentions and emotions by accumulating the intentions and emotions of the partner updated through the feedback received from the situational awareness agent. Because in-depth analysis consists of only the values of fragment analysis at the beginning of the conversation, preconceived notions about the subject are reflected, and as the conversation progresses, it is updated to match the intention and emotion of the actual subject.

The second is a self-emotion identification agent, which is an agent for modeling changes in AI's own emotions through sentences received from the other party. This agent, like the target language analysis agent, uses fragment analysis and in-depth analysis to understand the AI's own emotions from the other person's remarks and delivers them to the conversational situation determination agent. Modeling through a fragmentary reaction is a momentary emotional change that occurs through a strong reaction over a certain level of the other person. This affects sudden emotional changes or instantaneous expression of emotions caused by strong stimuli. Modeling through in-depth reaction models the emotional change that occurs and accumulates from the situation in which the conversation with the other party is in progress.

The third is the dialogue situation determination agent, which grasps the dialogue situation through the meaning and emotional state received from the target language analysis agent and the self-emotion identification agent, and the feedback of the decision-making agent, and delivers it to the decision-making agent do. In addition, feedback on the situation identified by the target language analysis agent and self-emotion identification agent is provided to update each in-depth analysis module. The fourth is a decision-making agent, which is an agent that determines the intention of the next statement, including the purpose of artificial intelligence's utterance, in the situation identified through the conversational situation-judging agent. The decision-making agent provides feedback to the reasoning engine of the target language analysis agent and the dialogue situation determination agent to achieve a subjective interpretation of a given situation. In order to achieve its own purpose, artificial intelligence makes decisions through simulation through the emotion model created by the target language analysis agent and the self-emotion identification agent.

The fifth is the response generating agent, which is an agent that generates the intention of the next statement determined through the decision-making agent at the level of the sentence. The sixth is a sentence expression agent. The sentence expression agent is an agent that transforms the resulting sentence of the response generating agent to fit the grammar and a specific accent. In the case of learning and training each of these agents, it can be modeled to enable progressively in-depth conversations by the re-learning and retrained contents, and the preset preemptive dialogue is like a parrot through continuous training and re-learning in addition to the same sentence. A database can be built so that diversity can be felt like a real person rather than an artificial intelligence speaker who only speaks the same thing.

Hereinafter, an operation process according to the configuration of the training service providing server of FIG. 2 will be described in detail with reference to FIGS. 3 and 4 as an example. However, it will be apparent that the embodiment is only one of various embodiments of the present invention and is not limited thereto.

Referring to FIG. 3A , a screen of a training service according to an embodiment of the present invention is shown. Referring to FIG. 3B , cognitive interventional therapy, that is, a service for an administrator to manage a plurality of users, is provided. As shown in FIG. 3C , a relationship structure between an occupational therapist (administrator) and a patient (user) is formed. 3D and 3E show content for training cognitive abilities in various fields, and through the task assignment function as shown in FIG. 3F, tasks are assigned to each individual as optimized content and a platform capable of analyzing the results this is provided

Referring to FIG. 4A , the training service according to an embodiment of the present invention can save human and time resources, provide immediate feedback, enable self-repetitive learning, and perform precise measurement such as time measurement. Based on the objective data, it is possible to systematically manage each user efficiently. 4B shows that users using the platform of the present invention are divided into users (individual users) and administrators (institutional users), and the user management flow system of the administrator is shown in FIG. 4C. Referring to Fig. 4D, it is shown that an embodiment of the present invention is provided as either individual training or group training, and Fig. 4E shows the process of training. From FIG. 4F, an example of an execution screen of an administrator is shown. As shown in FIG. 4G, the administrator may issue IDs of each user (participant), create a task template as shown in FIG. 4H, and assign it to each user. Fig. 4i shows a screen for managing assignment templates. A plurality of templates can be created or deleted, assignment templates can be created as shown in Fig. 4j, and can be assigned to each user as shown in Fig. 4k, and the result is also You can check it in real time. Tasks can be assigned as shown in FIG. 4L, and groups can be created and managed for each group as shown in FIG. 4M. In order to create a group, a simple method of pressing a confirmation button after inputting a group name as shown in FIG. 4N may be used. When a plurality of users in a group solve a problem as shown in FIG. 4O, a ranking listed by the evaluation score may be output, and as shown in FIG. 4P, the result of executing the cognitive ability training content by the user may be shown. can A screen for managing the progress of the task is shown in FIG. 4q, the completion rate of completing the task is output in a table, and a message sending function through text or messenger is included to the user who did not complete the task as shown in FIG. 4r, so that the management is easy. do. 4S shows a function that an administrator can create and send a question to know about a user in a survey format. In FIG. 4T, different login and billing methods for personal and institutional accounts, and when belonging to each institution It is shown that the method of joining can be performed differently. However, since this is a policy issue, it is not limited to the listed methods. When the administrator assigns the task in this way, the task to be performed by the user today is output as shown in FIG. 4V, and when the task is clicked, the cognitive ability training content is immediately executed. Figure 4w is a summary of the menu of the cognitive function training service according to an embodiment of the present invention, but it is not limited to the listed menu, and it will be obvious that it can be changed according to the embodiment.

