KR20240034089A - Method for image feature extraction robust to environmental changes - Google Patents

Abstract

According to an embodiment of the present disclosure, a method for training a neural network model for extracting features of an image, which is performed by one or more processors of a computing device, is disclosed. The method includes extracting a plurality of first feature vectors based on original image information; Obtaining modified image information based on the original image information; extracting a plurality of second feature vectors based on the modified image information; calculating one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors; calculating a loss function based on at least one of the one or more weights, the plurality of first feature vectors, or the plurality of second feature vectors; And it may include training a neural network model based on the loss function.

Description

Image feature extraction method robust to environmental changes {METHOD FOR IMAGE FEATURE EXTRACTION ROBUST TO ENVIRONMENTAL CHANGES}

This disclosure relates to a method of using a neural network model to extract features of an image, and more specifically, to a method of extracting desired features from an image using a neural network model.

In the case of existing face identification (Face Verification, Face Recognition) models, a method was used to simply extract the face and learn other information that is not related to ID (identity) information, such as the angle of the face, background, and light. There was a problem that ID information was extracted from information that was not related to ID information (Face Identity) or was small, and that ID information appeared differently depending on the conditions. For this reason, when performing face-related tasks (e.g., FaceSwap, Style Transfer, etc.), problems occurred that degraded model performance, and there were also problems with limitations in utilizing and developing the task. .

Therefore, there is a need for a method that can reduce or remove the influence of the angle of the face or the background during the process of extracting face identity information.

Meanwhile, the present disclosure has been derived based at least on the technical background examined above, but the technical problem or purpose of the present disclosure is not limited to solving the problems or shortcomings examined above. In other words, in addition to the technical issues discussed above, the present disclosure can cover various technical issues related to the content to be described below.

The present disclosure aims to provide a method for extracting desired features from an image using a neural network model.

Meanwhile, the technical problem to be achieved by the present disclosure is not limited to the technical problems mentioned above, and may include various technical problems within the scope of what is apparent to those skilled in the art from the contents described below.

A method performed by a computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The method includes extracting a plurality of first feature vectors based on original image information; Obtaining modified image information based on the original image information; extracting a plurality of second feature vectors based on the modified image information; calculating one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors; calculating a loss function based on at least one of the one or more weights, the plurality of first feature vectors, or the plurality of second feature vectors; And it may include training a neural network model based on the loss function.

Alternatively, the step of acquiring modified image information based on the original image information may include adding at least one of environment information, pose information, or hair information to the original image information. It may include a step of applying the modified image information.

Alternatively, the step of extracting a plurality of first feature vectors based on the original image information includes extracting a plurality of first feature vectors by respectively inputting the original image information into a plurality of ID (identity) extraction models. The step of extracting a plurality of second feature vectors based on the modified image information includes inputting the modified image information into the plurality of ID extraction models to extract a plurality of second feature vectors. It can be included.

Alternatively, the plurality of ID extraction models may be different models.

Alternatively, calculating one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors corresponds to the plurality of first feature vectors and the plurality of second feature vectors, respectively. step; measuring similarity between the corresponding plurality of first feature vectors and the plurality of second feature vectors; and calculating the one or more weights based on the measured similarity, wherein the similarity may include cosine similarity.

Alternatively, calculating a loss function based on at least one of the one or more weights, the plurality of first feature vectors, or the plurality of second feature vectors may comprise: a first loss function based on the plurality of first feature vectors; extracting a synthetic feature vector; extracting a second composite feature vector based on the plurality of second feature vectors and the one or more weights; and calculating a first loss function by comparing the first synthesized feature vector and the second synthesized feature vector, wherein the step of training a neural network model based on the loss function is based on the first loss function. This may include training the first neural network model.

Alternatively, extracting a first composite feature vector based on the plurality of first feature vectors includes concatenating the plurality of first feature vectors to extract a first composite feature vector; , The step of extracting a second composite feature vector based on the plurality of second feature vectors and the one or more weights includes calculating a weighted sum for the plurality of second feature vectors based on the one or more weights. It may include extracting a second synthetic feature vector by performing .

Alternatively, calculating a loss function based on at least one of the one or more weights, the plurality of first feature vectors, or the plurality of second feature vectors may comprise: the plurality of first feature vectors and the one or more weights extracting a third synthetic feature vector based on; extracting a third feature vector by inputting the original image information into a second neural network model; and calculating a second loss function by comparing the third feature vector and the third synthesized feature vector, wherein the step of training a neural network model based on the loss function includes: It may include training the second neural network model.

Alternatively, extracting a third composite feature vector based on the plurality of first feature vectors and the one or more weights may include performing a weighted sum operation on the plurality of first feature vectors based on the one or more weights. A step of extracting a third synthesized feature vector by performing a step of extracting a third synthesized feature vector, and the step of calculating a second loss function by comparing the third feature vector and the third synthesized feature vector is set based on the third synthesized feature vector. and calculating the second loss function by measuring an error with the third feature vector.

According to an embodiment of the present disclosure for realizing the above-described object, a computer program stored in a computer-readable storage medium is disclosed. When executed on one or more processors, the computer program causes the one or more processors to perform operations for extracting features of an image, the operations comprising: extracting a plurality of first feature vectors based on original image information. movement; Obtaining modified image information based on the original image information; extracting a plurality of second feature vectors based on the modified image information; calculating one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors; calculating a loss function based on at least one of the one or more weights, the plurality of first feature vectors, or the plurality of second feature vectors; and training a neural network model based on the loss function.

A computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The device includes at least one processor; and a memory, wherein the processor extracts a plurality of first feature vectors based on original image information; Obtain modified image information based on the original image information; extract a plurality of second feature vectors based on the modified image information; calculate one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors; calculate a loss function based on at least one of the one or more weights, the plurality of first feature vectors, or the plurality of second feature vectors; And it can be configured to learn a neural network model based on the loss function.

According to an embodiment of the present disclosure for realizing the above-described problem, a method for extracting features of an image performed by a computing device is disclosed. The method includes extracting a first feature vector by inputting original image information into a pre-trained neural network model, wherein the pre-trained neural network model utilizes a plurality of ID (identity) extraction models to extract the original image It may correspond to a pre-trained model based on a plurality of feature vectors extracted based on the information.

The present disclosure can provide a method for using a neural network model to extract features of an image, and can provide a method for extracting desired features from an image using a neural network model.

