KR20230090890A - Method for evaluating cognitive ability - Google Patents

Abstract

Disclosed is a method performed by a computing device including one or more processors according to some embodiments of the present disclosure, which may provide user-customized text data. The method may comprise: a step of acquiring an eye image by using a first pre-processing module that uses a face image as input, and extracting face information by using a second pre-processing module that uses the face image as input; a step of generating gaze position information of a user and gaze pattern information of the user from the eye image and the face information by using a gaze information generation module; a step of generating face movement information of the user from the eye image and the face information by using a face movement information generation module; a step of generating emotion state information of the user from the eye image and the face information by using an emotion state classification module; and a step of generating evaluation information on a cognitive ability of the user at least based on the gaze position information, the gaze pattern information, the face movement information, and the emotion state information.

Description

METHOD FOR Evaluating COGNITIVE ABILITY}

The present disclosure relates to artificial intelligence technology, and more particularly, to a method for evaluating cognitive abilities using artificial intelligence technology.

In modern society, it is easy to find many documents containing various information, which can be called a flood of information. Since the amount of such information is too vast and not systematically classified, it is necessary to precisely target and selectively extract only the necessary information in order to quickly and accurately acquire necessary information from the flood of such information.

In addition, with the development of multimedia technology, the learning ability of users continues to decline with the passage of time. Concomitantly, the reading comprehension and logical thinking ability of users familiar with digital devices and multimedia since childhood tend to decrease despite the necessity. Reading comprehension training that reads and accepts high-quality texts must be continuously performed across all age groups to prevent this sluggishness or deterioration of learning ability. need to be developed

Korean Patent Publication No. 10-2021-0089340

The present disclosure has been made in response to the above background art, and is intended to accurately evaluate a user's cognitive ability and/or reading ability in an efficient manner.

The technical problems of the present disclosure are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

According to some embodiments of the present disclosure for solving the above problems, a method performed by a computing device including at least one processor is disclosed. acquiring an eye image using a first pre-processing module that takes a face image as an input, and extracting face information using a second pre-processing module that takes the face image as an input; generating gaze position information of the user and gaze pattern information of the user from the eye image and the face information using a gaze information generation module; generating facial movement information of the user from the eye image and the face information by using a facial movement information generating module; generating emotional state information of the user from the eye image and the face information using an emotional state classification module; and generating evaluation information on the cognitive ability of the user based at least in part on the gaze position information, the gaze pattern information, the facial movement information, and the emotional state information.

Alternatively, the gaze information generation module extracts latent features for the eye image and face information using a convolutional neural network structure, and uses a multi-layer perceptron (MLP) structure to extract the user's latent features from the latent features. Generating location information of the gaze of the user, based on the location information of the user's gaze, calculating the movement speed and direction of the user's gaze, and based on the movement speed and direction of the user's gaze, Gaze pattern information may be generated.

Alternatively, the first pre-processing module, the second pre-processing module, the face motion information generating module, and the emotional state information generating module may include a convolutional neural network structure for extracting a latent feature and a module from the extracted latent feature. It may include a multi-layer perceptron (MLP) structure that generates an output.

Alternatively, the face image may be obtained in a process in which the user reads test text data to which a predetermined category and a predetermined level of difficulty are assigned.

Alternatively, the method may include: obtaining a memo image of the user generated in a process of the user reading the test text data; and acquiring memo location information and memo type information from the memo image using a third pre-processing module, wherein generating evaluation information on the user's cognitive ability includes the memo location information and the memo location information. It may additionally be based on shape information.

Alternatively, the generating of the evaluation information on the user's cognitive ability may include, using a first model, the gaze position information, the gaze pattern information, the memo position information, and the memo type information, the gaze and the memo generating first meta-information including comparison information between locations, gaze-retrogression information, handwriting analysis information, and memo quantitative information; and generating evaluation information on the cognitive ability of the user based on the first meta-information.

Alternatively, the first model measures distance information between the gaze position information and the memo position information in a corresponding time frame using a first rule-based module, thereby measuring distance information between the gaze position information and the memo position. calculating a degree of agreement and generating comparison information between the gaze and a memo position based on the calculated degree of agreement, and reading the test text data from the gaze position information by the user using a second rule-based module; The gaze retrogression information may be generated by determining a section in which the user's eye gaze is directed toward the rear of the test text data.

Alternatively, the first model generates the handwriting analysis information indicating precision of handwriting from the memo type information in a manner of obtaining self entropy for an output of the last layer of the first multi-layer perceptron, and the convolutional neural network Obtains a memo type latent feature from the memo type information, generates a matrix including the obtained memo type latent feature and memo size information, and inputs the generated matrix to the second multi-layer perceptron to obtain the memo Quantitative information can be generated.

Alternatively, the method may include: acquiring sensor data generated in a process of reading the test text data by the user; and generating a vibration pattern generated while the user reads the test text data from the sensor data using a fourth pre-processing module.

Alternatively, generating evaluation information on the user's cognitive ability may be additionally based on the vibration pattern.

Alternatively, the fourth pre-processing module converts the wave data extracted from the sensor data into at least one of a transformer structure to which an attention mechanism is applied, a convolutional neural network structure, and a rule-based structure including a low pass filter. A structure can be used to create the vibration pattern.

Alternatively, the generating of the evaluation information on the user's cognitive ability may include inputting the vibration pattern information into at least one of a transformer structure to which an attention mechanism is applied and a convolutional neural network structure, so that the test text data is stored in the test text data. generating a latent feature vector that reflects the user's emotional state and concentration during reading by the user; and generating evaluation information on the user's cognitive ability based on a latent feature vector that reflects the emotional state and concentration of the user in the process of reading the test text data.

Alternatively, the method may include: obtaining voice data uttered by the user in a process of the user reading the test text data; and generating time-based recording information from the voice data by using a fifth pre-processing module that converts the voice data into text data by combining the convolutional neural network structure and the transformer structure. Here, the generating of the evaluation information on the user's cognitive ability may be additionally based on the time-based recording information.

Alternatively, the generating of the evaluation information for the cognitive ability of the user may include comparing the recorded information for each time period with the test text data using a fourth model, displaying corresponding parts, and comparing the recorded information for each time period with the test text data. generating concentration information of the user by comparing the gaze position information with the gaze position information; and generating evaluation information on the user's cognitive ability based on the concentration information.

According to one embodiment of the present disclosure, a computer program stored in a computer readable storage medium is disclosed. The computer program, when executed by one or more processors, performs the following method, wherein the method acquires an eye image using a first preprocessing module that takes a face image as an input, and uses the face image as an input extracting face information using a second pre-processing module; generating gaze position information of the user and gaze pattern information of the user from the eye image and the face information using a gaze information generating module; generating facial movement information of the user from the eye image and the face information by using a facial movement information generating module; generating emotional state information of the user from the eye image and the face information using an emotional state classification module; and generating evaluation information on the cognitive ability of the user based at least in part on the gaze position information, the gaze pattern information, the facial movement information, and the emotional state information.

According to one embodiment of the present disclosure, a computing device including a storage unit and a processor is disclosed. The processor performs the following method, the method: acquiring an eye image using a first preprocessing module taking a face image as an input, and using a second preprocessing module taking the face image as an input extracting face information; generating gaze position information of the user and gaze pattern information of the user from the eye image and the face information using a gaze information generation module; generating facial movement information of the user from the eye image and the face information by using a facial movement information generating module; generating emotional state information of the user from the eye image and the face information using an emotional state classification module; and generating evaluation information on the cognitive ability of the user based at least in part on the gaze position information, the gaze pattern information, the facial movement information, and the emotional state information.

Techniques according to some embodiments of the present disclosure may generate test results for evaluating a user's cognitive ability and/or reading ability efficiently and accurately, and may further provide user-customized text data.

Effects obtainable in the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below. .

Various aspects are now described with reference to the drawings, wherein like reference numbers are used to collectively refer to like elements. In the following embodiments, for explanation purposes, numerous specific details are set forth in order to provide a thorough understanding of one or more aspects. However, it will be apparent that such aspect(s) may be practiced without these specific details.
1 illustratively illustrates a block diagram of a computing device in accordance with some embodiments of the present disclosure.
2 is a schematic diagram illustrating a network function according to some embodiments of the present disclosure.
3 illustratively illustrates a method for assessing cognitive ability and/or reading ability according to some embodiments of the present disclosure.
4 illustratively illustrates operation of a computing device for assessing cognitive ability and/or reading ability in accordance with some embodiments of the present disclosure.
5 illustratively illustrates operation of a computing device for assessing cognitive ability and/or reading ability in accordance with some embodiments of the present disclosure.
6 illustratively illustrates operation of a computing device for assessing cognitive ability and/or reading ability in accordance with some embodiments of the present disclosure.
7 depicts a simplified and general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

Various embodiments are now described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the present disclosure. However, it is apparent that these embodiments may be practiced without these specific details.

