KR102515987B1 - Apparatus and method for detecting learners' participation in an untact online class - Google Patents

Abstract

An apparatus and method for detecting a learner's non-face-to-face online class participation are disclosed. An apparatus for detecting non-face-to-face online class participation according to an embodiment is a personal electronic device equipped with an eye tracking device for detecting whether a learner receiving a non-face-to-face online class is participating in accordance with the progress of the online class, A data collection unit that receives the eye-tracking data acquired by the tracking device from the personal electronic device of each of the learners participating in the online class. A calculation unit that calculates the gaze fixation point, which is the fixed position, and a class participation state determination unit that determines whether each learner participates in class using the distribution of the gaze fixation points of the learners calculated by the calculation unit. .

Description

Apparatus and method for detecting learners' participation in an untact online class}

The present invention relates to a technology for detecting whether a learner is participating in a class, and relates to a technology for detecting whether a learner is participating in a non-face-to-face online class by using eye tracking data and utilizing the detected result. .

Recently, due to Corona 19, online school is being implemented and interest in the non-face-to-face industry is increasing. Accordingly, a trend change is taking place in the education industry that combines education and information and communication (ICT) technology, edutech, and non-face-to-face online classes are being conducted or promoted not only in schools, but also in various national and public educational institutions and private educational institutions. It is becoming.

Current non-face-to-face online classes are conducted under the premise of voluntary participation by learners who are not under the teacher's supervision or supervision, and in many cases, teachers cannot control the class situation. In addition, it is easy for learners to be exposed to an environment where they can use smartphones, game consoles, or computers even during online classes. As a result, in non-face-to-face online classes, the problem of low learning participation rate of learners is emerging.

In order to solve this problem, various technologies have been proposed to detect whether learners are normally participating in non-face-to-face online classes.

As an example, there is a method of displaying an interface for inputting a specific response on a learner's screen, and determining participation in learning according to whether an accurate response is input to the interface within a predetermined time. However, since this method detects participation only at the time when the corresponding interface is displayed, it is difficult to say that the teacher controls the class situation. In addition, due to the interface displayed on the class content, the progress of the online class may be hindered, and in some cases, an interface unrelated to the contents of the class content may be displayed.

As another example, there is a method of determining whether a learner is normally participating in a non-face-to-face online class by using eye tracking technology. For example, in Korea Patent Publication No. 10-2014-0146750, "Gaze-based Educational Contents Execution System" (Patent Document 1), a point to be observed is calculated based on the learner's gaze information, and the calculated learner's gaze position is calculated. A technique for controlling whether or not a video of instructional content continues to progress according to whether is an image of educational content (eg, a screen on which content is being displayed) is disclosed. However, since this method determines that the learner is normally participating in the class only when the learner is always watching the corresponding content, it is difficult to accurately determine whether the learner is participating in the flow of the ongoing class. For example, according to the method of Patent Document 1, while the learner is not watching the content while viewing the textbook or writing notes, it is determined that he or she is not participating in the class, or unlike most learners who are looking at the textbook, the content alone In the case of watching, most learners do not participate in the class, but it can be determined that only one person is participating in the class. In addition, even when the learner is gazing at a part other than the part being lectured (the area where the content is displayed) of the entire video at that time, it may be determined that the learner is participating in the class.

As another example, there is also a method of determining whether a learner is normally participating in a non-face-to-face online class using an image captured by a camera. For example, in Korea Patent Publication No. 10-2020-0050281, "Learning support system and method using artificial intelligence-based augmented reality and virtual reality" (Patent Document 2), video by the teacher's point of view along with the teacher's voice information and a technology for determining a student's learning state based on a virtual reality-based learning image generated using an image from a student's point of view together with audio information of the student. However, this method requires a large amount of data processing because the teacher's image and the student's image must be acquired and combined to generate a virtual reality-based learning image. In addition, it is only disclosed that information on the student's learning state is calculated based on the student's motion information collected during class, and specific criteria for determining the student's learning state are not disclosed.