1 is a block diagram of a computing device for extracting features of an image according to an embodiment of the present disclosure.
Figure 2 is a schematic diagram showing a network function according to an embodiment of the present disclosure.
Figure 3 is a flowchart showing a method of learning a neural network model for extracting features of an image according to an embodiment of the present disclosure.
Figure 4 is a schematic diagram illustrating the steps of extracting a plurality of first feature vectors and a plurality of second feature vectors according to an embodiment of the present disclosure.
Figure 5 is a schematic diagram illustrating the steps of calculating one or more weights, extracting a first composite feature vector and a second composite feature vector, and calculating a first loss function according to an embodiment of the present disclosure.
Figure 6 is a schematic diagram showing the steps of extracting a third feature vector and a third composite feature vector and calculating a second loss function according to an embodiment of the present disclosure.
Figure 7 is a flowchart showing a method of extracting a first feature vector by inputting original image information into a pre-trained neural network model according to an embodiment of the present disclosure.
Figure 8 is a brief, general schematic diagram of an example 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 disclosure. However, it is clear that these embodiments may be practiced without these specific descriptions.

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

Additionally, the term “or” is intended to mean an inclusive “or” and not an exclusive “or.” That is, unless otherwise specified or clear from context, “X utilizes A or B” is intended to mean one of the natural implicit substitutions. That is, either X uses A; X uses B; Or, if X uses both A and B, “X uses A or B” can apply to either of these cases. Additionally, the term “and/or” as used herein should be understood to refer to and include all possible combinations of one or more of the related listed items.

Additionally, the terms “comprise” and/or “comprising” should be understood to mean that the corresponding feature and/or element is present. However, the terms “comprise” and/or “comprising” should be understood as not excluding the presence or addition of one or more other features, elements and/or groups thereof. Additionally, unless otherwise specified or the context is clear to indicate a singular form, the singular terms herein and in the claims should generally be construed to mean “one or more.”

And, the term “at least one of A or B” should be interpreted to mean “a case containing only A,” “a case containing only B,” and “a case of combining A and B.”

Those skilled in the art will additionally understand that the various illustrative logical blocks, configurations, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented in electronic hardware, computer software, or both. It should be recognized that it can be implemented with combinations of. 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 software will depend on the specific application and design constraints imposed on the overall system. A skilled technician can implement the described functionality in a variety of ways for each specific application. However, such implementation decisions should not be construed as causing a departure from the scope of the present disclosure.

The description of the presented embodiments is provided to enable anyone 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. The general principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Therefore, the present invention is not limited to the embodiments presented herein. The present invention is to be interpreted in the broadest scope consistent with the principles and novel features presented herein.

In this disclosure, network function, artificial neural network, and neural network may be used interchangeably.

1 is a block diagram of a computing device for extracting features of an image 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 different components for performing the computing environment of the computing device 100, and only some of the disclosed components may configure the computing device 100.

The computing device 100 may include a processor 110, a memory 130, and a network unit 150.

The processor 110 may be composed of one or more cores, and may include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU) of a computing device. unit) may include a processor for data analysis and deep learning. The processor 110 may read a computer program stored in the memory 130 and perform data processing for machine learning 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 is used for learning neural networks, such as processing input data for learning in deep learning (DL), extracting features from input data, calculating errors, and updating the weights of the neural network using backpropagation. Calculations can be performed. At least one of the CPU, GPGPU, and TPU of the processor 110 may process learning of the network function. For example, CPU and GPGPU can work together to process learning of network functions and data classification using network functions. Additionally, in one embodiment of the present disclosure, the processors of a plurality of computing devices can be used together to process learning of network functions and data classification using network functions. Additionally, a computer program executed in a computing device according to an embodiment of the present disclosure may be a CPU, GPGPU, or TPU executable program.

The processor 110 according to an embodiment of the present disclosure may recognize original image information to extract features of the image, extract a plurality of first feature vectors based on the original image information, and extract a plurality of first feature vectors based on the original image information. Thus, operations to obtain modified image information can be performed. At this time, the original image information may include image information for extracting desired features, such as face photo information of a specific person. The original image information may include various information in addition to these examples.

According to an embodiment of the present disclosure, the processor 110 may extract a plurality of second feature vectors based on the modified image information, and the processor 110 may extract the plurality of first feature vectors and the plurality of feature vectors. One or more weights may be calculated based on the second feature vector. In addition, the processor 110 calculates a loss function based on at least one of the one or more weights, the plurality of first feature vectors, or the plurality of second feature vectors, and learns a neural network model based on the loss function. can be performed.

According to an embodiment of the present disclosure, the memory 130 may store any type of information generated or determined by the processor 110 and any type of information received by the network unit 150.

According to an embodiment of the present disclosure, the memory 130 is a flash memory type, hard disk type, multimedia card micro type, or card type memory (e.g. (e.g. SD or -Only Memory), and may include at least one type of storage medium among magnetic memory, magnetic disk, and optical disk. The computing device 100 may operate in connection with web storage that performs a storage function of the memory 130 on the Internet. The description of the memory described above is merely an example, and the present disclosure is not limited thereto.

The network unit 150 according to an embodiment of the present disclosure may use any type of known wired or wireless communication system.

For example, the network unit 150 may receive original image information from an external system. At this time, the information received from the database may be learning data or inference data for extracting features of the image using a neural network model. The original image information may include the information of the above-described examples, but is not limited to the above-described examples and may be configured in various ways within a range that a person skilled in the art can understand.

Additionally, the network unit 150 can transmit and receive information processed by the processor 110, a user interface, etc. through communication with other terminals. For example, the network unit 150 may provide a user interface generated by the processor 110 to a client (e.g. user terminal). Additionally, the network unit 150 may receive external input from a user authorized as a client and transmit it to the processor 110. At this time, the processor 110 may process operations such as output, modification, change, and addition of information provided through the user interface based on the user's external input received from the network unit 150.

Meanwhile, the computing device 100 according to an embodiment of the present disclosure is a computing system that transmits and receives information through communication with a client and may include a server. At this time, the client may be any type of terminal that can access the server. For example, the computing device 100, which is a server, may receive information for extracting image features from an external database, generate image feature extraction results, and provide a user interface regarding the image feature extraction results to the user terminal. . At this time, the user terminal outputs the user interface received from the computing device 100, which is a server, and can input or process information through interaction with the user.

In an additional embodiment, the computing device 100 may include any type of terminal that receives data resources generated by an arbitrary server and performs additional information processing.

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

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. A neural network can generally consist of a set of interconnected computational units, which can be referred to as nodes. These nodes may also be referred to as neurons. A neural network consists of at least one node. Nodes (or neurons) that make up neural networks may be interconnected by one or more links.