As used herein, the terms "unit", "module", "system", "component", and "model" and the like refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or the execution of software. , and they can be used interchangeably with each other. For example, "unit" can be, but is not limited to, a procedure running on a processor, a processor, an object, a thread of execution, a program, and/or a computer. For example, each of an application executed on the computing device and the computing device may correspond to “unit”. One or more components may reside within a processor and/or thread of execution. A component can be localized within a single computer. A component may be distributed between two or more computers. Also, these components can execute from various computer readable media having various data structures stored thereon. Components may be connected, for example, via signals with one or more packets of data (e.g., data and/or signals from one component interacting with another component in a local system, distributed system) to other systems and over a network such as the Internet. data being transmitted) may communicate via local and/or remote processes.

In addition, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or”. That is, unless otherwise specified or clear from the context, “X employs A or B” is intended to mean one of the natural inclusive substitutions. That is, X uses A; X uses B; Or, if X uses both A and B, "X uses either A or B" may apply to either of these cases. Also, the term "and/or" as used herein should be understood to refer to and include all possible combinations of one or more of the listed related items.

Also, the terms "comprises" and/or "comprising" should be understood to mean that the features and/or components are present. However, it should be understood that the terms "comprises" and/or "comprising" do not exclude the presence or addition of one or more other features, elements, and/or groups thereof. Also, unless otherwise specified or where the context clearly indicates that a singular form is indicated, the singular in this specification and claims should generally be construed to mean "one or more".

In addition, the term “at least one of A or B” should be interpreted as meaning “when only A is included”, “when only B is included”, and “when A and B are combined”.

Those skilled in the art will further understand that the various illustrative logical blocks, components, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented using electronic hardware, computer software, or combinations of both. It should be recognized that it can be implemented as To clearly illustrate the interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented in hardware or as software depends on the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of this disclosure.

The description of the presented embodiments is provided to enable any person skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art of this disclosure. The general principles defined herein may be applied to other embodiments without departing from the scope of this disclosure. Thus, the present invention is not limited to the embodiments presented herein. The present invention is to be accorded the widest scope consistent with the principles and novel features set forth herein.

1 illustratively illustrates a block diagram of a computing device in accordance with some embodiments of the present disclosure.

The computing device 100 may perform a method for evaluating a user's reading ability and/or cognitive ability based on an artificial neural network according to an embodiment of the present disclosure.

The configuration of the computing device 100 shown in FIG. 1 is only a simplified example. In one embodiment of the present disclosure, the computing device 100 may include other components for performing a computing environment of the computing device 100, and only some of the disclosed components may constitute the computing device 100. In addition, although the computing device 100 is represented as one entity in FIG. 1 , the method according to an embodiment of the present disclosure may be implemented by a plurality of computing devices 100 . For example, a method according to an embodiment of the present disclosure may be implemented based on processing and communication between the computing device 100 including a server and the computing device 100 including a user terminal.

The computing device 100 in the present disclosure may refer to any type of device having processing capability. As an example, the computing device 100 may be a server and/or a user terminal.

Computing device 100 may include any type of computer system or computer device, such as, for example, a microprocessor, mainframe computer, digital processor, portable device or device controller, and the like. This computing device 100 may include any type of server and/or user terminal. For example, the user terminal may include a PC (personal computer), a notebook (note book), a mobile terminal, a smart phone (smart phone), a tablet PC (tablet pc), etc. owned by the user, , can include all types of terminals that can access wired/wireless networks.

The computing device 100 may include a processor 110 , a storage unit 120 , and a communication unit 130 .

The computing device 100 of the present disclosure may obtain or generate outputs according to an embodiment of the present disclosure from a face image, a memo image, sensor data, and voice data using preprocessing modules.

A preprocessing module in the present disclosure may refer to any type of module for preprocessing data to be input to the models (eg, at least one of the first to fifth models) of the present disclosure. These pre-processing modules may include artificial intelligence-based models and/or rule-based models.

The computing device 100 of the present disclosure may generate evaluation information on the cognitive ability (or reading ability) of the user using one or more models (eg, at least one of the first to fifth models). there is.

The computing device 100 may select or generate recommendation text data to be provided to the user using a recommendation data generation model. The recommended text data may include text data having a level of difficulty or readability suitable for the user's level and/or including a subject capable of increasing the user's satisfaction.

The processor 110 may include one or more cores, and includes a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU) of a computing device. unit), or any type of processor for data processing, data analysis, and/or deep learning. The processor 110 may read the computer program stored in the storage unit 120 and perform the text data processing method according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor 110 may perform an operation for learning a neural network. The processor 110 processes input data for learning in machine learning or deep learning (DL: preprocessing such as denoising, tokenization, OCR, etc.), feature extraction from input data, Calculations for neural network learning, such as error calculation and weight update of the neural network using backpropagation, may be performed. At least one of the CPU, GPGPU, and TPU of the processor 110 may process learning of the network function. For example, the CPU and GPGPU can process learning of network functions and classification and generation of data using network functions. In addition, in an embodiment of the present disclosure, the learning of a network function and the classification and generation of data using the network function may be processed by using processors of a plurality of computing devices together. In addition, the computer program executed in the computing device 100 according to an embodiment of the present disclosure may be a CPU, GPGPU or TPU executable program.

According to an embodiment of the present disclosure, the storage unit 120 may store any type of information generated or determined by the processor 110 and any type of information received by the communication unit 130 . In the present disclosure, the storage unit 120 may be used interchangeably with a memory. For example, the storage unit 120 may store preprocessing modules and/or models used in the present disclosure. Also, the storage unit 120 may store evaluation information and/or may store text data for generating recommendation data based on the evaluation information.

According to an embodiment of the present disclosure, the storage unit 120 may be a flash memory type, a hard disk type, a multimedia card micro type, or a card type memory (eg SD or XD memory, etc.), RAM (Random Access Memory, RAM), SRAM (Static Random Access Memory), ROM (Read-Only Memory, ROM), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Memory) Read-Only Memory), a magnetic memory, a magnetic disk, and an optical disk may include at least one type of storage medium. The computing device 100 may operate in relation to a web storage that performs the storage function of the storage unit 120 on the Internet. The above description of the storage unit 120 is only an example, and the present disclosure is not limited thereto.

In FIG. 1, the storage unit 120 is shown as being included in the computing device 100, but according to an implementation aspect, the storage unit 120 exists outside the computing device 100 and communicates with the computing device 100. You can also send and receive data.

The communication unit 130 according to an embodiment of the present disclosure may communicate with other computing devices of any type through a network. A network according to an embodiment of the present disclosure may include any wired/wireless communication network capable of transmitting and receiving data and signals of any type, and may be included in the network represented in the present disclosure.

2 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure.

Throughout this specification, artificial intelligence-based models, artificial intelligence models, computational models, neural networks, network functions, and neural networks may be used interchangeably. For example, the first to fifth models in the present disclosure may belong to the category of artificial intelligence-based models as above. As another example, at least one of the preprocessing modules in the present disclosure may also belong to the category of the above AI-based model.

A neural network may consist of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network includes one or more nodes. Nodes (or neurons) constituting neural networks may be interconnected by one or more links.

Before input data is input to an input layer of an artificial intelligence model, preprocessing techniques according to an embodiment of the present disclosure may be performed. For example, at least one of the first to fifth preprocessing modules may output desired result data from a face image, memo image, sensor data, and/or voice data by performing a preprocessing operation.

In a neural network, one or more nodes connected through a link may form a relative relationship of an input node and an output node. The concept of an input node and an output node is relative, and any node in an output node relationship with one node may have an input node relationship with another node, and vice versa. As described above, an input node to output node relationship may be created around a link. More than one output node can be connected to one input node through a link, and vice versa.

In a relationship between an input node and an output node connected through one link, the value of data of the output node may be determined based on data input to the input node. Here, a link interconnecting an input node and an output node may have a weight. The weight may be variable, and may be changed by a user or an algorithm in order to perform a function desired by the neural network. For example, when one or more input nodes are interconnected by respective links to one output node, the output node is set to a link corresponding to values input to input nodes connected to the output node and respective input nodes. An output node value may be determined based on the weight.

