KR102627764B1 - Method for training the neural network model transforming images by using parial images - Google Patents

Abstract

According to an embodiment of the present disclosure, a method of converting an image using a neural network model, performed by one or more processors of a computing device, is disclosed. The method includes obtaining target image information and converted image information; extracting a plurality of target partial image information based on the target image information; extracting a plurality of converted partial image information based on the converted image information; calculating a plurality of loss functions based on the plurality of target partial image information and the plurality of converted partial image information; and training the neural network model based on the calculated plurality of loss functions.

Description

A method of learning a neural network model that transforms images using partial images {METHOD FOR TRAINING THE NEURAL NETWORK MODEL TRANSFORMING IMAGES BY USING PARIAL IMAGES}

This disclosure relates to a technology for learning a neural network model that converts an image using a partial image. More specifically, the image conversion quality can be improved by learning a neural network model by comparing a part of the target image with a part of the converted image. It's about technology.

In the case of existing face swap models, most are implemented using generative adversarial networks called GANs. At this time, the discriminator is generally designed to discriminate the entire face area, but there was a problem in that the entire face area was calculated as a single loss value and the neural network model was learned without considering the differences in each facial component. Because of this, there were many cases where partial areas such as the eyes or mouth were not converted robustly in various environments and quality was poor. Accordingly, there is an increasing demand for technology that can learn neural network models to extract more realistic results by allowing partial comparison of transformed areas on images.

Therefore, new technologies that can solve these problems are required.

The present disclosure aims to provide a technology that can improve image conversion quality by learning a neural network model by comparing a part of the image with a part of the converted image using a neural network model.

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 obtaining target image information and converted image information; extracting a plurality of target partial image information based on the target image information; extracting a plurality of converted partial image information based on the converted image information; calculating a plurality of loss functions based on the plurality of target partial image information and the plurality of converted partial image information; and training the neural network model based on the calculated plurality of loss functions.

Alternatively, the converted image information may include image information obtained through a generator included in the neural network model.

Alternatively, extracting a plurality of target partial image information based on the target image information may include extracting first segmentation information for the target image information; Receiving first condition information for the extracted first division information; and extracting first target partial image information based on the first condition information and the extracted first division information, and extracting a plurality of pieces of converted partial image information based on the converted image information. Extracting second segmentation information for the converted image information; receiving the first condition information for the extracted second division information; and extracting first converted partial image information based on the first condition information and the extracted second division information.

Alternatively, the step of extracting first target portion image information based on the first condition information and the extracted first division information may include first image information based on the first condition information and the extracted first division information. creating a target masking area; And extracting the first target partial image information using the generated first target masking area for the target image information, based on the first condition information and the extracted second division information. 1 The step of extracting converted partial image information includes generating a first converted masking area based on the first condition information and the extracted second division information; and extracting the first converted partial image information using the first converted masking area generated for the converted image information.

Alternatively, calculating a plurality of loss functions based on the plurality of target partial image information and the plurality of converted partial image information may include first condition information and a first extracted based on the target image information. 1 calculating a first loss function based on target partial image information and first converted partial image information extracted based on the first condition information and the converted image information; and a second target partial image information extracted based on second condition information and the target image information and second converted partial image information extracted based on the second condition information and the converted image information. It may include calculating a loss function.

Alternatively, based on first target partial image information extracted based on first condition information and the target image information and first converted partial image information extracted based on the first condition information and the converted image information. The step of calculating the first loss function includes inputting the first target portion image information into a first discriminator network to calculate a first true loss function; And inputting the first converted partial image information into a first discriminator network to calculate a first fake loss function, wherein a second target extracted based on second condition information and the target image information The step of calculating a second loss function based on partial image information and second converted partial image information extracted based on the second condition information and the converted image information includes converting the second target partial image information into a second loss function. Calculating a second truth loss function by inputting it into a discriminant network; And it may include calculating a second fake loss function by inputting the second converted partial image information into a second discriminator network.