Within a neural network, one or more nodes connected through a link may form a relative input node and output node relationship. The concepts of input node and output node are relative, and any node in an output node relationship with one node may be in an input node relationship with another node, and vice versa. As described above, input node to output node relationships can be created around links. One or more output nodes 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 the data of the output node may be determined based on the data input to the input node. Here, the link connecting the input node and the output node may have a weight. Weights may be variable and may be varied by the user or algorithm in order for the neural network to perform the desired function. For example, when one or more input nodes are connected to one output node by respective links, the output node is set to the values input to the input nodes connected to the output node and the links corresponding to each input node. The output node value can be determined based on the weight.

As described above, in a neural network, one or more nodes are interconnected through one or more links to form an input node and output node relationship within the neural network. The characteristics of the neural network can be determined according to the number of nodes and links within the neural network, the correlation between the nodes and links, and the value of the weight assigned to each link. For example, if the same number of nodes and links exist and two neural networks with different weight values of the links exist, the two neural networks may be recognized as different from each other.

A neural network may consist of a set of one or more nodes. A subset of nodes that make up a neural network can form a layer. Some of the nodes constituting the neural network may form one layer based on the distances from the first input node. For example, a set of nodes with a distance n from the initial input node may constitute n layers. The distance from the initial input node can be defined by the minimum number of links that must be passed to reach the node from the initial input node. However, this definition of a layer is arbitrary for explanation purposes, and the order of a layer within a neural network may be defined in a different way than described above. For example, a layer of nodes may be defined by distance from the final output node.

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

The neural network according to an embodiment of the present disclosure is a neural network in which 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 it progresses from the input layer to the hidden layer. You 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 in the input layer may be less than the number of nodes in the output layer, and the number of nodes decreases as it progresses from the input layer to the hidden layer. there is. In addition, the neural network according to another embodiment of the present disclosure may be 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 it progresses from the input layer to the hidden layer. You can. A neural network according to another embodiment of the present disclosure may be a neural network that is a combination of the above-described neural networks.

A deep neural network (DNN) may refer to a neural network that includes multiple hidden layers in addition to the input layer and output layer. Deep neural networks allow you to identify latent structures in data. In other words, it is possible to identify the potential structure of a photo, text, video, voice, or music (e.g., what object is in the photo, what the content and emotion of the text are, what the content and emotion of the voice are, etc.) . Deep neural networks include convolutional neural networks (CNN), recurrent neural networks (RNN), auto encoders, generative adversarial networks (GAN), and restricted Boltzmann machines (RBM). machine), deep belief network (DBN), Q network, U network, Siamese network, Generative Adversarial Network (GAN), etc. The description of the deep neural network described above is only an example and the present disclosure is not limited thereto.

In one embodiment of the present disclosure, the network function may include an autoencoder. An autoencoder may be a type of artificial neural network to output output data similar to input data. The autoencoder may include at least one hidden layer, and an odd number of hidden layers may be placed between input and output layers. The number of nodes in each layer may be reduced from the number of nodes in the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically and reduced from the bottleneck layer to the output layer (symmetrical to the input layer). Autoencoders can perform nonlinear dimensionality reduction. The number of input layers and output layers can be corresponded to the dimension after preprocessing of the input data. In an auto-encoder structure, the number of nodes in the hidden layer included in the encoder may have a structure that decreases 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, not enough information may be conveyed, so if it is higher than a certain number (e.g., more than half of the input layers, etc.) ) may be maintained.

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

Neural networks can be trained to minimize output errors. In neural network learning, learning data is repeatedly input into the neural network, the output of the neural network and the error of the target for the learning data 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. This is the process of updating the weight of each node in the neural network through backpropagation. In the case of teacher learning, learning data in which the correct answer is labeled in each learning data is used (i.e., labeled learning data), and in the case of non-teacher learning, the correct answer may not be labeled in each learning data. That is, for example, in the case of teacher learning regarding data classification, the learning data may be data in which each learning data is labeled with a category. Labeled training data is input to the neural network, and the error can be calculated by comparing the output (category) of the neural network with the label of the training data. As another example, in the case of non-teachable learning for data classification, the error can be calculated by comparing the input training data with the neural network output. The calculated error is backpropagated in the reverse direction (i.e., from the output layer to the input layer) in the neural network, and the connection weight of each node in each layer of the neural network can be updated according to backpropagation. The amount of change in the connection weight of each updated node may be determined according to the learning rate. The neural network's calculation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently depending on the number of repetitions of the learning cycle of the neural network. For example, in the early stages of neural network training, a high learning rate can be used to increase efficiency by allowing the neural network to quickly achieve a certain level of performance, and in the later stages of training, a low learning rate can be used to increase accuracy.

In the learning of neural networks, the training data can generally be a subset of real data (i.e., the data to be processed using the learned neural network), and thus the error for the training data is reduced, but the error for the real data is reduced. There may be an incremental learning cycle. Overfitting is a phenomenon in which errors in actual data increase due to excessive learning on training data. For example, a phenomenon in which a neural network that learned a cat by showing a yellow cat fails to recognize that it is a cat when it sees a non-yellow cat may be a type of overfitting. Overfitting can cause errors in machine learning algorithms to increase. To prevent such overfitting, various optimization methods can be used. To prevent overfitting, methods such as increasing the learning data, regularization, dropout to disable some of the network nodes during the learning process, and use of a batch normalization layer can be applied. You can.

Figure 3 is a flowchart showing a method of learning a neural network model for extracting features of an image according to an embodiment of the present disclosure.

The computing device 100 according to an embodiment of the present disclosure may directly obtain “original image information for extracting image features” or receive it from an external system. Original image information is information that is used to train a neural network model for extracting image features according to an embodiment of the present disclosure. The external system may be a server, database, etc. that stores and manages original image information. The computing device 100 may use original image information directly acquired or received from an external system as “input data for training a neural network model for extracting features of the image.”

According to an embodiment of the present disclosure, the computing device 100 may extract a plurality of first feature vectors based on original image information (S110-1). For example, the computing device 100 may input original image information into a plurality of identity (ID) extraction models to extract a plurality of first feature vectors. At this time, each of the plurality of first feature vectors may include “feature information to be extracted from the original image information (eg, feature information that can distinguish the identity of the person)” and a plurality of IDs (identity information) ) Each extraction model may include a neural network model for extracting “feature information capable of distinguishing the identity of a person” from the original image, and the plurality of ID extraction models may include different ID extraction models. there is. In addition, the model capable of extracting the first feature vector is not limited to the ID (identity) extraction model, and the “feature information to be extracted from the original image information” is “features that can distinguish the identity of a person.” It is not limited to “information,” and various information other than these examples may be included. The specific process of extracting a plurality of first feature vectors based on the original image information will be described later with reference to FIG. 4.