In addition, there are Non-Patent Document 1 and Non-Patent Document 2 as a study on detecting online learning participation in relation to gaze movement, and Non-Patent Document 3 to Non-Patent Document 5 as a study on learners' participation in learning.

Korean Patent Publication No. 10-2014-0146750 Korean Patent Publication No. 10-2020-0050281

S. Aslan, Z. Cataltepe, I. Diner, O. Dundar, A.A. Esme, R. Ferens, M. Yener, Learner Engagement Measurement and Classification in 1:1 Learning (International Conference on Machine Learning and Applications, Detroit, 2014) L.B. Krithika, P.G.G. Lakshmi, Student emotion recognition system (SERS) for e-learning improvement based on learner concentration metric. Procedia Computer Science 85, 767-776 (2016) N. Bosch, Detecting Student Engagement: Human Versus Machine (Conference on User Modeling Adaptation and Personalization, Halifax, 2016) N. Bosch, S.K. D'Mello, R.S. Baker, J. Ocumpaugh, V. Shute, M. Ventura, W. Zhao, Detecting Student Emotions in ComputerEnabled Classrooms (International Joint Conference on Artificial Intelligence, New York, 2016) J.A. Fredrick, P.C. Blumenfeld, A.H. Paris, School engagement: Potential of the concept, state of the evidence. Rev. Educ. Res. 74(1), 59-109 (2004)

One problem to be solved by the present invention is to use the learner's eye-tracking data in a situation where non-face-to-face real-time online classes are conducted to more accurately determine the learner's participation in learning, the learner's non-face-to-face online class participation To provide a detection device and method.

Another problem to be solved by the present invention is that in a situation where non-face-to-face real-time online classes are conducted, learners' participation in non-face-to-face online classes can adaptively determine the learning participation status of learners according to the progress of the class. To provide a detection device and method.

One embodiment of the present invention for solving the above problems is a device for detecting whether a learner receiving a non-face-to-face online class with a personal electronic device equipped with an eye tracking device is participating in accordance with the progress of the online class, A data collection unit that receives the eye-tracking data obtained by the eye-tracking device from the personal electronic device of each of the learners participating in the online class, and each of the learners using the eye-tracking data collected by the data collection unit. A calculation unit that calculates a gaze fixation point, which is a location where the gaze of the learner is fixed, and a class participation state determination unit that determines whether each learner participates in class using the distribution of the gaze fixation point of the learners calculated by the calculation unit. include

According to one aspect of the embodiment, the class participation state determination unit determines, among the gaze fixation points of each of the learners, the gaze fixation point corresponding to the largest number of learners as the gaze region, and then the gaze fixation point of the specific learner is located in the gaze region. Based on whether or not the learner is included, it is possible to determine whether or not the learner participates in the class.

According to another aspect of the embodiment, the calculation unit also calculates the fixation time at the gaze fixation point of each of the learners, and the class participation state determination unit uses the gaze movement information by the gaze fixation point and the fixation time. Thus, it is determined whether each of the learners participates in the class. In this case, the class participation state determination unit may determine whether each of the learners participate in the class by using content information of the online class including the class type and execution time as well.

The class participation state determination unit may be implemented using an online class participation detection model based on unsupervised learning artificial intelligence.

According to another aspect of the embodiment, the class participation state determining unit may further include a feedback unit for transmitting a notification message to the personal electronic device of a learner determined to be not participating in class.

Another embodiment of the present invention for solving the above problems is a method for detecting whether a learner receiving a non-face-to-face online class with a personal electronic device equipped with an eye tracking device is participating in accordance with the progress of the online class, Receiving the gaze tracking data obtained by the gaze tracking device from the personal electronic device of each of the learners participating in the online class, using the gaze tracking data collected in the collection step, the gaze of each learner Calculating gaze fixation points, which are fixed positions, and determining whether each learner participates in class using the distribution of the gaze fixation points of the learners calculated in the calculation step.