Alternatively, the neural network model includes one or more of the first discriminant network or the second discriminant network, and training the neural network model based on the calculated plurality of loss functions comprises: training the first discriminant network based on one or more of a real loss function or the first fake loss function; Alternatively, it may include training the second discriminant network based on one or more of the second real loss function or the second fake loss function.

Alternatively, training the neural network model based on the calculated plurality of loss functions may include calculating an average of the calculated plurality of loss functions; calculating an average of the calculated plurality of loss functions as a final loss function; And it may include training the neural network model based on the calculated final loss function.

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 the computer program is executed on one or more processors, it causes the one or more processors to perform operations for converting an image using a neural network model, and the operations include: obtaining target image information and converted image information. ; extracting a plurality of target partial image information based on the target image information; extracting a plurality of converted partial image information based on the converted image information; calculating a plurality of loss functions based on the plurality of target partial image information and the plurality of converted partial image information; and training the neural network model based on the calculated plurality of loss functions.

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 acquires target image information and converted image information; extracting a plurality of target partial image information based on the target image information; extracting a plurality of converted partial image information based on the converted image information; calculate a plurality of loss functions based on the plurality of target partial image information and the plurality of converted partial image information; And it can be configured to learn a neural network model based on the calculated plurality of loss functions.

The present disclosure can provide a technology that can improve image conversion quality by utilizing a neural network model to learn a neural network model by comparing a part of the image with a part of the converted image.

1 is a block diagram of a computing device for converting an image using a neural network model 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.
FIG. 3 is a flowchart illustrating a method of training a neural network model for converting an image using a partial image according to an embodiment of the present disclosure.
Figure 4 is a schematic diagram illustrating a process for obtaining a converted image according to an embodiment of the present disclosure.
FIG. 5A is a schematic diagram illustrating a process for extracting a plurality of target partial image information based on target image information according to an embodiment of the present disclosure.
FIG. 5B is a schematic diagram illustrating a process of extracting a plurality of converted partial image information based on converted image information according to an embodiment of the present disclosure.
FIG. 6 is a schematic diagram illustrating a process of calculating a plurality of loss functions based on a plurality of target partial image information and a plurality of converted partial image information according to an embodiment of the present disclosure.
7 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 recognize that the various illustrative logical blocks, components, modules, circuits, means, logic, and algorithm steps described in connection with the embodiments disclosed herein may be implemented using electronic hardware, computer software, or a combination of both. It must be recognized that it can be implemented with 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 converting an image using a neural network model 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 configurations for performing the computing environment of the computing device 100, and only some of the disclosed configurations 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 model. The processor 110 processes input data for learning in deep learning (DL), extracts features from input data, calculates errors, and learns neural network models such as updating the weights of the neural network model using backpropagation. Calculations can be performed for . At least one of the CPU, GPGPU, and TPU of the processor 110 may process learning of the neural network model. For example, the CPU and GPGPU can work together to process neural network model learning and data classification using the neural network model. Additionally, in one embodiment of the present disclosure, the processors of a plurality of computing devices can be used together to process learning of a neural network model and data classification using the neural network model. 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.

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 includes Public Switched Telephone Network (PSTN), x Digital Subscriber Line (xDSL), Rate Adaptive DSL (RADSL), Multi Rate DSL (MDSL), and VDSL ( A variety of wired communication systems can be used, such as Very High Speed DSL), Universal Asymmetric DSL (UADSL), High Bit Rate DSL (HDSL), and Local Area Network (LAN).

In addition, the network unit 150 presented in the present disclosure includes Code Division Multi Access (CDMA), Time Division Multi Access (TDMA), Frequency Division Multi Access (FDMA), Orthogonal Frequency Division Multi Access (OFDMA), and SC-FDMA ( A variety of wireless communication systems can be used, such as Single Carrier-FDMA) and other systems.