As described above, in the neural network, one or more nodes are interconnected through one or more links to form an input node and output node relationship in the neural network. Characteristics of the neural network may be determined according to the number of nodes and links in the neural network, an association between the nodes and links, and a weight value assigned to each link. For example, when there are two neural networks having the same number of nodes and links and different weight values of the links, the two neural networks may be recognized as different from each other.

A neural network may be composed of a set of one or more nodes. A subset of nodes constituting a neural network may constitute a layer. Some of the nodes constituting the neural network may form one layer based on distances from the first input node. For example, a set of nodes having a distance of n from the first input node may constitute n layers. The distance from the first input node may be defined by the minimum number of links that must be passed through to reach the corresponding node from the first input node. However, the definition of such a layer is arbitrary for explanation, and the order of a layer in a neural network may be defined in a method different from the above. For example, a layer of nodes may be defined by a distance from a final output node.

An initial input node may refer to one or more nodes to which data is directly input without going through a link in relation to other nodes among nodes in the neural network. Alternatively, in a relationship between nodes based on a link in a neural network, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the neural network. Also, the hidden node may refer to nodes constituting the neural network other than the first input node and the last output node.

In the neural network according to an embodiment of the present disclosure, the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as the number of nodes progresses from the input layer to the hidden layer. can In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes of the input layer may be less than the number of nodes of the output layer and the number of nodes decreases as the number of nodes increases from the input layer to the hidden layer. there is. In addition, the neural network according to another embodiment of the present disclosure is a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as the number of nodes increases from the input layer to the hidden layer. can A neural network according to another embodiment of the present disclosure may be a neural network in the form of a combination of the aforementioned neural networks.

A deep neural network (DNN) may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer. Deep neural networks can reveal latent structures in data. That is, pictures, texts, videos, voices, textual data of voices, subliminal structure of music (e.g. what objects are in a picture, what is the content and emotion of the text, what is the content and emotion of the voice, etc. ) can be identified. Deep neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), auto encoders, generative adversarial networks (GANs), and restricted boltzmann machines (RBMs). machine), a deep belief network (DBN), a Q network, a U network, a Siamese network, a Generative Adversarial Network (GAN), and the like. The description of the deep neural network described above is only an example, and the present disclosure is not limited thereto.

For example, the artificial intelligence-based model in one embodiment of the present disclosure may include, for example, a multi-perceptron-based artificial intelligence model. A multi-perceptron-based model may include an input layer, hidden layers, and an output layer. For example, a multi-perceptron-based model may include an input layer including input neurons capable of receiving linguistic features and latent features, a plurality of hidden layers, and an output layer outputting results.

In one embodiment of the present disclosure, the network function may include an autoencoder. An autoencoder may be a type of artificial neural network for outputting output data similar to input data. An auto-encoder may include at least one hidden layer, and an odd number of hidden layers may be disposed between input and output layers. The number of nodes of each layer may be reduced from the number of nodes of the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically with the reduction from the bottleneck layer to the output layer (symmetrical to the input layer). Autoencoders can perform non-linear dimensionality reduction. The number of input layers and output layers may correspond to dimensions after preprocessing of input data. In the auto-encoder structure, the number of hidden layer nodes included in the encoder may decrease as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes located between the encoder and decoder) is too small, a sufficient amount of information may not be conveyed, so more than a certain number (e.g., more than half of the input layer, etc.) ) may be maintained.

The neural network may be trained using at least one of supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Learning of the neural network may be a process of applying knowledge for the neural network to perform a specific operation to the neural network.

A neural network can be trained in a way that minimizes output errors. In the learning of the neural network, the learning data is repeatedly input into the neural network, the output of the neural network for the training data and the error of the target are calculated, and the error of the neural network is transferred from the output layer of the neural network to the input layer in the direction of reducing the error. It is a process of updating the weight of each node of the neural network by backpropagating in the same direction. In the case of teacher learning, the learning data in which the correct answer is labeled is used for each learning data (ie, the labeled learning data), and in the case of comparative teacher learning, the correct answer may not be labeled in each learning data. That is, for example, learning data in the case of teacher learning regarding data classification may be data in which each learning data is labeled with a category. Labeled training data is input to a neural network, and an error may be calculated by comparing an output (category) of the neural network and a label of the training data. As another example, in the case of comparative history learning for data classification, an error may be calculated by comparing input learning data with a neural network output. The calculated error is back-propagated in a reverse direction (ie, from the output layer to the input layer) in the neural network, and the connection weight of each node of each layer of the neural network may be updated according to the back-propagation. The amount of change in the connection weight of each updated node may be determined according to a learning rate. The neural network's computation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently according to the number of iterations of the learning cycle of the neural network. For example, a high learning rate may be used in the early stage of neural network training to increase efficiency by allowing the neural network to quickly obtain a certain level of performance, and a low learning rate may be used in the late stage to increase accuracy.

In neural network learning, generally, training data can be a subset of real data (ie, data to be processed using the trained neural network), and therefore, errors for training data are reduced, but errors for real data are reduced. There may be incremental learning cycles. Overfitting is a phenomenon in which errors for actual data increase due to excessive learning on training data. For example, a phenomenon in which a neural network that has learned a cat by showing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing the error of machine learning algorithms. Various optimization methods can be used to prevent such overfitting. To prevent overfitting, methods such as increasing the training data, regularization, inactivating some nodes in the network during learning, and using a batch normalization layer should be applied. can

A computer readable medium storing a data structure according to an embodiment of the present disclosure is disclosed.

Data structure can refer to the organization, management, and storage of data that enables efficient access and modification of data. Data structure may refer to the organization of data to solve a specific problem (eg, data retrieval, data storage, data modification in the shortest time). A data structure may be defined as a physical or logical relationship between data elements designed to support a specific data processing function. A logical relationship between data elements may include a connection relationship between user-defined data elements. A physical relationship between data elements may include an actual relationship between data elements physically stored in a computer-readable storage medium (eg, a persistent storage device). The data structure may specifically include a set of data, a relationship between data, and a function or command applicable to the data. Through an effectively designed data structure, a computing device can perform calculations while using minimal resources of the computing device. Specifically, the computing device can increase the efficiency of operation, reading, insertion, deletion, comparison, exchange, and search through an effectively designed data structure.

The data structure can be divided into a linear data structure and a non-linear data structure according to the shape of the data structure. A linear data structure may be a structure in which only one data is connected after one data. Linear data structures may include lists, stacks, queues, and decks. A list may refer to a series of data sets in which order exists internally. The list may include a linked list. A linked list may be a data structure in which data are connected in such a way that each data is connected in a single line with a pointer. In a linked list, a pointer can contain information about connection to the next or previous data. A linked list can be expressed as a singly linked list, a doubly linked list, or a circular linked list depending on the form. A stack can be a data enumeration structure that allows limited access to data. A stack can be a linear data structure in which data can be processed (eg, inserted or deleted) at only one end of the data structure. The data stored in the stack may be a LIFO-Last in First Out (Last in First Out) data structure. A queue is a data listing structure that allows limited access to data, and unlike a stack, it can be a data structure (FIFO-First in First Out) in which data stored later comes out later. A deck can be a data structure that can handle data from either end of the data structure.

The nonlinear data structure may be a structure in which a plurality of data are connected after one data. The non-linear data structure may include a graph data structure. A graph data structure can be defined as a vertex and an edge, and an edge can include a line connecting two different vertices. A graph data structure may include a tree data structure. The tree data structure may be a data structure in which one path connects two different vertices among a plurality of vertices included in the tree. That is, it may be a data structure that does not form a loop in a graph data structure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. Hereinafter, a neural network is unified and described. The data structure may include a neural network. And the data structure including the neural network may be stored in a computer readable medium. The data structure including the neural network may also include preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation function associated with each node or layer of the neural network, and neural network It may include a loss function for learning of . A data structure including a neural network may include any of the components described above. That is, the data structure including the neural network includes preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation function associated with each node or layer of the neural network, and neural network. It may be configured to include all or any combination thereof, such as a loss function for learning of . In addition to the foregoing configurations, the data structure comprising the neural network may include any other information that determines the characteristics of the neural network. In addition, the data structure may include all types of data used or generated in the computational process of the neural network, but is not limited to the above. A computer readable medium may include a computer readable recording medium and/or a computer readable transmission medium. A neural network may consist of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network includes one or more nodes.