According to the above-described embodiment of the present invention, since the motion of the gaze including the fixation point of each learner's gaze calculated based on the learner's gaze tracking data and the fixation time at the fixed point is utilized, the learner's learning Participation status can be judged more accurately. In particular, by using the distribution of gaze fixation points of learners in a specific time range, each learner's participation in learning is determined based on whether some gaze fixation points are different from most learners or whether the gaze movement is out of the normal range, A more accurate and adaptive judgment is possible. In addition, by providing appropriate feedback to students or teachers using these judgment results, the participation rate of learners in non-face-to-face online classes can be improved.

1 is a block diagram schematically showing the schematic configuration of a non-face-to-face online class participation detection device according to an embodiment of the present invention.
2 is a schematic diagram for explaining an example of a method of calculating a fixed position of the gaze using gaze coordinate values.
3 is an example of a diagram schematically illustrating a movement pattern of a specific learner's gaze.

Hereinafter, preferred embodiments and examples of the present invention will be described with reference to the drawings. However, the following embodiments and examples are merely illustrative of preferred configurations of the present invention, and the scope of the present invention is not limited to these configurations. And in the following description, the hardware configuration and software configuration of the device, processing flow, manufacturing conditions, size, material, shape, etc. are not intended to limit the scope of the present invention to these unless specifically described. .

According to the present invention, based on eye tracking data of learners participating in non-face-to-face online classes, it is determined whether each learner is normally participating in online classes. Generally, the state of the learner's gaze (fixation status, location of the fixation point, fixation time, movement pattern, etc.) means not only visual movement but also selective and active perception for problem solving. Information to understand the cognitive process of learners can be obtained. In particular, if the fixed time at the corresponding point is known along with the point at which the eye gazes, it is possible to infer information such as how long attention is paid to a certain learning content and skipping without reading in real time.

However, depending on the contents of the content or the teaching method of the instructor (teacher), the point at which the learner should gaze (fixed point) or gaze time (fixed time) may vary. For example, when an instructor explains something contained in the content, the learner should gaze at that part of the content on the display, but when the instructor explains with reference to the textbook or asks the learner to solve problems or take notes. On the back, the learner should gaze at the teaching materials or notes on the desk rather than the displayed content, so that it can be seen that he is participating in the course of the class. In other words, if the learner gazes at the content in a different location that has nothing to do with the part the lecturer is explaining even though they are staring at the display screen, or if they stare at the display screen or a third location even though they are supposed to gaze at the textbook or notebook, the learner It is difficult to see that they are participating in line with the progress of online classes.

However, in an online class that lasts for a predetermined amount of time, the fixation time may vary along with the location (fixed point of gaze) that the learner should gaze at depending on the progress of the class, the instructor's class method, and the contents of the displayed content. . If online classes are conducted with recorded lecture videos, it may be possible to determine in advance the location where learners should stare at a specific time by referring to the content or lecture content, but in the case of conducting non-face-to-face online classes in real time, However, since the position where the learner should gaze is flexible during the online class, it is almost impossible to pre-determine the fixation time along with the fixed point of the learner's gaze at a specific time.

In an embodiment of the present invention described later, in consideration of these characteristics of non-face-to-face real-time online classes, it is assumed that among the learners participating in the online class at a specific time, the learners participating according to the progress of the class will usually occupy the majority. Under this method, learners responding to multiple gaze movements are judged to be participating in the class, but learners who deviate from multiple gaze movements are judged not to be participating in the class, thereby basically determining the learner's class participation status. .

More specifically, in the embodiment of the present invention, it is determined whether a specific learner is participating in an online class by using the distribution of learners at the fixed point of the gaze (ie, how many learners correspond to the fixed point). . For example, if more than half, preferably more than 2/3 of the learners of a predetermined group (eg, 30) gaze at the same point, it is determined that the learners who gaze at the corresponding point normally participate in the class, but other A learner who gazes at a point may be judged not to have participated in the class. Alternatively, depending on the embodiment, the gaze time (fixed time of the gaze) for the corresponding point may be additionally used as a judgment criterion. Alternatively, the motion of the gaze considering all of the fixation points of the plurality of gazes, the fixation time at each fixation point, and changes over time may be used as a criterion for determination.