In the present disclosure, the network unit 150 may be configured regardless of the communication mode, such as wired or wireless, and may be composed of various communication networks such as a personal area network (PAN) and a wide area network (WAN). You can. Additionally, the network may be the well-known World Wide Web (WWW), or may use wireless transmission technology used for short-distance communication, such as Infrared Data Association (IrDA) or Bluetooth. The techniques described in this disclosure can also be used in other networks mentioned above.

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.

Reinforcement learning is a learning method that trains a neural network model based on the reward calculated for the action selected by the neural network model so that the neural network model can determine a better action based on the state. A state is a set of values that indicate what the situation is at the current point in time, and can be understood as an input to a neural network model. Action refers to a decision based on the options that a neural network model can take, and can be understood as the output of a neural network model. Reward refers to the benefit that follows when a neural network model performs an action, and represents the value evaluated for the current state and action. Reinforcement learning can be understood as learning through trial and error in that rewards are given for actions. The reward given to the neural network model during the reinforcement learning process may be the cumulative reward of the results of multiple actions. Through reinforcement learning, it is possible to create a neural network model that maximizes the return, such as the reward itself or the total sum of the rewards, by considering rewards according to various states and actions.

FIG. 3 is a flowchart illustrating a method of training a neural network model for converting an image using a partial image according to an embodiment of the present disclosure.

Referring to FIG. 3, the computing device 100 according to an embodiment of the present disclosure may directly obtain information for converting an image using a neural network model or receive it from an external system. The external system may be a server, database, etc. that stores and manages information for converting images using a neural network model. The computing device 100 may use target image information, converted image information, condition information, source image information, etc. directly acquired or received from an external system as input data for converting an image using a neural network model.

The computing device 100 may acquire target image information and converted image information (S110). At this time, the target image information may include actual data used in the process of learning a generative adversarial network that generates the image. Additionally, target image information may include image information used to obtain a converted image. For example, the computing device 100 may obtain converted image information through a generator included in a neural network model. In the process of acquiring the converted image, technologies such as Conditional Generative Adversarial Networks (cGAN), Unpaired Image-to-Image Translation, and Conditional Variational Autoencoder (cVAE) can be used. Additionally, in the process of acquiring the converted image, it can be used to convert specific attributes in latent space using GAN Inversion technology. The technology for obtaining a converted image using a neural network model 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 the range understandable by those skilled in the art. The specific process for this is explained in detail in FIG. 4 below.

Additionally, the computing device 100 may extract a plurality of target partial image information based on the target image information obtained through step S110 (S120-1). For example, the computing device 100 extracts first segmentation information for the target image information, receives first condition information for the extracted first segmentation information, and inputs the first condition information. First target portion image information may be extracted based on condition information and the extracted first segmentation information. Specifically, the computing device 100 generates a first target masking area based on the first condition information and the extracted first division information, and uses the generated first target masking area for the target image information. Thus, the first target portion image information can be extracted. Expressions such as first and second disclosed throughout this specification are only meant to distinguish between components and are not limited thereto. The specific process of extracting a plurality of target portion image information is described in detail in FIG. 5A below.

Additionally, the computing device 100 may extract a plurality of pieces of converted partial image information based on the converted image information obtained through step S110 (S120-2).

For example, the computing device 100 extracts second segmentation information for the converted image information, receives the first condition information for the extracted second segmentation information, and receives the first condition information for the extracted second segmentation information. First converted partial image information may be extracted based on the first condition information and the extracted second division information. Specifically, the computing device 100 generates a first converted masking area based on the first condition information and the extracted second division information, and creates the first converted masking area for the converted image information. The first converted partial image information can be extracted using the area. Expressions such as first and second disclosed throughout this specification are only meant to distinguish between components and are not limited thereto. The specific process of extracting a plurality of converted partial image information is described in detail in FIG. 5B below.