As described above, the method of providing a dementia prevention cognitive function training service capable of remote home training of FIGS. 2 to 4 is described above with respect to the method of providing a dementia prevention cognitive function training service capable of remote home training through FIG. 1. Since it is the same as the content or can be easily inferred from the described content, the following description will be omitted.

5 is a diagram illustrating a process in which data is transmitted/received between components included in the dementia prevention cognitive function training service providing system capable of remote at-home training of FIG. 1 according to an embodiment of the present invention. Hereinafter, an example of a process in which data is transmitted and received between each component will be described with reference to FIG. 5, but the present application is not limited to such an embodiment, and the example shown in FIG. 5 according to various embodiments described above will be described. It is apparent to those skilled in the art that the data transmission/reception process may be changed.

Referring to FIG. 5 , the training service providing server receives registration of at least one user terminal from the manager terminal ( S5100 ).

In addition, the training service providing server generates a task template including at least one cognitive function training content of the selected type and difficulty when the type and difficulty of at least one cognitive function training content pre-stored in the manager terminal is selected (S5200) ) When the manager terminal designates at least one registered user terminal and the task exposure time, the cognitive function training content included in the task template is allocated to the designated at least one user terminal (S5300).

The order between the above-described steps ( S5100 to S5300 ) is merely an example and is not limited thereto. That is, the order between the above-described steps ( S5100 to S5300 ) may be mutually changed, and some of these steps may be simultaneously executed or deleted.

As described above, the method of providing a dementia prevention cognitive function training service capable of remote home training of FIG. 5 is described above with respect to the method of providing a dementia prevention cognitive function training service capable of remote home training through FIGS. 1 to 4 above. Since it is the same as the content or can be easily inferred from the described content, the following description will be omitted.

The method of providing a dementia prevention cognitive function training service capable of remote at-home training according to an embodiment described with reference to FIG. 5 is in the form of a recording medium including instructions executable by a computer, such as an application or program module executed by a computer can also be implemented. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer-readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The method of providing a dementia prevention cognitive function training service capable of remote at-home training according to an embodiment of the present invention described above includes an application basically installed in a terminal (which may include a program included in a platform or operating system basically installed in the terminal) ), and may be executed by an application (ie, a program) installed directly on the master terminal by a user through an application providing server such as an application store server, an application, or a web server related to the corresponding service. In this sense, the above-described dementia prevention cognitive function training service providing method capable of remote at-home training according to an embodiment of the present invention is implemented as an application (ie, a program) installed by default in a terminal or directly installed by a user and installed in the terminal. It may be recorded on a computer-readable recording medium, such as.

The above description of the present invention is for illustration, and those of ordinary skill in the art to which the present invention pertains can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a dispersed form, and likewise components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention. do.