However, in this determination, the content of the learning content or the teaching method of the instructor (teacher) may be additionally considered. In addition, such determination may be performed using a predetermined model based on artificial intelligence (eg, a class participation detection model). According to an embodiment of the present invention described later, the class participation detection model clusters states in similar situations based on unsupervised learning, and repeats modification (update) of the clusters to adapt to new experiences to obtain classified optimal values. One way is to perform reinforcement learning.

Hereinafter, with reference to the accompanying drawings, preferred embodiments of the present invention will be described.

1 is a block diagram schematically showing the schematic configuration of a non-face-to-face online class participation detection device according to an embodiment of the present invention. Referring to FIG. 1, the non-face-to-face online class participation detection device 100 (hereinafter referred to as 'detection device') includes a data collection unit 110, a calculation unit 120, and a class participation state determination unit 130. . And the detection device 100 may further include a feedback unit 140.

Such a detection device 100 may be implemented, for example, through a computer program executable on a computer processor and a recording medium (ie, a computer-readable recording medium) in which the program is stored. The type of recording medium on which the computer program is stored is not particularly limited, and may be permanently or temporarily stored on the corresponding medium. In addition, the individual components constituting the detection device 100 are logically divided based on their respective functions, and each may be implemented as a separate function module or two or more components may be implemented as one function module.

In addition, the detection device 100 may further include a database (not shown) for storing data input to each component and/or data generated as a result of processing in each component. Alternatively, the detection device 100 may use a storage means of a computer in which a computer processor executing it is installed as a database. Therefore, the flow of data indicated by arrows in FIG. 1 is only for intuitive understanding of the technical characteristics of the present invention, and substantially through the input and output of data occurring between each component and the database, the flow of data indicated by arrows in FIG. flow can be implemented.

The detection device 100 of FIG. 1 is a specific online class (eg, content 1, etc.) It is a device for detecting whether you are participating according to the progress, that is, the class participation status. Here, learners participate in non-face-to-face online classes using personal electronic devices such as personal computers, notebook computers, tablet computers, smart phones, etc. that can be integrated with or connected to monitors. In addition, one or more cameras and an eye tracking device (eg, an application that can be installed in the personal electronic device) are installed in the personal electronic device.

The gaze tracking device is a means capable of calculating the coordinates of the point where the learner's gaze is directed (hereinafter referred to as 'gaze coordinates') from an image acquired by a camera of a personal electronic device. The gaze tracking device may output, for example, the location of the gaze of the learner who is watching the display of the learner's personal electronic device (ie, the location of the gaze on the display being watched) as a coordinate value. In addition, when the learner's gaze is directed to a place other than the display, a relative position relative to the display may be output as a coordinate value.

According to an embodiment of the present invention, there is no particular limitation on the type of gaze tracking device or its gaze tracking method, and any means capable of generating the learner's eye gaze coordinates as an output from a camera image is sufficient. The eye tracking device may be, for example, ITU Gaze Tracker distributed by IT University of Denmark. ITU Gaze Tracker uses pupil and corneal light tracking for gaze tracking, and is a gaze tracking program developed as an open source project for building a low-cost eye tracking system. However, this is just an example, and may be a different type of eye tracking device that is currently commercially used and used, or a more improved device that is newly developed and used for eye tracking in the future.

The data collection unit 110 receives eye tracking data for each of the N learners participating in the online class. More specifically, the data collection unit 110 receives eye-tracking data acquired by the eye-gaze tracking device installed in the personal electronic device of each of the N learners from the personal electronic device. Here, the gaze tracking data refers to a set of temporal coordinate values for the position of the learner's gaze acquired by the gaze tracking device. That is, the eye tracking data indicates a series of coordinate values for gaze positions measured at predetermined time intervals, and the size of the measurement time interval is not particularly limited. In addition, the time interval does not always have to be constant during the course of the online class, and may change according to the progress of the class.