The data structure may include data input to the neural network. For example, the data structure may include data obtained as a result of one-hot encoding. A data structure including data input to the neural network may be stored in a computer readable medium. Data input to the neural network may include training data input during a neural network learning process and/or input data input to a neural network that has been trained. Data input to the neural network may include pre-processed data and/or data subject to pre-processing. Pre-processing may include a data processing process for inputting data to a neural network. Accordingly, the data structure may include data subject to pre-processing and data generated by pre-processing. The foregoing data structure is only an example, and the present disclosure is not limited thereto.

The data structure may include the weights of the neural network. (In this specification, weights and parameters may be used in the same meaning.) Also, a data structure including weights of a neural network may be stored in a computer readable medium. A neural network may include a plurality of weights. The weight may be variable, and may be changed by a user or an algorithm in order to perform a function desired by the neural network. For example, when one or more input nodes are interconnected by respective links to one output node, the output node is set to a link corresponding to values input to input nodes connected to the output node and respective input nodes. A data value output from an output node may be determined based on the weight. The foregoing data structure is only an example, and the present disclosure is not limited thereto.

As a non-limiting example, the weights may include weights that are varied during neural network training and/or weights for which neural network training has been completed. The variable weight in the neural network learning process may include a weight at the time the learning cycle starts and/or a variable weight during the learning cycle. The weights for which neural network learning has been completed may include weights for which learning cycles have been completed. Accordingly, the data structure including the weights of the neural network may include a data structure including weights that are variable during the neural network learning process and/or weights for which neural network learning is completed. Therefore, it is assumed that the above-described weights and/or combinations of weights are included in the data structure including the weights of the neural network. The foregoing data structure is only an example, and the present disclosure is not limited thereto.

The data structure including the weights of the neural network may be stored in a computer readable storage medium (eg, a memory or a hard disk) after going through a serialization process. Serialization can be the process of converting a data structure into a form that can be stored on the same or another computing device and later reconstructed and used. A computing device may serialize data structures to transmit and receive data over a network. The data structure including the weights of the serialized neural network may be reconstructed on the same computing device or another computing device through deserialization. The data structure including the weights of the neural network is not limited to serialization. Furthermore, the data structure including the weights of the neural network is a data structure for increasing the efficiency of operation while minimizing the resource of the computing device (for example, B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree). The foregoing is only an example, and the present disclosure is not limited thereto.

The data structure may include hyper-parameters of the neural network. Also, the data structure including the hyperparameters of the neural network may be stored in a computer readable medium. A hyperparameter may be a variable variable by a user. Hyperparameters include, for example, learning rate, cost function, number of learning cycle iterations, weight initialization (eg, setting the range of weight values to be targeted for weight initialization), hidden unit number (eg, the number of hidden layers and the number of nodes in the hidden layer). The foregoing data structure is only an example, and the present disclosure is not limited thereto.

3 illustratively illustrates a method for performing an evaluation of a user's cognitive ability and/or reading ability according to some embodiments of the present disclosure.

The flowchart shown in FIG. 3 is shown for illustrative purposes, and additional steps may be present or some of the steps included in the flowchart shown in FIG. 3 may be omitted.

The method of FIG. 3 may be performed by, for example, the computing device 100 of the present disclosure.

Cognitive ability in the present disclosure is the ability to discriminate a specific state, and may mean, for example, the ability to grasp the contents of a corresponding text in the process of reading specific text data. For example, cognitive ability can be used interchangeably with reading ability.

In one embodiment, the computing device 100 may acquire an eye image using a first pre-processing module that takes a face image as an input, and extracts face information using a second pre-processing module that takes the face image as an input. Yes (310). For example, the computing device 100 may generate text data to be read by a user. For example, the face image may be acquired while the user reads test text data to which a predetermined category and a predetermined level of difficulty are assigned.

In one embodiment, the computing device 100 may obtain a photographed image of the user's face information while the user reads the text data. For example, a user may receive text data assigned to a specific category such as politics, economy, culture, etc. and a specific level of difficulty such as first grade of elementary school, middle school, high school, and educated adult through a smart device such as a smartphone or tablet. can While the user reads the text data or solves a problem with the text data, the user's face image, voice data, memo information, vibration pattern of the smart device, gaze information, etc. are transferred to a computing device (smart device or server, etc.) can be obtained by

In one embodiment, the first pre-processing module may include an image segmentation model trained to take a face image as an input and output an eye image. The second pre-processing module may include an image segmentation model learned to output face information (eg, face features) by taking a face image as an input. As an example, the above-described image segmentation model may include an instance segmentation model and/or a semantic segmentation model.

In another embodiment, the first pre-processing module may include an object detection model in an image trained to take a face image as an input and output an eye image. The second preprocessing module may include an object feature extraction model in an image learned to output face information (eg, face features) by taking a face image as an input.

In an embodiment, the computing device 100 may generate the user's gaze position information and the user's gaze pattern information from the eye image and face information using the gaze information generation module (320). The computing device 100 may determine the movement speed and direction of the gaze from the eye image and face information. Determining the movement speed and direction of the gaze from the eye image and face information may be implemented using at least one of a rule-based module and a convolutional neural network. For example, the computing device 100 may track location information and pattern information of the user's gaze based on the obtained eye angle information of the eye image and the obtained angle information of the face image. For example, the computing device 100 may determine the position of the gaze and the movement speed of the gaze by acquiring the movement of the eyeball on the eye image. For example, gaze position information may be determined based on an angle between the eyes and/or the face and a screen on which text data is displayed. This gaze position information may refer to information indicating which part of the text data the user is currently gazing at (or reading). Gaze pattern information refers to information representing a pattern of reading text data of a user. The gaze pattern information includes, for example, a pattern of reading left and right twice, a pattern of reading alternately up and down, a pattern of reading while moving from left to right and top to bottom, reading time for specific text data and/or text data It may include a pattern that reads back to the back (the direction of the past or previous text data) and the like.

Eye-gaze tracking in an embodiment of the present disclosure generates an eye image dividing an eye part of a user's face from an image of the user, creates pupil information on the eye image, and generates the eye image of the user from the image of the user. Face information is extracted, and face location information (eg, the center point of the face), distance information between the face and the screen (eg, face size), and face rotation information (eg, face according to eye, nose, and mouth positions) are extracted. Rotation information) may be implemented in a manner of obtaining. Based on the obtained information, the computing device 100 may perform deep learning-based user gaze tracking. Deep learning-based user eye tracking is to track the user's eye gaze by inputting the user's pupil information, face position information, distance information between the face and the screen, and/or face rotation information, and to determine the degree of interest in a specific text and/or It may also mean outputting results related to comprehension.

In an embodiment, the computing device 100 may generate facial movement information of the user from the eye image and face information using the facial movement information generating module (330). For example, the computing device 100 may estimate the movement of the user's facial muscles and/or the resulting emotional state by determining positions and changes of facial feature points using the facial movement information generating module.

In an embodiment, the computing device 100 may generate emotional state information of the user from the eye image and face information using the emotional state classification module (340). The emotional state information may refer to the user's emotion on the test text data expressed in the face and/or eyes while the user reads the test text data or solves a problem related thereto. For example, the emotional state information may include comfort, difficulty, difficulty, frustration, and/or enjoyment.

In one embodiment, the computing device 100 may generate evaluation information for the user's cognitive ability based at least in part on gaze position information, gaze pattern information, facial movement information, and/or emotional state information (350). .

Evaluation information on cognitive ability includes, for example, an evaluation of high interest in text data belonging to a specific category, an evaluation of feeling comfortable with text data belonging to a specific difficulty level, and anxiety when solving problems with text data belonging to a specific difficulty level. and/or an evaluation of feeling difficulty in reading text data belonging to a specific level of difficulty. As another example, the evaluation information on the cognitive ability may include information indicating that the corresponding user belongs to a category in the field of science and technology and is a suitable user for text data having a level of difficulty at the level of a sophomore in high school.