Additionally, the computing device 100 may calculate a plurality of loss functions based on a plurality of target partial image information extracted through step S120-1 and a plurality of converted partial image information extracted through step S120-2. There is (S130). For example, the computing device 100 may include first target portion image information extracted based on first condition information and the target image information, and first transformation extracted based on the first condition information and the converted image information. Calculating a first loss function based on the partial image information, and calculating a first loss function based on second condition information and second target partial image information extracted based on the target image information, the second condition information, and the converted image information. A second loss function may be calculated based on the extracted second converted partial image information. Specifically, the computing device 100 calculates a first truth loss function by inputting the first target partial image information to a first discriminator network, and inputs the first converted partial image information to the first discriminant network. A first false loss function can be calculated by inputting the second target partial image information into a second discriminator network to calculate a second true loss function, and the second converted partial image information The second Fake loss function can be calculated by inputting into the second discriminant network. The specific process for this is explained in Figure 6 below.

Additionally, the computing device 100 may train the neural network model based on a plurality of loss functions calculated in step S130 (S140).

For example, the computing device 100 calculates the average of the plurality of loss functions calculated through step S130, calculates the average of the calculated plurality of loss functions as a final loss function, and adds the calculated final loss function to the final loss function. Based on this, the neural network model can be trained.

According to another embodiment of the present disclosure, the computing device 100 determines the first decision included in the neural network model based on one or more of the first real loss function or the first fake loss function. A network can be trained, and the second discriminant network included in the neural network model can be trained based on one or more of the second real loss function or the second fake loss function.

By learning a neural network model based on the plurality of loss functions, when the neural network model performs image conversion, the technical effect of obtaining a more realistic conversion result of the features of the partial image can be obtained. A specific process in which the plurality of loss functions are calculated and the neural network model is learned based on the calculated plurality of loss functions will be described with reference to FIG. 6 below.

Figure 4 is a schematic diagram illustrating a process for obtaining a converted image according to an embodiment of the present disclosure.

The computing device 100 according to an embodiment of the present disclosure may extract the latent vector 13 of the source image based on the source image information 11 (12). At this time, the source image information 11 may be information about an image whose identity is desired to be input when converting the target image information 10 and obtaining the converted image information 15. For example, if you want to input the identity of the eyes, nose, mouth, etc. in the image of a specific person (Robert Downey Jr. as an example in FIG. 4) when converting the target image information 10, the image of the specific person from which the identity is to be extracted Information may be included in the source image information 11. For example, the computing device 100 inputs the source image information 11 into an ID embedding network (e.g., ArcFace, ShphereFace, etc.) to apply facial contours of the eyes, nose, mouth, lower duct, etc. when converting the image. The latent vector 13 of the source image containing the features can be extracted.

The computing device 100 according to an embodiment of the present disclosure may perform encoding by inputting the target image information 10 into the generation network 14 (generator) included in the neural network model. At this time, the target image information 10 may include an image with characteristics desired to be preserved during image conversion. For example, referring to FIG. 4, the target image information 10 includes posture, expression, texture, background, gaze direction, etc., excluding the facial outline features of the eyes, nose, mouth, and lower crown that can be applied when converting the image. It can include images of people who have it. By performing encoding on the target image information 10, it is possible to perform an operation with the latent vector 13 of the source image when performing image conversion through the generator network 14 included in the neural network model. Thereafter, the encoded target image information and the latent vector 13 of the source image can be input into the generation network 14 included in the neural network model through conditional normalization to obtain the converted image information 15. there is. The converted image information 15 is a result image to which the features of the source image information 11 to be applied when converting the image are applied, and the features to be preserved when converting the image of the target image information 10 are preserved. may include. For example, referring to FIG. 4, the facial contour features of Robert Downey Jr.'s eyes, nose, mouth, and lower body, which are desired to be applied when converting the image, are applied to the face image of an ordinary person (10) and are included in the neural network model. Through this, it is possible to obtain a “face image of an ordinary person (15) in which the facial contour features of Robert Downey Jr.'s eyes, nose, mouth, and lower jaw are applied, and the remaining features are maintained.” At this time, the generation network 14 of the neural network model used for image conversion can utilize technologies such as the generation model of GAN (Generative Adversarial Networks), Unpaired Image-to-Image Translation, Variational Autoencoder (VAE), and Deepfake. , GAN Inversion technology can also be used to transform specific attributes in latent space. Image conversion technology using a neural network model 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 the range understandable by those skilled in the art. Meanwhile, according to an embodiment of the present disclosure, the computing device may extract a plurality of target partial image information based on target image information and a plurality of converted partial image information based on converted image information. , the specific process for this will be described later with reference to FIGS. 5A and 5B.