Claims (10)

In the training service providing method executed in the training service providing server,
Receiving registration of at least one user terminal from the manager terminal;
generating a task template including at least one cognitive function training content of the selected type and difficulty when the type and difficulty of at least one pre-stored cognitive function training content are selected in the manager terminal; and
allocating cognitive function training content included in the task template to the designated at least one user terminal when the manager terminal specifies the registered at least one user terminal and the task exposure time;
A method of providing a dementia prevention cognitive function training service capable of remote at-home training comprising a.
The method of claim 1,
After receiving the registration of at least one user terminal from the manager terminal,
when the manager terminal clusters the registered at least one user terminal into a group, tagging a group name to identify at least one user terminal included in the group;
further comprising,
When the same task is assigned to at least one user terminal included in the group, the performance result of the task is shared between at least one user terminal included in the group, dementia prevention cognitive function training capable of remote home training HOW TO PROVIDE SERVICES.
The method of claim 1,
When the manager terminal specifies the registered at least one user terminal and the task exposure time, after allocating the cognitive function training content included in the task template to the designated at least one user terminal,
when the at least one user terminal outputs the cognitive function training content included in the task template, collecting at least one event occurring in the at least one user terminal;
Calculating an evaluation score based on the difficulty of the cognitive function training content included in the task template, correct answer, hint usage, task execution time, opportunity usage, concentration, help freedom and positive addition points;
The method of providing a dementia prevention cognitive function training service capable of remote home training further comprising a.
4. The method of claim 3,
The positive addition point is,
When the cognitive function training content included in the task template is social activity training content, based on the amount of speech recognized by the user terminal,
When the cognitive function training content included in the task template is physical activity training content, based on the amount of physical activity measured by the user terminal or an IoT-based terminal interlocked with the user terminal,
When the cognitive function training content included in the task template is emotional activity training content, dementia prevention cognitive function training capable of remote home training is based on the user's emotional distribution, which is an object in the image frame collected from the camera of the user terminal HOW TO PROVIDE SERVICES.
4. The method of claim 3,
The concentration is
Eye tracking based on the size of the pupil, the position of the pupil, eye fixation, saccade, fixation duration, and fixation count in the image frame input through the camera of the user terminal (EyeTracking) A method of providing a cognitive function training service for dementia prevention that is calculated by the algorithm, remote home training is possible.
4. The method of claim 3,
The freedom of assistance is that a face other than a face having a corresponding facial feature point of the user in the image frame received from the camera of the user terminal is recognized, or an assistant voice other than the user's voice among the voices recognized by the user terminal is recognized A method of providing a dementia prevention cognitive function training service capable of remote at-home training, in which it is determined by classifying whether or not by a speaker identification algorithm.
4. The method of claim 3,
After the step of calculating the evaluation score based on the difficulty of the cognitive function training content included in the task template, correct answer, hint usage, task execution time, opportunity use, concentration, help free and positive added points,
reassigning the pre-stored step-by-step task corresponding to the evaluation score to the at least one user terminal;
The method of providing a dementia prevention cognitive function training service capable of remote home training further comprising a.
The method of claim 1,
When the manager terminal specifies the registered at least one user terminal and the task exposure time, after allocating the cognitive function training content included in the task template to the designated at least one user terminal,
monitoring the progress rate of the cognitive function training content included in the task template in the at least one user terminal and updating it in real time;
transmitting a reminder call or message to the at least one user terminal when the remaining task exposure time of the task exposure time enters less than a preset time and the progress rate does not satisfy a preset completion rate;
further comprising,
When monitoring the progress rate of the cognitive function training content in the manager terminal to visualize and display the degree of task performance by user, group, and task, dementia prevention cognitive function training service providing method capable of remote home training.
The method of claim 1,
After receiving the registration of at least one user terminal from the manager terminal,
when registering a cognitive screening test result from the at least one user terminal, sharing the cognitive screening test result with the manager terminal;
The method of providing a dementia prevention cognitive function training service capable of remote home training further comprising a.
The method of claim 1,
The at least one user terminal is interlocked with a natural language processing-based conversational artificial intelligence speaker,
After receiving the registration of at least one user terminal from the manager terminal,
collecting voice conversations between the interactive artificial intelligence speaker and the user from the interactive artificial intelligence speaker linked with the at least one user terminal;
storing a result of analyzing the collected voice conversations by mapping the result of the analysis with the at least one user terminal;
The method of providing a dementia prevention cognitive function training service capable of remote home training further comprising a.

Images