In a further embodiment of the present disclosure, the computing device 100 obtains a user's memo image generated in the process of the user reading test text data, and obtains memo location information and memo location information from the memo image using a third pre-processing module. Memo type information may be obtained. The computing device 100 may obtain memo information about the user in a process of reading the user's test text data. For example, in a process of reading test text data through a smart device, a user interface may be created so that a user can leave a memo on the text data. Based on the memo information input by the user, the computing device 100 may generate evaluation information on the user's cognitive ability. For example, evaluation information on cognitive ability may be generated based on memo location information and memo type information. The computing device 100 performs comparison between the gaze position information and the memo position information, generates gaze retrograde information, generates handwriting analysis information about the memo, and generates quantitative information about the memo, thereby increasing the cognitive ability of the user. Evaluation information can be created.

In a further embodiment of the present disclosure, the computing device 100 may obtain sensor data generated in the course of the user reading the test text data. For example, the computing device 100 may obtain sensor data including shaking information, hand shaking information, and/or movement information of the smart device being manipulated by the user. The computing device 100 may generate a vibration pattern generated while the user reads the test text data from the sensor data by using the fourth preprocessing module. A user's emotional state and/or concentration state may be determined based on the vibration pattern obtained from the sensor data, and accordingly, the computing device 100 may evaluate the user's cognitive ability for the test text data.

In a further embodiment of the present disclosure, the computing device 100 may obtain voice data uttered by the user while the user reads the test text data. In one embodiment, for example, a user interface may be created so that a user can leave voice data for text data in a process of reading test text data through a smart device. In this case, the user can interact with the test text data in the form of reading the test text data currently being viewed by voice. The computing device 100 may generate concentration information for the test text data for the user by comparing the test text data with the recorded information for each time period. The computing device may evaluate the user's cognitive ability based on the concentration information.

4 illustratively illustrates operation of a computing device for assessing cognitive ability and/or reading ability in accordance with some embodiments of the present disclosure.

As shown in FIG. 4 , the computing device 100 may acquire a facial image 410 captured during a process of reading the user's test text data. Computing device 100 generates an eye image 430 from first pre-processing module 420A by inputting face image 410 to first pre-processing module 420A and second pre-processing module 420B and second pre-processing module 420B. Facial information 440 may be generated from module 420B.

In one embodiment, the first pre-processing module 420A and the second pre-processing module 420B have a convolutional neural network structure for extracting latent features and a multi-layer perceptron (MLP) structure for generating module outputs from the extracted latent features. It can be done. For example, the first pre-processing module 420A may extract feature points corresponding to eyes from a face image (or an upper body image, etc.) obtained while a user solves a problem. The first preprocessing module 420A may detect or segment an eye or a region corresponding to the eyeball in the image. For example, the second pre-processing module 420B may extract feature points corresponding to the face region from a face image (or upper body image, etc.) obtained while the user solves the problem. The second pre-processing module 420B may detect or segment a face or a region corresponding to the face in the image. As described above, the technique according to an embodiment of the present disclosure parallelly inputs image information obtained in the process of reading text data to a plurality of pre-processing modules, thereby obtaining eye information and face information from the same image information. information can be obtained. Accordingly, gaze tracking information for text data of the user may be generated in a more accurate manner through matching between the eyes and the face.

In one embodiment of the present disclosure, the eye image 430 and the face information 440 are input to the gaze information generating module 450A, the facial movement information generating module 450B, and the emotional state classification module 450C, respectively. Gaze position information and gaze pattern information 460 , facial muscle movement information 470 , and emotional state information 480 may be output by the modules 450A, 450B, and 450C. As described above, when a face image obtained in the process of reading text data is input, the user's gaze position information, gaze pattern information, facial muscle movement information, and emotional state through modules according to an embodiment of the present disclosure. information can be obtained. User-customized evaluation information corresponding to text data can be generated in a more accurate manner through such gaze position information, gaze pattern information, facial muscle movement information, and emotional state information.

In one embodiment of the present disclosure, the convolutional neural network structure and the multilayer perceptron structure may be operated in a form connected to each other in the gaze information generation module 450A, the face movement information generation module 450B, and the emotional state classification module 450C. there is. The convolutional neural network architecture may generate latent features for determining gaze positions/patterns, latent features for determining facial muscle movements, and/or latent features for determining emotional states. The multilayer perceptron structure can output gaze position information, gaze pattern information, facial muscle movement information, and emotional state information based on latent features. For example, the gaze information generation module 450A may further include a rule-based module and/or a multi-layered convolutional neural network that generates a gaze position and/or pattern based on the movement speed and direction of the gaze. For example, the computing device 100 may track location information and pattern information of the user's gaze based on the obtained eye angle information of the eye image and the obtained angle information of the face image. For example, the computing device 100 may determine the position of the gaze and the movement speed of the gaze by acquiring the movement of the eyeball on the eye image. For example, gaze position information may be determined based on an angle between the eyes and/or the face and a screen on which text data is displayed. This gaze position information may refer to information indicating which part of the text data the user is currently gazing at (or reading). Gaze pattern information refers to information indicating a pattern of reading text data of a user. The gaze pattern information includes, for example, a pattern of reading left and right twice, a pattern of reading alternately up and down, a pattern of reading while moving from left to right and top to bottom, reading time for specific text data and/or text data It may include a pattern that is read again in the backward or reverse direction of . Backward in the present disclosure may mean a direction to the past or previously read text data. Backward in the present disclosure may mean backward for text data.

Through modules according to an embodiment of the present disclosure, gaze position information, gaze pattern information, facial muscle movement information, and emotional state information of a user may be obtained in an efficient manner through an input of a face image. User-customized evaluation information corresponding to text data can be generated in a more accurate manner through such gaze position information, gaze pattern information, facial muscle movement information, and emotional state information. That is, from the user's point of view, just by reading the output text data, it can be automatically determined whether it is text data that the user can be interested in and/or whether the text data is suitable for the user and the level of difficulty. Evaluation information may be generated.

The modules shown in FIG. 4 may be integrated into the computing device 100 and operated, or at least some of the modules may exist outside the computing device 100 to exchange data while communicating with the computing device 100. . Additionally, an implementation method in which some of the modules shown in FIG. 4 may be performed in a user terminal and the other parts are performed in a server is also possible.

5 illustratively illustrates operation of a computing device for assessing cognitive ability and/or reading ability in accordance with some embodiments of the present disclosure.

The computing device 100 may include a third pre-processing module 540A that receives the memo image 510 and outputs memo location information and memo type information 550 . The computing device 100 may include a fourth preprocessing module 540B that receives sensor data 520 and outputs vibration pattern information 560 . The computing device 100 may include a fifth pre-processing module 540C that receives voice data 530 and outputs time-based recording information 570 .

The computing device 100 may obtain the memo image of the user generated while the user reads the test text data. The computing device 100 may allow a user to leave a memo on the text data in the process of solving the text data. The computing device 100 may perform a more accurate evaluation of the cognitive ability of the user by additionally considering the memo left in the process of solving the text data.

In an embodiment, the computing device 100 may obtain memo location information and memo type information 550 from the memo image 510 using the third pre-processing module 540A. From the memo image 510, location information on the smart device where the memo was left, location information between memos, location information between the memo and the text, comparison information between the time of writing the memo and the time of reading the text, the accuracy and/or clarity of the handwriting of the memo Memo type information indicating , etc. may be obtained.

In one embodiment, the computing device 100 may generate text data corresponding to the memo image 510 using the third pre-processing module 540A. In this case, the third pre-processing module 540A may include an optical character recognition (OCR) function. The computing device 100 may evaluate the user's concentration or the user's cognitive ability based on the degree of correspondence or discrepancy between the text data extracted from the memo image 510 and the test text data read by the user.

The third preprocessing module 540A may refer to a model in which a convolutional neural network and a multilayer perceptron are connected. In the convolutional neural network, latent features of the memo image are extracted, and in the multilayer perceptron, location information of the memo, thickness information of the memo, and/or memo shape information may be output based on the latent features. In one embodiment, evaluation information on the user's cognitive ability may be determined based on memo location information and memo type information.

As shown in FIG. 5 , the fourth preprocessing module 540B may receive sensor data 520 and output a vibration pattern 560 . The sensor data 520 may include hand shake information generated while the user reads the test text data and/or location movement information of the smart device. A user's state (eg, emotional state, etc.) in the process of reading text data may be estimated through the sensor data 520 . The fourth pre-processing module 540B uses at least one of a transformer structure to which an attention mechanism is applied, a convolutional neural network structure, and a rule-based structure including a low pass filter for the wave data extracted from sensor data. Thus, a vibration pattern can be created. Wave information on hand shake or wave information on smart phone shake is extracted from the sensor data, and a vibration pattern may be generated from the extracted wave information. The vibration pattern may include a pattern (hand-shake speed, hand-shake intensity, hand-shake time, motion strength, motion speed, motion time, etc. of a smart device) for describing any type of vibration.