The computing device 100 may obtain modified image information based on the original image information (S110-2). For example, the computing device 100 may obtain modified image information by applying at least one of environment information, pose information, or hair information to the original image information. . At this time, the environmental information may include information such as the background, clothes worn, lighting, and brightness, and the pose information may include information such as the angle of the face, gaze direction, and facial expression of the person, and the body hair information. may include information such as hair, beard, eyebrows, etc. Meanwhile, the environment information, the pose information, or the hair information may include various types of information in addition to these examples. Additionally, the computing device 100 may extract a plurality of second feature vectors based on the modified image information (S110-3). For example, the computing device 100 may extract a plurality of second feature vectors by respectively inputting the modified image information into the plurality of identity (ID) extraction models. At this time, each of the plurality of second feature vectors includes “feature information to be extracted from the original image information” and “information other than the feature information to be extracted from the original image information (for example, environment information, pose information, body hair information, etc. )” may be included. In addition, each of the plurality of ID (identity) extraction models may include a neural network model for extracting “feature information capable of distinguishing the identity of a person” from the transformed image information, and the plurality of first feature vectors It may include the same model as the model that extracts. Meanwhile, the model capable of extracting the second feature vector is not limited to the ID (identity) extraction model, and the “information other than the feature information to be extracted from the original image information” includes environment information and pose. ) information, or body hair information, and various information other than these examples may be included. The specific process in which modified image information is acquired based on the original image information and a plurality of second feature vectors are extracted will be described later with reference to FIG. 4.

The computing device 100 sets one or more weights based on “a plurality of first feature vectors obtained through step S110-1” and “a plurality of second feature vectors obtained through steps S110-2 to S110-3.” It can be calculated (S120). For example, the computing device 100 corresponds to each of the plurality of first feature vectors and the plurality of second feature vectors, and determines a similarity between the corresponding plurality of first feature vectors and the plurality of second feature vectors. may be measured, and the one or more weights may be calculated based on the measured similarity. At this time, the similarity may include cosine similarity, but various similarity measurement results may be included in addition to cosine similarity. A specific process of calculating one or more weights based on a plurality of first feature vectors and a plurality of second feature vectors will be described later with reference to FIG. 5.

The computing device 100 may include “one or more weights calculated through step S120,” “a plurality of first feature vectors obtained through step S110-1,” or “a plurality of first feature vectors obtained through steps S110-2 to S110-3.” A loss function may be calculated based on at least one of the “second feature vectors” (S130). For example, according to an embodiment of the present disclosure, the computing device 100 extracts a first composite feature vector based on the plurality of first feature vectors and adds the plurality of second feature vectors and the one or more weights to the plurality of second feature vectors and the one or more weights. Extract a second synthetic feature vector based on the A first loss function may be calculated by comparing the first synthesized feature vector and the second synthesized feature vector. Additionally, the computing device 100 may extract a first composite feature vector by concatenating a plurality of first feature vectors, and calculate a weighted sum (() for the plurality of second feature vectors based on the one or more weights. A weighted sum) operation can be performed to extract the second synthetic feature vector. The specific process of extracting the first synthetic feature vector and the second synthetic feature vector and calculating the first loss function will be described later with reference to FIG. 5.

Meanwhile, according to another embodiment of the present disclosure, the computing device 100 extracts a third synthesized feature vector based on the plurality of first feature vectors and the one or more weights, and applies the original image information to a second neural network model. A third feature vector can be input and extracted, and a second loss function can be calculated by comparing the third feature vector and the third synthesized feature vector. In this case, the computing device 100 may extract a third synthesized feature vector by performing a weighted sum operation on the plurality of first feature vectors based on the one or more weights. Additionally, the computing device 100 may set the third synthesized feature vector as a standard and measure the error with the third feature vector to calculate the second loss function. The specific process of extracting the third feature vector and the third synthesized feature vector and calculating the second loss function will be described later with reference to FIG. 6.

The computing device 100 may train a neural network model based on the loss function calculated in step S130 (S140). For example, the computing device 100 may train a first neural network model based on the first loss function and train the second neural network model based on the second loss function. At this time, the first neural network model and the second neural network model may be different from each other. In addition, the first neural network model or the second neural network model that is the object of learning corresponds to one model among the plurality of ID (identity) extraction models, or is a separate model different from the plurality of ID (identity) extraction models. It may be a model of Through this learning process, when the neural network model extracts “feature information to be extracted from the original image information” from the original image information, the influence of “information other than the feature information to be extracted from the original image information” is reduced. direction can be learned. For example, the “embedding result of the original image information” and the “embedding result of the image information in which information unrelated to the ID feature information is additionally transformed” are learned to be similar to each other, and as a result, the ID feature information and Neural network models can be learned in a way that reduces the influence of irrelevant information.

Meanwhile, in the present disclosure, terms such as “first,” “second,” and “third” are used to distinguish one component from another component and maintain consistency throughout the specification, and these terms The scope of rights should not be limited by those. The specific process of learning the neural network model based on the first and second loss functions calculated through step S130 will be described later with reference to FIGS. 5 and 6.

Figure 4 is a schematic diagram illustrating the steps of extracting a plurality of first feature vectors and a plurality of second feature vectors according to an embodiment of the present disclosure.

According to an embodiment of the present disclosure, the computing device 100 may input the original image information 11 into the neural network model 12 to extract a plurality of first feature vectors 13. For example, the neural network model 12 may include a plurality of ID (identity) extraction models, and the plurality of ID extraction models may be different ID extraction models, and various models other than these examples may be included. there is. Each of the plurality of first feature vectors 13 may include “feature information to be extracted from the original image information (eg, feature information capable of distinguishing the identity of a person),” and the plurality of The ID (identity) extraction model may include a neural network model for extracting “feature information capable of distinguishing a person's identity” from the original image information 11. Additionally, the plurality of first feature vectors 13 may be extracted from a plurality of different ID extraction models. The “feature information that you want to extract from the original image information” is not limited to “feature information that can distinguish the identity of a person,” and various information other than these examples may be included.

In addition, the computing device 100 adds to the original image information 11 “information other than the feature information desired to be extracted from the original image information (eg, environment information, pose information, or body hair information, etc.)” (14-1). By applying (14), modified image information (15) can be obtained. For example, after the original image information 11 is input to the encoder and encoded, the computing device 100 applies (adds) pose information and body hair information to the encoded information in a latent space. Modified encoding information can be generated, and a decoding process is performed on the modified encoding information using a decoder, so that the angle of the head is turned in the original image information (11) and a “transformed” image with beard and hair is added. “Image information (15)” can be obtained.