FIG. 5A is a schematic diagram illustrating a process for extracting a plurality of target partial image information based on target image information according to an embodiment of the present disclosure.

The computing device 100 according to an embodiment of the present disclosure extracts first segmentation information 22 for the target image using target image information 10 obtained directly or received from an external system. It can be done (21). At this time, deep learning-based segmentation technology can be used to extract the first segmentation information 22 for the target image information 10. The deep learning-based segmentation technology used at this time is Fully Convolutional Network (FCN). ), U-Net, DeepLab, Mask RCNN, etc. can be used.

The computing device 100 according to an embodiment of the present disclosure may receive first condition information for the extracted first division information 22. The first condition information refers to information that can be applied when extracting target partial image information from the first division information 22 for the target image information 10. For example, among the first segmentation information 22 for the target image information 10, one or more of the facial contours of the eyes, nose, mouth, or lower canal are input as the first condition information to extract target partial image information. It can be applied to . Specifically, when the first condition information is input to the eye area, when extracting the target image information, the eye area among the target image information 10 may be extracted as the first target image information. Additionally, according to another embodiment of the present disclosure, when the second condition information is input as an area for the nose, the area for the nose among the target image information 10 may be extracted as the second target partial image information. However, the area for the eyes or the area for the nose is only an example, and various other examples may be included in the condition information. Meanwhile, the first or second condition information may include information input by the user during learning and may be randomly assigned. In addition, the expressions such as first, second, etc. are only meant to distinguish between components and are not limited thereto.

In addition, the computing device 100 according to an embodiment of the present disclosure may generate a first target masking area 23-1 based on the first condition information and the extracted first division information 22, , A second target masking area 23-2 can be generated based on the second condition information and the extracted first division information 22, and the generated first target for the target image information 10. The first target partial image information 25-1 can be extracted using the masking area 23-1, and the generated second target masking area 23-2 is used for the target image information 10. The second target portion image information (25-2) can be extracted (24). Specifically, in FIG. 5A, when the area for the eyes is input as the first condition information among the first division information 22, a first target masking area 23-1 is created in which masking is applied to the area excluding the eyes. It can be. Additionally, when the nose area is input as the second condition information, a second target masking area 23-2 can be created in which masking is applied to the area excluding the nose. However, the area for the eyes or the area for the nose is only an example of the first or second condition information, and two or more areas such as the eyes, mouth, nose, and facial contour of the lower canal are designated as the first or second condition information. In addition, various examples can be entered as condition information. Additionally, the computing device 100 may extract first target partial image information 25-1 through element-wise product between the generated first target masking area 23-1 and the target image information 10. And the computing device 100 can extract the second target partial image information (25-2) through element-wise product between the generated second target masking area (23-2) and the target image information (10). There is (24). For example, in FIG. 5A, the target partial image information 25-1 of the eye area is extracted through an element-wise product between the target masking area 23-1 of the generated eye area and the target image information 10. The target portion image information 25-2 of the nose area may be extracted through an element-wise product between the generated target masking area 23-2 of the nose area and the target image information 10. The target partial image information 25-3 of the mouth area can be extracted through an element-wise product between the target masking area 23-3 of the generated mouth area and the target image information 10. At this time, when the target partial image information (25-1, 25-2, 25-3) is extracted, the generated target masking areas (23-1, 23-2, 23-3) and the target image In addition to the element-wise product between the information 10, various calculation methods that can be used by a person skilled in the art can be used. Meanwhile, the extracted target partial image information (25-1, 25-2, 25-3) can be divided into facial components such as eyes, nose, mouth, etc. and used individually in the process of learning the neural network model. . Additionally, truth loss functions may be calculated based on the extracted target partial image information 25-1, 25-2, and 25-3, and the specific process for this will be described later with reference to FIG. 6.