As shown in FIG. 5 , the fifth pre-processing module 540C may generate time-based recording information 570 from voice data 530 . The voice data 530 may refer to voice data 530 uttered by the user in response to text data currently read by the user in the process of reading the text data. The computing device 100 obtains voice data 530 uttered by the user while the user reads the test text data, and converts the voice data 530 to text data by combining the convolutional neural network structure and the transformer structure. Time-based recording information 570 can be generated from the voice data 530 using the fifth pre-processing module 540C that converts to .

The modules 540A, 540B, and 540C shown in FIG. 5 may be integrated into the computing device 100 and operated, or at least some of the modules 540A, 540B, and 540C exist outside the computing device 100. Thus, data can be exchanged while communicating with the computing device 100 . Additionally, an implementation method in which some of the modules 540A, 540B, and 540C shown in FIG. 5 may be performed in a user terminal and other parts are performed in a server is also possible.

As described above, a technique for evaluating a user's cognitive ability according to an embodiment of the present disclosure includes a memo image 510, sensor data 520, and voice data obtained in a process in which a user reads test text data. By further utilizing 530, customized cognitive ability evaluation for the user's test text data can be made.

6 illustratively illustrates operation of a computing device for assessing cognitive ability and/or reading ability in accordance with some embodiments of the present disclosure.

The computing device 100 receives gaze position information and gaze pattern information 460 and memo position information and memo shape information 550, and receives information on the degree of coincidence between the gaze and the memo position 620, information on the number of gaze retractions 630, and handwriting. It may include a first model 610A that outputs precision information 640, memo method information, and memo quantity information 650.

The computing device 100 uses the first model 610A to determine the relationship between the gaze position and the memo position from the gaze position information 460, the gaze pattern information 460, the memo position information 550, and the memo type information 550. First meta information including comparison information 620 , gaze retraction information 630 , handwriting analysis information 640 , and memo quantity information 650 may be generated. The first meta information is factors extracted based on the user's face image and memo image and can be used to evaluate the user's cognitive ability. The computing device 100 may generate evaluation information on the user's cognitive ability by using the fifth model 610E based on the first meta-information.

In one embodiment, the first model 610A measures the distance information between the gaze position information 460 and the memo position information 550 in the corresponding time frame using the first rule-based module, thereby determining the position of the gaze. The degree of agreement 620 between the information and the location of the memo may be calculated. For example, each time frame may be assigned to the gaze position information 460 and the memo position information 550 . The gaze position information 460 to which the corresponding time frame is allocated and the memo position information 550 may be matched with each other, and through such matching, the degree of agreement or difference between the gaze position information 460 and the memo position information 550 is calculated. It can be. The computing device 100 may generate comparison information between the line of sight and the position of the memo based on the calculated coincidence. Through this correspondence, difference, or comparison information, it is analyzed whether the gaze of the user and the memo on the test text data match, and the user's learning concentration or cognitive ability can be evaluated according to this analysis.

For example, the “degree of discrepancy between the position of the gaze and the position of the memo” may be determined from the position and shape 550 of the memo. A technique for determining such discrepancies may include capturing the position of the gaze, the position of the memo, and the shape of the memo in units of time frames (eg, about 10 to 30 frames/sec) using an image recognition algorithm or the like. . In addition, the technique for judging this discrepancy estimates the difference between the position of the memo and the line of sight by Euclidean distance, and from the handwritten form of the memo, the thickness of the line, the shape of the line (straight line or wavy pattern) , how straight it is if it is a straight line, etc.) and put it into a multilayer perceptron-based regression model to find out the degree of attention.

The computing device 100 uses the second rule-based module, in a process in which the user reads the test text data from the gaze position information 460, the user's eye gaze is directed toward the rear of the test text data (eg, backward direction). Section) may be determined to generate the gaze retrograde information 630 . The computing device 100 may determine whether the gaze position is moving in the forward direction of the text or in the reverse direction of the text, based on the gaze position information 460 . Based on the number of times 630 of gaze reversal, the computing device 100 may evaluate the user's suitability or interest in the corresponding test text data. For example, if the number of backtracking 630 is relatively large, it may be determined that the corresponding user does not match the level of difficulty of the corresponding text data.

A rule-based module in the present disclosure may refer to a module that attempts to generate an output (eg, solves a problem) for input data using a predetermined rule (eg, condition).

In a further embodiment of the present disclosure, the first model 610A uses handwriting analysis information (handwriting analysis information indicating precision of handwriting from memo type information 550 in a manner of obtaining self entropy for the output of the last layer of the first multi-layer perceptron). 650) can be created. The first model 610A may generate information 640 including sharpness of handwriting and/or precision of handwriting by receiving an input regarding the shape 550 of the memo. In this case, the computing device 100 determines the relevance of the user's memo to text data, the accuracy of the user's memo, and/or the reliability of the user's memo, and evaluates the user's cognitive ability based on the determined information. can do. The first model 610A may obtain a memo type latent feature from the memo type information 550 using a convolutional neural network. In addition, the first model 610A generates a matrix including the acquired memo shape latent features and memo size information, and inputs the generated matrix to the second multi-layer perceptron to generate the memo quantitative information 650. can Quantitative information on memos may include the amount of memos generated by the user, and/or the amount of memos related to text data among the memos created by the user, and the amount of meaningful memos among the memos created by the user. Not limited. For example, as the amount of memos related to text data increases, a positive result may be derived when evaluating a user's cognitive ability. As another example, when the amount of memos exceeds a critical value, it may be determined that concentration on text data is low.

In one embodiment, by way of example and not limitation, the first model 610A includes a rule-based module for measuring the distance between the location of the gaze and the location of the memo, and a rule-based module for determining whether the direction of the gaze is moving backwards in text from the memo image. It may include an artificial intelligence-based module for estimating the sharpness of the text, and an artificial intelligence-based module for extracting latent features of the shape of the memo and generating quantitative features of the memo.

In one embodiment of the present disclosure, the computing device 100 generates the user's text suitability evaluation information 660 from the facial muscle movement information 470 and the emotional state information 480 using the second model 610B. can do. In one embodiment, by way of example and not limitation, the second model 610B may mean a module of a multilayer perceptron structure. The second model 610B utilizes motion information 470 of the facial muscles acquired based on the face image and current emotional state information 480 of the user to generate an output for determining whether the current text data is suitable for the user. can create In an additional embodiment, the second model 610B may refer to a model for classifying suitability between text data and users.

For example, a time frame may be allocated to motion information 470 of facial muscles and emotional state information 480 of a user obtained based on a face image. The computing device 100 analyzes facial muscle motion information 470 having a corresponding time frame and emotional state information 480 of the user to determine whether the user is interested in the category currently assigned to the text data (if so, how much). whether) or whether the user is suitable for the level of difficulty assigned to the current text data (if so, how suitable) information can be generated.

In one embodiment of the present disclosure, the computing device 100 may generate text suitability evaluation information 670 of the user from the vibration pattern 560 using the third model 610C. The aforementioned vibration pattern 560 may be extracted from sensor data. The computing device 100 inputs the vibration pattern information 560 into at least one structure of a transformer structure or a convolutional neural network structure to which an attention mechanism is applied, and the user's emotion in the process of reading the test text data by the user. A latent feature vector reflecting state and concentration can be created. The computing device 100 may generate evaluation information on the user's cognitive ability based on a latent feature vector that reflects the emotional state and concentration of the user during the text data reading process. The third model 610C inputs the vibration pattern information extracted by the fourth preprocessing module to a transformer structure module to which an attention mechanism is applied or a multilayer convolutional neural network structure module to generate a latent feature vector that reflects the user's emotional state and concentration. In this way, it is possible to provide basic information for reflecting the user's emotional state and concentration.

In one embodiment, the fourth model 610D compares the recorded information by time 570 and the test text data to display corresponding parts, and compares the recorded information by time 570 and the gaze position information 460 to display the above data. User concentration information may be generated. For example, the fourth model 610D may determine the first degree of similarity by comparing text information included in the time-by-time recording information 570 with test text data. Additionally, the fourth model 610D may determine the second degree of similarity by comparing the recording information 570 for each time period with the gaze position information 460 . How much the user concentrates on the text data may be calculated through the first similarity and/or the second similarity. As described above, evaluation information on the user's cognitive ability may be automatically generated based on the information on the text data being read and the information on the gaze position through the fourth model 610D. The fifth model 610E may generate evaluation information on the user's cognitive ability based on the concentration information.