According to one embodiment of the present disclosure, the computing device 100 may input the modified image information 15 into the neural network model 12 to extract a plurality of second feature vectors 16. For example, the computing device 100 may input the modified image information 15 into a plurality of ID extraction models to extract a plurality of second feature vectors 16. Additionally, each of the plurality of second feature vectors 16 may include “feature information to be extracted from the original image information” and “information other than the feature information to be extracted from the original image information.” For example, “feature information you want to extract from original image information” is “human In the case of “feature information capable of distinguishing identity,” the plurality of second feature vectors 16 are “information unnecessary to distinguish the identity of a person (e.g., face angle, position of lighting, “Presence or absence of hair, etc.)” may be included. For example, the plurality of second feature vectors 16 may include feature vectors extracted by inputting image information such as face angle, lighting position, and hair deformation into a plurality of ID extraction models. Meanwhile, as an additional embodiment, the computing device 100 may generate the plurality of second feature vectors 16 by inputting different transformed images into the plurality of ID extraction models. In other words, the plurality of second feature vectors 16 include a plurality of first modified image information in which the face angle is modified, second modified image information in which the position of the lighting is modified, and third modified image information in which the hair is modified. It may also include feature vectors each input and extracted into the ID extraction model. The process of calculating one or more weights based on the plurality of first feature vectors 13 and the plurality of second feature vectors 16, calculating a loss function based on this, and learning a neural network model through the loss function includes: Hereinafter, it will be described in detail with reference to FIG. 5.

Figure 5 is a schematic diagram illustrating the steps of calculating one or more weights, extracting a first composite feature vector and a second composite feature vector, and calculating a first loss function according to an embodiment of the present disclosure.

According to an embodiment of the present disclosure, the computing device 100 may calculate one or more weights 22 based on the plurality of first feature vectors 13 and the plurality of second feature vectors 16 (21) ). In this case, the computing device 100 may correspond to the plurality of first feature vectors 13 and the plurality of second feature vectors 16, respectively. For example, among the plurality of second feature vectors 16, the number 1 is assigned to the second feature vector extracted from the first ID (identity) extraction model, and the number 1 is assigned to the number 1 extracted from the second and third ID extraction models, respectively. The second feature vectors may be assigned numbers 2 and 3, respectively. In addition, the same process may be performed for each of the plurality of first feature vectors 13, so that numbers 1, 2, and 3 may be assigned to each of the three first feature vectors, and the computing device 100 may assign numbers 1, 2, and 3. The plurality of first feature vectors 13 and the plurality of second feature vectors 16 can each correspond to feature vectors having the same number.

Meanwhile, according to another embodiment, the letter a is assigned to the second feature vector output based on “image information transformed based on environmental information,” “image information transformed based on pose information,” “ Letters b and c may be assigned to each second feature vector output based on “image information transformed based on body hair information.” In addition, letters a, b, and c may be assigned to three first feature vectors among the plurality of first feature vectors 13, and the computing device 100 determines that feature vectors with the same assigned letters correspond to each other. By doing so, the plurality of first feature vectors 13 and the plurality of second feature vectors 16 can correspond to each other.

Additionally, the computing device 100 may measure the similarity between the corresponding plurality of first feature vectors 13 and the plurality of second feature vectors 16, and based on the measured similarity, the one or more Weights 22 can be calculated. For example, the cosine similarity between the “(No. 1) second feature vector extracted from the first ID (identity) extraction model” and the corresponding “(No. 1) first feature vector” is measured, and the measured result is “Similarity 1” can be measured as (0.3), and the measured “Similarity 1 (=0.3)” can be calculated as “Weight w1”. Additionally, this process can be used to calculate the “second ID (identity) It can be performed on the “(2) second feature vector extracted from the extraction model” and the “(3) second feature vector extracted from the 3rd ID (identity) extraction model,” respectively, and as a result of the performance, “similarity 2” and its corresponding “weight w2” can be calculated as (0.5), and “similarity 3” and its corresponding “weight w3” can be calculated as (0.7).

Meanwhile, according to another embodiment, the cosine between “(a) second feature vector” output based on the “image information transformed based on environmental information” and the corresponding “(a) first feature vector” Similarity is measured, and the measured result “similarity (a)” can be measured as (0.3), and the measured “similarity (a) (=0.3)” can be calculated as “weight w(a)”. there is. In addition, this process produces “(b) second feature vector” and “(c) second feature vector” output based on “image information transformed based on pose information” and “image information transformed based on body hair information,” respectively. 2 feature vectors,” and as a result, “similarity (b)” and its corresponding “weight w(b)” are calculated as (0.5), and “similarity (c)” and its corresponding “weight w(b)” are calculated as (0.5). The resulting “weight w(c)” can be calculated as (0.7).

Meanwhile, the similarity between the corresponding plurality of first feature vectors 13 and the plurality of second feature vectors 16 is measured, and the weight 22 is calculated based on the measured similarity. The weight of the first or second feature vectors (13, 16), which contain a lot of “information other than the feature information that you want to extract from the original image information,” may be calculated low, respectively, and the “feature information that you want to extract from the original image information.” The weight of the first or second feature vectors 13 and 16 that contain a large number of can be calculated to be high, respectively. Through this, during the extraction process of the first feature vector 13 and the second feature vector 16, the concentration on “feature information to be extracted from the original image information (eg, ID feature information)” increases and the environment (facial expression) is increased. , hairstyle, face angle, etc.) is minimized/removed, so extraction of “feature information to be extracted from the original image information (e.g., ID feature information)” can be performed robustly to the environment. there is.

According to an embodiment of the present disclosure, the computing device 100 may extract the first composite feature vector 24 by concatenating the plurality of first feature vectors 13 (23). Through the process of step 23, the plurality of first feature vectors 13 can be extracted into one vector. Additionally, the computing device 100 extracts a second composite feature vector 26 by performing a weighted sum operation on the plurality of second feature vectors 16 based on the one or more weights 22. It can be done (25). For example, when the one or more weights 22 are calculated as “weight w1 (=0.3)”, “weight w2 (=0.5)”, and “weight w3 (=0.7)”, the second composite feature vector (26) ) is the “(1) second feature vector” * “weight w1 (=0.3)”, the “(2) second feature vector” * “weight w2 (=0.5)”, and the “(3) ) It can be extracted as the result of a sum operation between “second feature vector” * “weight w3 (=0.7)”. In another embodiment, the one or more weights 22 are “weight w(a)(=0.3)”, “weight w(b)(=0.5)”, “weight w(c)(=0.7)” When calculated, the second composite feature vector 26 is the “(a) second feature vector”*“weight w(a)(=0.3)” and the “(b) second feature vector”*“weight w(b)(=0.5)”, and the “(c) second feature vector”*“weight w(c)(=0.7)” can be extracted as a result of performing a sum operation.