Meanwhile, FIG. 5B is a schematic diagram illustrating a process of extracting a plurality of converted partial image information based on converted image information according to an embodiment of the present disclosure.

The computing device 100 according to an embodiment of the present disclosure provides second segmentation information 32 for the converted image using converted image information 15 directly acquired or received from an external system. It can be extracted (31). At this time, deep learning-based segmentation technology can be used to extract the second segmentation information 32 for the converted image information 15. The deep learning-based segmentation technology used at this time is Fully Convolutional Network ( FCN), U-Net, DeepLab, Mask RCNN, etc. can be used.

The computing device 100 according to an embodiment of the present disclosure may receive the first condition information for the extracted second division information 32. The first condition information refers to information that can be applied when extracting target partial image information from the first division information 22 for the target image information 10, and the first condition information is the converted image information. The same can be applied when extracting the converted partial image information from the second division information 32 for (15). For example, the eye area among the first segmentation information 22 for the target image information 10 may be input as the first condition information and applied when extracting target partial image information. In addition, the eye area among the second division information 32 for the converted image information 15 may be input as the first condition information and applied in the same way when extracting the converted partial image information. Through this, in the process of learning the neural network model for image conversion, the corresponding two-part image information (target image and converted image) can be used as learning data to individually determine facial components such as eyes, nose, mouth, etc. there is. Specifically, when the first condition information is input to the eye area, the computing device 100 extracts the eye area from the target image information 10 as the first target partial image information when extracting target partial image information. Also, when the computing device 100 extracts the converted partial image information, the area for the eyes among the converted image information 15 can be extracted as the first converted partial image information. Likewise, when the second condition information is input to the nose area, when the computing device 100 extracts target image information, the area for the nose among the target image information 10 can be extracted as the second target image information. Also, when the computing device 100 extracts the converted partial image information, the area for the nose among the converted image information 15 can be extracted as the second converted partial image information. However, the area for the eyes or the area for the nose is only an example, and various other conditions may be included in the first or second condition information. Meanwhile, condition information may include information input by the user during learning and may be randomly assigned.

In addition, the computing device 100 according to an embodiment of the present disclosure may generate the first converted masking area 33-1 based on the first condition information and the extracted second division information 32. And, the first converted partial image information 35-1 can be extracted using the first converted masking area 33-1 created for the converted image information 15 (34). For example, if the area for the eyes is input as the first condition information among the second division information 32, a first converted masking area 33-1 in which masking is applied to the area excluding the eyes will be created. You can. Additionally, when the nose area is input as the second condition information, a second converted masking area 33-2 may be created in which masking is applied to the area excluding the nose area.

In addition, the computing device 100 generates first converted partial image information 35-1 through element-wise multiplication between the generated first converted masking area 33-1 and the converted image information 15. It can be extracted, and the computing device 100 generates the second converted partial image information (35-) through element-wise multiplication between the generated second converted masking area (33-2) and the converted image information (15). 2) can be extracted (34). For example, in FIG. 5B, converted partial image information 35-1 of the eye area is obtained through element-wise product between the generated converted masking area 33-1 of the eye area and the converted image information 15. can be extracted, and the converted partial image information 35-2 of the nose area is obtained through element-wise product between the converted masking area 33-2 of the generated nose area and the converted image information 15. It can be extracted, and the converted partial image information 35-3 of the mouth area is extracted through an element-wise product between the converted masking area 33-3 of the generated mouth area and the converted image information 15. It can be. At this time, when the converted partial image information (35-1, 35-2, 35-3) is extracted, the generated converted masking areas (33-1, 33-2, 33-3) and the In addition to the element-wise product between the converted image information 15, various computational methods available to a person skilled in the art can be used. Meanwhile, the extracted converted partial image information (35-1, 35-2, 35-3) can be divided into facial components such as eyes, nose, mouth, etc. and used individually in the process of learning the neural network model. there is. Additionally, false loss functions may be calculated based on the extracted converted partial image information 35-1, 35-2, and 35-3, and the detailed process for this will be described later with reference to FIG. 6.