The calculation unit 120 calculates a gaze fixation point, which is a position where each learner's gaze is fixed, using the gaze tracking data collected by the data collection unit 120 . This is to determine the area the learner is learning, such as whether the learner is looking at the content and/or what part of the content the learner is looking at, through the gaze position. According to one aspect of the present embodiment, the calculation unit 120 calculates that the gaze coordinate values included in the eye tracking data are included within a predetermined area for a predetermined time (eg, a predetermined number of continuous gaze coordinate values). If there is, the corresponding area may be determined as the gaze fixation point.

2 is a schematic diagram for explaining an example of a method of calculating a fixed position of the gaze using gaze coordinate values. In FIG. 2 , numbers ① to ⑭ indicate positions of gaze fixation points sequentially collected by the data collection unit 110 .

Referring to FIG. 2, first, when four consecutive gaze coordinate values (① to ④) are included in a square of a predetermined size (eg, 30 × 30 pixel size) within a rectangular display, the region is set as a gaze fixation point. to decide Here, 4 is exemplary, and the number may vary. In addition, when the subsequent gaze coordinate values (⑤ to ⑭) are new positions outside the corresponding gaze fixation point, the gaze coordinate values (⑤) are added to the extended area (e.g., 50 × 50 pixel size) including the location of the current gaze fixation point. ~ ⑦, ⑩) are included, these new positions are merged with the current gaze fixation point so that the extended area becomes the new gaze fixation point. On the other hand, if the gaze coordinate values (⑧, ⑨, ⑪ to ⑭) are not included in the extended area (e.g., 50 × 50 pixel size) that includes the position of the current gaze fixation point, these new positions are ignored (⑧, ⑨) Or, if four consecutive gaze fixation points are located within a new square of a predetermined size, that is, a 30 × 30 pixel size, the new square is used as a criterion (⑪ to ⑭) for a new gaze fixation point.

According to one aspect of the present embodiment, the calculation unit 120 may also calculate the fixation time at each gaze fixation point using the gaze tracking data collected by the data collection unit 120 . The fixation time indicates the time the learner gazed at the corresponding fixation point (the content of the corresponding area), and can be used to determine the degree of attention or the learning state of the learner.

In addition, the fixation time is also helpful in identifying eye movement patterns. 3 is an example of a diagram schematically illustrating a movement pattern of a specific learner's gaze. Referring to FIG. 3 , in order to show the movement pattern of the gaze, the fixation time of the gaze at each gaze fixation point is indicated by the size of a circle. That is, the larger the size of the circle, the longer the fixation time.

The use of such eye movement patterns is to use the fact that the number of fixation points and fixation time among gaze information show a deep correlation with the learner's learning degree and cognitive load. Therefore, using the eye movement pattern, it can be used as an element to determine the learning state of the learner, and through this, the degree of visual attention of the learner to the corresponding area of the content can be grasped.

The class participation state determination unit 130 determines whether each of the learners normally participates in line with the progress of the class, using the distribution of gaze fixation points of the learners calculated by the calculation unit 120 .

Here, the 'distribution of gaze fixation points of learners' refers to the number of learners at each gaze fixation point obtained at an arbitrary time. More specifically, the calculation unit 120 calculates gaze fixation points for each of the N learners at a specific time. At this time, some or all of the gaze fixation points of each of the N learners may overlap, and the gaze fixation points of some learners may not overlap with other learners. Here, the 'distribution of gaze fixation points of learners' is statistical data indicating how many learners correspond to each of the gaze fixation points obtained from the calculation unit 120 .

As an example, the class participation state determining unit 130, when there are a relatively large number of learners having gaze fixation points in a specific area among a predetermined group of learners (eg, N learners), determines this area as the gaze area, Based on whether each learner's gaze fixation point corresponds to the gaze region, the class participation state of the learner may be determined. That is, if the gaze fixation point of a specific learner corresponds to the gaze area, it is determined that the learner is normally participating in the class, but if the gaze fixation point of the specific learner does not correspond to the gaze area, it may be determined that the learner does not normally participate in the class. In this case, it may be determined that a learner whose gaze fixation point is not calculated also does not normally participate in the class, but this is just an example.