According to an embodiment of the present disclosure, the computing device 100 calculates a first loss function 27 by comparing the first synthesized feature vector 24 and the second synthesized feature vector 26, and calculates the first loss function 27. The first neural network model 12 can be trained based on the loss function 27. For example, the error of the first synthesized feature vector 24 and the second synthesized feature vector 26 is calculated as a first loss function 27, and based on the first loss function 27, the 1 A neural network model 12 can be learned. Through this, when the first neural network model 12 extracts “feature information desired to be extracted from the original image information” from the original image information 11, “information other than the feature information desired to be extracted from the original image information” It can be learned in a way that reduces its influence. For example, “embedding results of inputting the original image information 11 into the first neural network model 12” and “image information 15 in which information unrelated to the ID feature information is additionally transformed. Since the “embedding results” input to the first neural network model 12 are learned to be similar to each other, the neural network model 12 can be learned in such a way that the influence of information unrelated to the ID feature information is consequently reduced. It works. The specific process of extracting a third feature vector and a third composite feature vector based on the one or more weights 22 and a plurality of first or second feature vectors 13 and 16, respectively, and calculating the second loss function is This is explained below with reference to FIG. 6.

Figure 6 is a schematic diagram showing the steps of extracting a third feature vector and a third composite feature vector and calculating a second loss function according to an embodiment of the present disclosure.

According to an embodiment of the present disclosure, the computing device 100 may extract a third synthesized feature vector 32 based on the plurality of first feature vectors 13 and the one or more weights 22 (31) ). The process of calculating the one or more weights 22 may refer to the description of FIG. 5 as previously described in detail. Specifically, the computing device 100 performs a weighted sum operation on the plurality of first feature vectors 13 based on the one or more weights 22 to generate a third synthesized feature vector 32. It can be extracted (31). For example, if the one or more weights 22 are calculated as “weight w1 (=0.3)”, “weight w2 (=0.5)”, and “weight w3 (=0.7)”, the third composite feature vector (26) ) is the “(No. 1) first feature vector” * “weight w1 (=0.3)”, the “(No. 2) first feature vector” * “weight w2 (=0.5)”, and the “(No. 3) ) It can be extracted as the result of a sum operation between “first feature vector” * “weight w3 (=0.7)”. In another embodiment, the one or more weights 22 are “weight w(a)(=0.3)”, “weight w(b)(=0.5)”, “weight w(c)(=0.7)” When calculated, the third composite feature vector 26 is the “(a) first feature vector”*“weight w(a)(=0.3)” and the “(b) first feature vector”*“weight w(b)(=0.5)”, and the “(c) first feature vector”*“weight w(c)(=0.7)” can be extracted as a result of performing a sum operation.

However, the numbers representing the one or more weights 22 are only examples, and the one or more weights 22 are not limited to this and may be calculated using various numerical values.

According to an embodiment of the present disclosure, the computing device 100 may extract the third feature vector 34 by inputting the original image information 11 into the second neural network model 33. For example, the second neural network model 33 may include an identity (ID) extraction model, but various other models may also be included. The third feature vector 34 may include “feature information to be extracted from the original image information.”

According to an embodiment of the present disclosure, the computing device 100 may calculate a second loss function 35 by comparing the third feature vector 34 and the third composite feature vector 32. For example, the computing device 100 sets the third synthetic feature vector 32 as a standard, measures the error between the set standard and the third feature vector 34, and calculates the second loss function 35. It can be calculated. Additionally, the computing device 100 may train the second neural network model 33 based on the second loss function 35. Through this, the second neural network model 33 can be learned in a way that the error between the set standard and the third feature vector 34 is reduced, and the learned second neural network model 33 is sensitive to external factors. By learning to minimize/remove the influence on the original image information (11), it is possible to better extract “feature information to be extracted from the original image information.” Below, in Figure 7, the process of extracting the first feature vector using a pre-trained neural network model is explained in detail.

Figure 7 is a flowchart showing a method of extracting a first feature vector by inputting original image information into a pre-trained neural network model according to an embodiment of the present disclosure.

Referring to FIG. 7, according to an embodiment of the present disclosure, a neural network model is pre-trained based on a plurality of feature vectors extracted based on the original image information 11 by utilizing a plurality of ID (identity) extraction models. (S210). Specifically, the original image information 11 or the modified image information 15 are respectively input to a plurality of ID (identity) extraction models to generate the plurality of first feature vectors 13 and the plurality of second feature vectors. (16), or the third feature vector 34 may be extracted, respectively, and the extracted plurality of first feature vectors 13, the plurality of second feature vectors 16, or the third feature vector A neural network model (12,33) may be pre-trained based on at least one of (34). For the specific learning process, the description of FIGS. 3 to 6 may be referred to.

According to an embodiment of the present disclosure, the original image information 11 may be input to the pre-trained neural network models 12 and 33 to extract a first feature vector (S220). For example, a first feature vector containing “feature information capable of distinguishing the identity of a person” may be extracted from the original image information 11. At this time, by using the pre-trained neural network model (12,33), the influence of “information other than characteristic information that can distinguish a person’s identity (e.g., environment, pose, body hair information, etc.)” is reduced. Feature vectors may be extracted.

According to an embodiment of the present disclosure, a computer-readable medium storing a data structure is disclosed. Data structure can refer to the organization, management, and storage of data to enable efficient access and modification of data. Data structure can refer to the organization of data to solve a specific problem (e.g., retrieving data, storing data, or modifying data in the shortest possible time). A data structure may be defined as a physical or logical relationship between data elements designed to support a specific data processing function. Logical relationships between data elements may include connection relationships between user-defined data elements. Physical relationships between data elements may include actual relationships between data elements that are physically stored in a computer-readable storage medium (e.g., a persistent storage device). A data structure may specifically include a set of data, relationships between data, and functions or instructions applicable to the data. Effectively designed data structures allow computing devices to perform computations while minimizing the use of the computing device's resources. Specifically, computing devices can increase the efficiency of operations, reading, insertion, deletion, comparison, exchange, and search through effectively designed data structures.