In one embodiment of the present disclosure, the fifth model 610E calculates the user's cognitive ability evaluation information 690 based on the outputs 620, 630, 640, 650, 660, 670, and 680 of the aforementioned models. can create For example, the fifth model 610E may calculate the cognitive ability evaluation score 690 by inputting outputs of the first to fourth models into a transformer structure module or a multilayer perceptron module in which the attention mechanism is reflected.

In one embodiment of the present disclosure, the first model 610A to the fifth model 610E may exist integrated into one computing device 100 . As another example, at least some of the first model 610A to fifth model 610E may exist outside the computing device 100 to communicate with the computing device 100 and transmit/receive data. At least some of the first model 610A to fifth model 610E may exist in the server and the other part may exist in the user terminal.

In a further embodiment of the present disclosure, the computing device 100 may generate a classification result for the user's test text data based on the cognitive ability evaluation score 690 . For example, the classification result may include a result on a topic and/or a result on difficulty (readability). For example, if the test text data is a topic about the financial field and has a level of difficulty of an educated adult, a classification result including the user's interest and difficulty in the financial field is output based on the cognitive ability evaluation score (690). It can be.

In one embodiment of the present disclosure, a classification result or a cognitive ability evaluation score may be input to a recommendation data generation model (not shown). The recommendation data generation model may select text data having a similar subject and difficulty to those included in the classification result (or suitable for the user) among pre-stored text data as the recommendation text data.

As another example, the recommendation data generation model may be trained based on a generative adversarial network (GAN). The recommendation data generation model may be an artificial intelligence-based generation model that receives learning text data labeled with a cognitive ability evaluation score 690 or a classification result (topic and difficulty) and generates learning text data having a similar topic and difficulty. . Also, a discriminator model (not shown) may be trained to perform a task of discriminating text data generated by the recommendation data generating model. Accordingly, as the recommendation data generation model and the discriminator model learn competitively, the quality of text data generated in the recommendation data generation model may be improved.

In an embodiment, the recommended text data may mean text data for reinforcement of weak parts of the user, as a result of eye tracking or cognitive ability evaluation. For example, the fact that a specific user with a difficulty level of the second grade of elementary school has difficulties when learning scientific texts (or difficulties when learning specific text data) can be used to evaluate eye tracking or cognitive abilities. can be obtained through Therefore, when the text data that the user has difficulties in is input into the recommendation data generation model, the subject (e.g., scientific passage) and difficulty (e.g., 2nd grade level of elementary school) of the text data are obtained, and the obtained topic and difficulty Based on the recommendation data generation model, it is possible to generate scientific fingerprints at the second grade level of elementary school as recommendation text data. The generation of such recommended text data may be performed by searching a database storing a large amount of text data and selecting suitable text data.

According to an embodiment of the present disclosure, when specific text data is input, a classification result for the corresponding text data is automatically generated, and furthermore, recommended text data matching the input text data can be generated according to the classification result. do. Accordingly, text data that is a problem in the user's learning process or that can attract the user's attention is selected in combination with the cognitive ability evaluation result and / or the eye tracking function, and recommended text data similar to the selected text data is generated. Therefore, user-customized text data can be provided.

7 depicts a simplified and general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

Although the present disclosure has been described above as being generally embodied by a computing device, those skilled in the art will understand that the present disclosure may be combined with computer-executable instructions and/or other program modules that may be executed on one or more computers and/or may be implemented in hardware and software. It will be appreciated that it can be implemented as a combination.

Generally, program modules include routines, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types. In addition, those skilled in the art will understand that the methods of the present disclosure can be applied to single-processor or multiprocessor computer systems, minicomputers, mainframe computers as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, and the like ( It will be appreciated that each of these may be implemented with other computer system configurations, including those that may be operative in connection with one or more associated devices.

The described embodiments of the present disclosure may also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Computers typically include a variety of computer readable media. Computer readable media can be any medium that can be accessed by a computer, including volatile and nonvolatile media, transitory and non-transitory media, removable and non-transitory media. Includes removable media. By way of example, and not limitation, computer readable media may include computer readable storage media and computer readable transmission media. Computer readable storage media are volatile and nonvolatile media, transitory and non-transitory, 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. includes media Computer readable storage media may include RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage device, magnetic cassette, magnetic tape, magnetic disk storage device or other magnetic storage device. device, or any other medium that can be accessed by a computer and used to store desired information.

A computer readable transmission medium typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism. Including all information delivery media. The term modulated data signal means a signal that has one or more of its characteristics set or changed so as to encode information within the signal. By way of example, and not limitation, computer readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer readable transmission media.

An exemplary environment 1100 implementing various aspects of the present disclosure is shown including a computer 1102, which includes a processing unit 1104, a system memory 1106, and a system bus 1108. do. System bus 1108 couples system components, including but not limited to system memory 1106 , to processing unit 1104 . Processing unit 1104 may be any of a variety of commercially available processors. Dual processor and other multiprocessor architectures may also be used as the processing unit 1104.

System bus 1108 may be any of several types of bus structures that may additionally be interconnected to a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures. System memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) 1112 . A basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, or EEPROM, and is a basic set of information that helps transfer information between components within computer 1102, such as during startup. contains routines. RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.

The computer 1102 may also include an internal hard disk drive (HDD) 1114 (eg, EIDE, SATA) - the internal hard disk drive 1114 may also be configured for external use within a suitable chassis (not shown). Yes—a magnetic floppy disk drive (FDD) 1116 (e.g., for reading from or writing to a removable diskette 1118), and an optical disk drive 1120 (e.g., a CD-ROM) for reading disc 1122 or reading from or writing to other high capacity optical media such as DVDs). The hard disk drive 1114, magnetic disk drive 1116, and optical disk drive 1120 are connected to the system bus 1108 by a hard disk drive interface 1124, magnetic disk drive interface 1126, and optical drive interface 1128, respectively. ) can be connected to The interface 1124 for external drive implementation includes at least one or both of USB (Universal Serial Bus) and IEEE 1394 interface technologies.

These drives and their associated computer readable media provide non-volatile storage of data, data structures, computer executable instructions, and the like. In the case of computer 1102, drives and media correspond to storing any data in a suitable digital format. Although the description of computer readable media above refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art can use zip drives, magnetic cassettes, flash memory cards, cartridges, etc. It will be appreciated that other tangible media readable by the computer, such as the like, may also be used in the exemplary operating environment and any such media may include computer executable instructions for performing the methods of the present disclosure.

A number of program modules may be stored on the drive and RAM 1112, including an operating system 1130, one or more application programs 1132, other program modules 1134, and program data 1136. All or portions of the operating system, applications, modules and/or data may also be cached in RAM 1112. It will be appreciated that the present disclosure may be implemented in a variety of commercially available operating systems or combinations of operating systems.

A user may enter commands and information into the computer 1102 through one or more wired/wireless input devices, such as a keyboard 1138 and a pointing device such as a mouse 1140. Other input devices (not shown) may include a microphone, IR remote control, joystick, game pad, stylus pen, touch screen, and the like. Although these and other input devices are often connected to the processing unit 1104 through an input device interface 1142 that is connected to the system bus 1108, a parallel port, IEEE 1394 serial port, game port, USB port, IR interface, may be connected by other interfaces such as the like.

A monitor 1144 or other type of display device is also connected to the system bus 1108 through an interface such as a video adapter 1146. In addition to the monitor 1144, computers typically include other peripheral output devices (not shown) such as speakers, printers, and the like.

Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 1148 via wired and/or wireless communications. Remote computer(s) 1148 may be a workstation, computing device computer, router, personal computer, handheld computer, microprocessor-based entertainment device, peer device, or other common network node, and generally includes It includes many or all of the components described for, but for simplicity, only memory storage device 1150 is shown. The logical connections shown include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, such as a wide area network (WAN) 1154 . Such LAN and WAN networking environments are common in offices and corporations and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to worldwide computer networks, such as the Internet.