According to an embodiment of the present invention, there is no particular limitation on the number of learners whose gaze area should be the gaze fixation point in order to determine the gaze area. For example, if more than half of all N learners gaze at the same area (that is, gaze fixation points coincide or overlap), or preferably 2/3 or more or 3/4 or more learners gaze at the same area , the corresponding region may be determined as the gaze region.

As another example, the class participation state determination unit 130 may also use the distribution of gaze fixation points of learners and the fixation time at the corresponding gaze fixation point to determine the class participation state. In this case, the class participation state determination unit 130 comprehensively considers whether the fixation point of each learner's gaze corresponds to the gaze region and the fixed time during which the gaze is fixed on the gaze region to determine the student's class participation state. You may.

As another example, the class participation state determination unit 130 may also use the distribution of gaze fixation points of the learners, the fixation time at the gaze fixation point, and its change over time to determine the class participation state. In this case, the class participation state determining unit 130 considers the movement of the learner's gaze, such as whether the fixation point of each learner's gaze corresponds to the gaze area, the fixation time when the gaze is fixed on the gaze area, and the change in the gaze fixation point. Thus, the class participation status of the student may be determined.

According to one aspect of the present embodiment, the class participation state determination unit 130 may also use online class content information displayed on the screen to determine the learner's class participation state. To this end, the class participation state determining unit 130 may receive content information about content used in online classes from the outside or collect content information from data displayed on the screen. Such content information may include, but is not limited to, subject, execution (lecture) time, content ID, and the like. In addition, when a plurality of contents are mixed and configured on one screen, each content information may be collected in units of individual areas (learning areas) of the screen. According to this, in determining the learner's class participation status, it is possible to determine the class participation status more adaptively according to the type of class subject or school age by also considering content information.

The class state participation determination unit 130 may be implemented using an online class participation detection model based on unsupervised learning artificial intelligence. According to this, since artificial intelligence technology is used to determine class participation, more effective and accurate judgment is possible. In addition, since unsupervised learning is used, there is no need to prepare a large amount of training data in advance for AI learning, and no prior knowledge for initialization is required, and only the central value of similar states is stored, so the size of the data is also small In addition, since the corresponding model can be updated with actually performed judgment data, more accurate judgment is possible as actual use continues.

In order to build an online class participation detection model based on unsupervised learning artificial intelligence, first, an unsupervised learning-based class participation map is created. The class participation map is a key algorithm that can generate various learning participation behaviors of learners participating in non-face-to-face online classes using unsupervised learning. To this end, a class participation pattern is first generated based on unsupervised learning, and unsupervised learning algorithms such as Self-Organizing Map (SOM) and Generative Adversarial Networks (GAN) may be used. According to this, a class participation map may be generated by first extracting feature values for candidate variables from the obtained data, and then clustering the extracted feature values according to the non-face-to-face online class participation behavior for each class of class. According to this, it is possible to reduce the cost of creating a database.

In the feature extraction step, feature values are extracted for candidate variables obtained from the collected data, such as the location of the gaze fixation point, gaze fixation time, content gaze frequency, and gaze fixation. value can be extracted. At this time, the calculated measured values are normalized to the range of [-1, 1] or [0, 1] because the ratio of the maximum value and the minimum value is different from each other and the value may be very large. As a result, the ratio of the maximum value to the minimum value of the normalized measurement values is standardized, and a measurement value outside the corresponding range can be determined as an error, which is helpful in removing data noise.

Then, redundancy removal and dimensionality reduction of feature values are performed. More specifically, since features unrelated to gaze movement features or highly correlated features may be mixed among the features, the dimension is reduced by removing redundancy of the selected candidate variables (features). At this time, in order to extract a core feature from among a plurality of features, a dimensionality reduction algorithm such as principal component analysis (PCA) may be used or a principal component analysis and feature selection algorithm may be combined.

In addition, by using an unsupervised learning algorithm including SOM or GAN, the extracted feature values are clustered according to the non-face-to-face online class participation behavior for each subject category, and an 'unsupervised learning-based class participation map' is created for reinforcement learning. Do it. During reinforcement learning, several episodes must be experienced. Since the target state does not transition to the next state at the end of one episode, it cannot be treated like other clustered states. By adding a cluster centered on the target state, The number of nodes constituting the 'unsupervised learning-based class participation map' is "the number of clusters × the number of possible actions + 1". Accuracy of judgment can be improved by continuously updating the corresponding model through comparison with the previously generated unsupervised learning-based class participation map by reinforcement learning.