Data structures can be divided into linear data structures and non-linear data structures depending on the type of data structure. A linear data structure may be a structure in which only one piece of data is connected to another piece of data. Linear data structures may include List, Stack, Queue, and Deque. A list can refer to a set of data that has an internal order. The list may include a linked list. A linked list may be a data structure in which data is connected in such a way that each data is connected in a single line with a pointer. In a linked list, a pointer may contain connection information to the next or previous data. Depending on its form, a linked list can be expressed as a singly linked list, a doubly linked list, or a circularly linked list. A stack may be a data listing structure that allows limited access to data. A stack can be a linear data structure in which data can be processed (for example, inserted or deleted) at only one end of the data structure. Data stored in the stack may have a data structure (LIFO-Last in First Out) where the later it enters, the sooner it comes out. A queue is a data listing structure that allows limited access to data. Unlike the stack, it can be a data structure (FIFO-First in First Out) where data stored later is released later. A deck can be a data structure that can process data at both ends of the data structure.

A non-linear data structure may be a structure in which multiple pieces of data are connected behind one piece of data. Nonlinear data structures may include graph data structures. A graph data structure can be defined by vertices and edges, and an edge can include a line connecting two different vertices. Graph data structure may include a tree data structure. A tree data structure may be a data structure in which there is only one path connecting two different vertices among a plurality of vertices included in the tree. In other words, it may be a data structure that does not form a loop in the graph data structure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. Below, it is described in a unified manner as a neural network. Data structures may include neural networks. And the data structure including the neural network may be stored in a computer-readable medium. Data structures including neural networks also include data preprocessed for processing by a neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data acquired from the neural network, activation functions associated with each node or layer of the neural network, neural network It may include a loss function for learning. A data structure containing a neural network may include any of the components disclosed above. In other words, 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 acquired from the neural network, activation functions associated with each node or layer of the neural network, neural network It may be configured to include all or any combination of the loss function for learning. In addition to the configurations described above, a data structure containing a neural network may include any other information that determines the characteristics of the neural network. Additionally, the data structure may include all types of data used or generated in the computational process of a neural network and is not limited to the above. Computer-readable media may include computer-readable recording media and/or computer-readable transmission media. A neural network can generally consist of a set of interconnected computational units, which can be referred to as nodes. These nodes may also be referred to as neurons. A neural network consists of at least one node.

The data structure may include data input to the neural network. A data structure containing data input to a neural network may be stored in a computer-readable medium. Data input to the neural network may include learning data input during the neural network learning process and/or input data input to the neural network on which training has been completed. Data input to the neural network may include data that has undergone pre-processing and/or data subject to pre-processing. Preprocessing may include a data processing process to input data into a neural network. Therefore, the data structure may include data subject to preprocessing and data generated by preprocessing. The above-described 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 with the same meaning.) And the data structure including the weights of the neural network may be stored in a computer-readable medium. A neural network may include multiple weights. Weights may be variable and may be varied by the user or algorithm in order for the neural network to perform the desired function. For example, when one or more input nodes are connected to one output node by respective links, the output node is set to the values input to the input nodes connected to the output node and the links corresponding to each input node. Based on the weight, the data value output from the output node can be determined. The above-described data structure is only an example and the present disclosure is not limited thereto.

As an example and not a limitation, the weights may include weights that are changed during the neural network learning process and/or weights for which neural network learning has been completed. Weights that change during the neural network learning process may include weights that change at the start of the learning cycle and/or weights that change during the learning cycle. Weights for which neural network training has been completed may include weights for which a learning cycle has been completed. Therefore, the data structure including the weights of the neural network may include weights that are changed during the neural network learning process and/or the data structure including the weights for which neural network learning has been completed. Therefore, the above-mentioned weights and/or combinations of each weight are included in the data structure including the weights of the neural network. The above-described 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 (e.g., memory, 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 a different computing device and later reorganized and used. Computing devices can transmit and receive data over a network by serializing data structures. Data structures containing the weights of a serialized neural network can be reconstructed on the same computing device or on a different 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 to increase computational efficiency while minimizing the use of computing device resources (e.g., in non-linear data structures, B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree) may be included. The foregoing is merely an example and the present disclosure is not limited thereto.

The data structure may include hyper-parameters of a neural network. And the data structure including the hyperparameters of the neural network can be stored in a computer-readable medium. A hyperparameter may be a variable that can be changed by the user. Hyperparameters include, for example, learning rate, cost function, number of learning cycle repetitions, weight initialization (e.g., setting the range of weight values subject to weight initialization), Hidden Unit. It may include a number (e.g., number of hidden layers, number of nodes in hidden layers). The above-described data structure is only an example and the present disclosure is not limited thereto.

Figure 8 is a brief, general schematic diagram of an example computing environment in which embodiments of the present disclosure may be implemented.

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

Typically, program modules include routines, programs, components, data structures, etc. that perform specific tasks or implement specific abstract data types. Additionally, those skilled in the art will understand that the methods of the present disclosure can be used on single-processor or multiprocessor computer systems, minicomputers, mainframe computers, as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, etc. It will be appreciated that other computer system configurations may be implemented, including, but not limited to, each of which may operate in connection with one or more associated devices.

The described embodiments of the present disclosure can 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, and such computer-readable media includes volatile and non-volatile 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 refers to volatile and non-volatile media, transient and non-transitory media, removable and non-removable, 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, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage. This includes, but is not limited to, a device, or any other medium that can be accessed by a computer and used to store desired information.

A computer-readable transmission medium typically implements computer-readable instructions, data structures, program modules, or other data on a modulated data signal, such as a carrier wave or other transport mechanism. Includes all information delivery media. The term modulated data signal refers to a signal in which one or more of the characteristics of the signal have been set or changed 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 example environment 1100 is shown that implements various aspects of the present disclosure, 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 processors and other multiprocessor architectures may also be used as processing unit 1104.

System bus 1108 may be any of several types of bus structures that may further be interconnected to a memory bus, peripheral bus, and 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. The basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, and EEPROM, and is a basic input/output system that helps transfer information between components within the computer 1102, such as during startup. Contains routines. RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.

Computer 1102 may also include an internal hard disk drive (HDD) 1114 (e.g., 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 the disk 1122 or reading from or writing to other high-capacity optical media such as DVDs). Hard disk drive 1114, magnetic disk drive 1116, and optical disk drive 1120 are connected to system bus 1108 by hard disk drive interface 1124, magnetic disk drive interface 1126, and optical drive interface 1128, respectively. ) can be connected to. The interface 1124 for implementing an external drive includes at least one or both of Universal Serial Bus (USB) 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. For computer 1102, drive and media correspond to storing any data in a suitable digital format. Although the description of computer-readable media above refers to removable optical media such as HDDs, removable magnetic disks, and CDs or DVDs, those of ordinary skill in the art will also recognize zip drives, magnetic cassettes, flash memory cards, and cartridges. It will be appreciated that other types of computer-readable media may also be used in the example operating environment, and that any such media may contain computer-executable instructions for performing the methods of the present disclosure. .