When used in a LAN networking environment, computer 1102 connects to local network 1152 through wired and/or wireless communication network interfaces or adapters 1156. Adapter 1156 may facilitate wired or wireless communications to LAN 1152, which also includes a wireless access point installed therein to communicate with wireless adapter 1156. When used in a WAN networking environment, computer 1102 may include a modem 1158, be connected to a communicating computing device on WAN 1154, or establish communications over WAN 1154, such as over the Internet. have other means. A modem 1158, which may be internal or external and a wired or wireless device, is connected to the system bus 1108 through a serial port interface 1142. In a networked environment, program modules described for computer 1102, or portions thereof, may be stored on remote memory/storage device 1150. It will be appreciated that the network connections shown are exemplary and other means of establishing a communication link between computers may be used.

Computer 1102 is any wireless device or entity that is deployed and operating in wireless communication, eg, printers, scanners, desktop and/or portable computers, portable data assistants (PDAs), communication satellites, wireless detectable tags associated with It operates to communicate with arbitrary equipment or places and telephones. This includes at least Wi-Fi and Bluetooth wireless technologies. Thus, the communication may be a predefined structure as in conventional networks or simply an ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) makes it possible to connect to the Internet without wires. Wi-Fi is a wireless technology, such as a cell phone, that allows such devices, eg, computers, to transmit and receive data both indoors and outdoors, i.e. anywhere within coverage of a base station. Wi-Fi networks use a radio technology called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks can operate in the unlicensed 2.4 and 5 GHz radio bands, for example, at 11 Mbps (802.11a) or 54 Mbps (802.11b) data rates, or in products that include both bands (dual band) .

Those skilled in the art will understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, instructions, information, signals, bits, symbols and chips that may be referenced in the above description are voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields s or particles, or any combination thereof.

Those skilled in the art will understand that the various illustrative logical blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein are electronic hardware, (for convenience) , may be implemented by various forms of program or design code (referred to herein as software) or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the particular application and the design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

Various embodiments presented herein may be implemented as a method, apparatus, or article of manufacture using standard programming and/or engineering techniques. The term article of manufacture includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, computer-readable storage media include magnetic storage devices (eg, hard disks, floppy disks, magnetic strips, etc.), optical disks (eg, CDs, DVDs, etc.), smart cards, and flash memory devices (eg, EEPROM, cards, sticks, key drives, etc.), but are not limited thereto. Additionally, various storage media presented herein include one or more devices and/or other machine-readable media for storing information.

It is to be understood that the specific order or hierarchy of steps in the processes presented is an example of exemplary approaches. Based upon design priorities, it is to be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of this disclosure. The accompanying method claims present elements of the various steps in a sample order, but are not meant to be limited to the specific order or hierarchy presented.

The description of the presented embodiments is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art of this disclosure, and the general principles defined herein may be applied to other embodiments without departing from the scope of this disclosure. Thus, the present disclosure is not to be limited to the embodiments presented herein, but is to be interpreted in the widest scope consistent with the principles and novel features presented herein.

Claims (15)

A method performed by a computing device comprising at least one processor, comprising:
acquiring an eye image using a first pre-processing module that takes a face image as an input, and extracting face information using a second pre-processing module that takes the face image as an input;
generating gaze position information of the user and gaze pattern information of the user from the eye image and the face information using a gaze information generating module;
generating facial movement information of the user from the eye image and the face information by using a facial movement information generating module;
generating emotional state information of the user from the eye image and the face information using an emotional state classification module; and
generating evaluation information for the cognitive ability of the user based at least in part on the gaze position information, the gaze pattern information, the facial motion information, and the emotional state information;
including,
method. According to claim 1,
The gaze information generation module,
Extracting latent features for the eye image and face information using a convolutional neural network structure, and generating location information of the user's gaze from the latent features using a multi-layer perceptron (MLP) structure;
Based on the user's gaze position information, calculating the movement speed and direction of the user's gaze, and
Generating gaze pattern information of the user based on the moving speed and direction of the user's gaze;
method. According to claim 1,
The first pre-processing module, the second pre-processing module, the face motion information generating module, and the emotional state information generating module,
Including a convolutional neural network structure for extracting latent features and a multi-layer perceptron (MLP) structure for generating outputs of modules from the extracted latent features,
method. According to claim 1,
The face image is obtained in a process in which the user reads test text data to which a predetermined category and a predetermined level of difficulty are assigned.
method. According to claim 4,
obtaining a memo image of the user generated in a process of the user reading the test text data; and
obtaining memo location information and memo type information from the memo image using a third pre-processing module;
Including more,
The generating of the evaluation information on the user's cognitive ability may be further based on the memo location information and the memo type information.
method. According to claim 5,
Generating evaluation information on the user's cognitive ability,
Using the first model, from the gaze position information, the gaze pattern information, the memo position information, and the memo shape information, comparison information between gaze and memo position, gaze retrogression information, handwriting analysis information, and memo quantitative information including generating first meta information; and
generating evaluation information on the user's cognitive ability based on the first meta-information;
including,
method. According to claim 6,
The first model,
Using a first rule-based module, a degree of agreement between the gaze position information and the memo position is calculated by measuring distance information between the gaze position information and the memo position information in a corresponding time frame, and Based on this, comparison information between the line of sight and the position of the memo is generated, and
Using a second rule-based module, determining a section in which the user's gaze is directed backward in the test text data in a process in which the user reads the test text data from the gaze position information to generate the gaze retrogression information ,
method. According to claim 7,
The first model,
generating the handwriting analysis information representing precision of handwriting from the memo type information in a manner of obtaining self entropy for an output of the last layer of the first multi-layer perceptron; and
Using a convolutional neural network, a memo type latent feature is obtained from the memo type information, a matrix including the obtained memo type latent feature and memo size information is generated, and the generated matrix is applied to a second multi-layer perceptron. input to generate the memo quantitative information,
method. According to claim 5,
obtaining sensor data generated in a process in which the user reads the test text data; and
generating a vibration pattern generated while the user reads the test text data from the sensor data using a fourth pre-processing module;
Including more,
Generating the evaluation information on the user's cognitive ability is further based on the vibration pattern,
method. According to claim 9,
The fourth preprocessing module,
Generating the vibration pattern by using at least one of a transformer structure, a convolutional neural network structure, or a rule-based structure including a low pass filter to which an attention mechanism is applied to the wave data extracted from the sensor data ,
method. According to claim 10,
Generating evaluation information on the user's cognitive ability,
By inputting the vibration pattern information into at least one of a transformer structure or a convolutional neural network structure to which an attention mechanism is applied, a latent feature vector reflecting the user's emotional state and concentration in the process of reading the test text data is generated. doing; and
generating evaluation information on the user's cognitive ability further based on a latent feature vector reflecting the user's emotional state and concentration in the process of reading the test text data;
including,
method. According to claim 5,
obtaining voice data uttered by the user while the user reads the test text data; and
generating time-based recording information from the voice data using a fifth pre-processing module that converts the voice data into text data by combining a convolutional neural network structure and a transformer structure;
Including more,
The generating of the evaluation information on the cognitive ability of the user is additionally based on the time-based recording information.
method. According to claim 12,
Generating evaluation information on the user's cognitive ability,
generating concentration information of the user by comparing the recorded information for each time period and the test text data to display corresponding parts and comparing the recorded information for each time period with the gaze position information using a fourth model; and
generating evaluation information on the user's cognitive ability based on the concentration information;
including,
method. A computer program stored in a computer-readable storage medium, wherein the computer program performs the following method when executed by one or more processors, the method comprising:
acquiring an eye image using a first pre-processing module that takes a face image as an input, and extracting face information using a second pre-processing module that takes the face image as an input;
generating gaze position information of the user and gaze pattern information of the user from the eye image and the face information using a gaze information generation module;
generating facial movement information of the user from the eye image and the face information by using a facial movement information generating module;
generating emotional state information of the user from the eye image and the face information using an emotional state classification module; and
generating evaluation information for the cognitive ability of the user based at least in part on the gaze position information, the gaze pattern information, the facial motion information, and the emotional state information;
including,
A computer program stored on a computer readable storage medium. A computing device including a storage unit and a processor,
The processor performs the following method, the method comprising:
acquiring an eye image using a first pre-processing module that takes a face image as an input, and extracting face information using a second pre-processing module that takes the face image as an input;
generating gaze position information of the user and gaze pattern information of the user from the eye image and the face information using a gaze information generating module;
generating facial movement information of the user from the eye image and the face information by using a facial movement information generating module;
generating emotional state information of the user from the eye image and the face information using an emotional state classification module; and
generating evaluation information for the cognitive ability of the user based at least in part on the gaze position information, the gaze pattern information, the facial motion information, and the emotional state information;
including,
computing device.

Images