The feedback unit 140 provides feedback such as a notification message to the learners based on the determination result of the class participation state determination unit 130 . For example, the feedback unit 140 may transmit a notification message including contents such as attention warning, current status, and penalty information only to the learner determined by the class participation state determination unit 130 to be not participating in the class. In this case, the notification message may be transmitted to a personal electronic device used by the learner for online classes and displayed in a pop-up form on the screen or transmitted to a mobile terminal such as a smart phone of the learner. Depending on the embodiment, the feedback unit 140 may transmit the judgment result of the class participation status determination unit 130 in the form of a predetermined message, such as a notification message, not only to individual learners but also to teachers (instructors) or parents.

The above description is only an example and should not be construed as being limited thereto. The technical idea of the present invention should be specified only by the invention described in the claims below, and all technical ideas within the equivalent range should be interpreted as being included in the scope of the present invention. Therefore, it is obvious to those skilled in the art that the above-described embodiment can be implemented in various forms.

Claims (7)

In the device for detecting whether a plurality of learners receiving non-face-to-face online classes are participating in accordance with the progress of the online class with a personal electronic device equipped with an eye tracking device,
a data collection unit receiving the eye-tracking data acquired by the eye-tracking device from the personal electronic device of each of the plurality of learners participating in the online class;
a calculation unit that calculates a gaze fixation point, which is a location where the gaze of each of the plurality of learners is fixed, using the gaze tracking data collected by the data collection unit; and
A class participation state determining unit for determining whether or not each of the plurality of learners participate in class using the distribution of the gaze fixation points of the plurality of learners calculated by the calculation unit;
The class participation state determining unit determines, among the gaze fixation points of each of the plurality of learners, the gaze fixation point corresponding to the largest number of learners as the gaze region, and then, among the plurality of learners, the gaze fixation point of a specific learner is located in the gaze region. Non-face-to-face online class participation detection device, characterized in that for determining whether the learner participates in the class based on whether it is included.
delete According to claim 1,
The calculation unit also calculates the fixation time at the gaze fixation point of each of the plurality of learners,
The non-face-to-face online class participation detection device, characterized in that the class participation state determination unit determines whether each of the plurality of learners participates in the class using the gaze movement information by the gaze fixation point and the fixation time.
According to claim 3,
The non-face-to-face online class participation, characterized in that the class participation status determination unit determines whether each of the plurality of learners participate in the class by using the content information of the online class, including the type and execution time of the class subject. detection device.
According to claim 4,
The non-face-to-face online class participation detection device, characterized in that the class participation state determination unit is implemented using an online class participation detection model based on unsupervised learning artificial intelligence.
According to claim 1,
Non-face-to-face online class participation detection device, characterized in that it further comprises a feedback unit for transmitting a notification message to the personal electronic device of the learner determined that the class participation state determination unit is not participating in the class.
A method for detecting whether a plurality of learners receiving a non-face-to-face online class are participating in accordance with the progress of the online class, using a personal electronic device equipped with an eye tracking device,
receiving eye-tracking data obtained by the eye-tracking device from the personal electronic device of each of the plurality of learners participating in the online class;
Calculating a gaze fixation point, which is a location where the gaze of each of the plurality of learners is fixed, using the gaze tracking data collected in the receiving step; and
Using the distribution of gaze fixation points of each of the plurality of learners calculated in the calculating step, determining whether each of the plurality of learners participates in class,
In the step of determining, the gaze fixation point corresponding to the largest number of learners among the gaze fixation points of each of the plurality of learners is determined as the gaze region, and then the gaze fixation point of a specific learner among the plurality of learners is included in the gaze region. Non-face-to-face online class participation detection method, characterized in that for determining whether the learner participates in the class based on whether or not

Images