A number of program modules may be stored in drives 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 on various commercially available operating systems or combinations of operating systems.

A user may enter commands and information into computer 1102 through one or more wired/wireless input devices, such as a keyboard 1138 and a pointing device such as mouse 1140. Other input devices (not shown) may include microphones, IR remote controls, joysticks, game pads, stylus pens, touch screens, etc. These and other input devices are connected to the processing unit 1104 through an input device interface 1142, which is often connected to the system bus 1108, but may also include a parallel port, an IEEE 1394 serial port, a game port, a USB port, an IR interface, It can be connected by other interfaces, etc.

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

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, portable computer, microprocessor-based entertainment device, peer device, or other conventional network node, and is generally connected to computer 1102. For simplicity, only memory storage device 1150 is shown, although it includes many or all of the components described. The logical connections depicted include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, such as a wide area network (WAN) 1154. These LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, such as the Internet.

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

Computer 1102 may be associated with any wireless device or object deployed and operating in wireless communications, such as a printer, scanner, desktop and/or portable computer, portable data assistant (PDA), communications satellite, wirelessly detectable tag. Performs actions to communicate with any device or location and telephone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, communication may be a predefined structure as in a conventional network or may simply be ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) allows connection to the Internet, etc. without wires. Wi-Fi is a wireless technology, like cell phones, that allows these devices, such as computers, to send and receive data indoors and outdoors, anywhere within the coverage area of a base station. Wi-Fi networks use wireless 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, the Internet, and wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks can operate in the unlicensed 2.4 and 5 GHz wireless bands, for example, at data rates of 11 Mbps (802.11a) or 54 Mbps (802.11b), 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, commands, information, signals, bits, symbols and chips that may be referenced in the above description include voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields. It can be expressed by particles 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 may be used in electronic hardware, (for convenience) It will be understood that it 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 interoperability of hardware and software, various illustrative components, blocks, modules, circuits and steps have been described above generally with respect to their functionality. Whether this functionality is implemented as hardware or software depends on the specific application and design constraints imposed on the overall system. A person skilled in the art of this disclosure may implement the described functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of this disclosure.

The 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 (e.g., hard disks, floppy disks, magnetic strips, etc.), optical disks (e.g., CDs, DVDs, etc.), smart cards, and flash. Includes, but is not limited to, memory devices (e.g., EEPROM, cards, sticks, key drives, etc.). 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 illustrative approaches. It is to be understood that the specific order or hierarchy of steps in processes may be rearranged within the scope of the present disclosure, based on design priorities. The appended method claims present elements of the various steps in a sample order but are not meant to be limited to the particular 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, and the general principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Thus, the present disclosure is not limited to the embodiments presented herein but is to be interpreted in the broadest scope consistent with the principles and novel features presented herein.

Claims (9)

A method of learning a neural network model for extracting features of an image, performed by a computing device,
extracting a plurality of first feature vectors based on original image information;
Obtaining modified image information based on the original image information;
extracting a plurality of second feature vectors based on the modified image information;
calculating one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors;
extracting a third composite feature vector based on the plurality of first feature vectors and the one or more weights;
extracting a third feature vector by inputting the original image information into a neural network model;
Comparing the third feature vector and the third synthesized feature vector to calculate a loss function; and
training the neural network model based on the loss function;
Including,
method.
According to claim 1,
The step of acquiring modified image information based on the original image information includes:
Obtaining modified image information by applying at least one of environment information, pose information, and hair information to the original image information;
Including,
method.
According to claim 1,
The step of extracting a plurality of first feature vectors based on the original image information includes:
Including the step of extracting a plurality of first feature vectors by inputting the original image information into a plurality of ID (identity) extraction models, respectively,
The step of extracting a plurality of second feature vectors based on the modified image information includes:
Comprising the step of extracting a plurality of second feature vectors by inputting the modified image information into the plurality of ID extraction models, respectively.
method.
According to claim 3,
The plurality of ID extraction models are different models,
method.
According to claim 1,
Calculating one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors includes:
Corresponding each of the plurality of first feature vectors and the plurality of second feature vectors;
measuring similarity between the corresponding plurality of first feature vectors and the plurality of second feature vectors; and
Comprising the step of calculating the one or more weights based on the measured similarity,
The similarity includes cosine similarity,
method.
According to claim 1,
Extracting a third synthesized feature vector based on the plurality of first feature vectors and the one or more weights includes:
Extracting a third synthesized feature vector by performing a weighted sum operation on the plurality of first feature vectors based on the one or more weights,
Comparing the third feature vector and the third synthesized feature vector to calculate a loss function includes:
Comprising the step of setting the third synthesized feature vector as a standard and measuring the error with the third feature vector to calculate the loss function,
method.
A computer program stored in a computer-readable storage medium, wherein the computer program, when executed by one or more processors, causes the one or more processors to perform operations for extracting features of an image, the operations comprising:
Extracting a plurality of first feature vectors based on original image information;
Obtaining modified image information based on the original image information;
extracting a plurality of second feature vectors based on the modified image information;
calculating one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors;
extracting a third composite feature vector based on the plurality of first feature vectors and the one or more weights;
extracting a third feature vector by inputting the original image information into a neural network model;
calculating a loss function by comparing the third feature vector and the third synthesized feature vector; and
An operation of training the neural network model based on the loss function;
Including,
A computer program stored on a computer-readable storage medium.
As a computing device,
at least one processor; and
Memory
Including,
The at least one processor,
extract a plurality of first feature vectors based on the original image information;
Obtain modified image information based on the original image information;
extract a plurality of second feature vectors based on the transformed image information;
calculate one or more weights based on the plurality of first feature vectors and the plurality of second feature vectors;
extract a third composite feature vector based on the plurality of first feature vectors and the one or more weights;
inputting the original image information into a neural network model to extract a third feature vector;
Compare the third feature vector and the third composite feature vector to calculate a loss function; and
configured to train the neural network model based on the loss function,
Computing device.
A method for extracting features of an image, performed by a computing device, comprising:
Including the step of extracting a first feature vector by inputting original image information into a pre-trained neural network model,
The pre-trained neural network model is,
A plurality of first feature vectors are extracted based on original image information using a plurality of ID (identity) extraction models, and a first composite feature vector extracted based on the plurality of first feature vectors;
a plurality of second feature vectors extracted based on transformed image information;
a third feature vector extracted based on the original image information;
one or more weights calculated based on the plurality of first feature vectors and the plurality of second feature vectors;
a third synthesized feature vector extracted based on the plurality of first feature vectors and the one or more weights; and
Corresponding to a pre-trained model based on a loss function calculated by comparing the third feature vector and the third synthesized feature vector,
method.