FIG. 6 is a schematic diagram illustrating a process of calculating a plurality of loss functions based on a plurality of target partial image information and a plurality of converted partial image information according to an embodiment of the present disclosure.

Referring to FIG. 6, the computing device 100 inputs the extracted first target portion image information 25-1 into the first discrimination network 40-1 to generate a first real loss function 41-1. 1) can be calculated, and the extracted second target part image information (25-2) can be input into the second discriminator network (40-2) to calculate the second truth loss function (41-2). And, the extracted third target portion image information 25-3 can be input into the third discriminator network 40-3 to calculate a third real loss function 41-3. For example, the computing device 100 inputs the extracted target portion image information 25-1 of the eye region to the eye region determination network 40-1 to create an eye region true loss function 41-1. ) can be calculated, and the nose region true loss function (41-2) can be calculated by inputting the extracted target portion image information (25-2) of the nose region into the nose region discrimination network (40-2). In addition, the extracted target portion image information 25-3 of the mouth area can be input to the mouth area determination network 40-3 to calculate the mouth area real loss function 41-3.

In addition, the computing device 100 according to an embodiment of the present disclosure inputs the extracted first converted partial image information 35-1 into the first determination network 40-1. Thus, the first false loss function (51-1) can be calculated, and the extracted second converted partial image information (35-2) is input to the second discriminator network (40-2) to obtain a second The Fake loss function 51-2 can be calculated, and the extracted third converted partial image information 25-3 is input to the third discriminator network 40-3 to generate the third Fake loss function. The loss function (51-3) can be calculated. For example, the computing device 100 inputs the extracted partial image information 35-1 of the eye region into the eye region determination network 40-1 to create an eye region fake loss function 51. -1) can be calculated, and the converted partial image information (35-2) of the extracted nose region is input to the nose region determination network (40-2) to create a nose region fake loss function (51-2). ) can be calculated, and the converted partial image information (35-3) of the extracted mouth area is input to the mouth area determination network (40-3) to generate a mouth area fake loss function (51-3). It can be calculated. At this time, the eye, nose, and mouth regions are examples, and are not limited thereto, and discriminant networks for various regions may be used in the process of calculating a plurality of loss functions.

Additionally, the computing device 100 determines the first discriminant network 40 based on one or more of the first real loss function 41-1 or the first false loss function 51-1. -1), and the second discriminant network (40-) based on one or more of the second true loss function (41-2) or the second false loss function (51-2). 2) can be learned. For example, when the target portion image information 25-1 of the eye region is input to the eye region determination network 40-1, the computing device 100 inputs the input data to the eye region determination network 40-1. The eye region determination network 40-1 can be trained based on the eye region real loss function 41-1 to determine that is real. Additionally, when the converted partial image information 35-1 of the eye region is input to the eye region determination network 40-1, the computing device 100 determines that the eye region determination network 40-1 receives the input data. The eye region discrimination network 40-1 can be trained based on the eye region fake loss function 51-1 to determine that it is fake. Additionally, the same process can be applied to the nose area discrimination network 40-2 or the mouth area discrimination network 40-3, and the nose area discrimination network 40-2 or the mouth area discrimination network 40-3 is It can be learned. Through this, the determination networks 40-1, 40-2, and 40-3 determine whether the partial image information of each input region is image information converted through the generation network of FIG. 4 or target image information. It can be learned to distinguish well by facial area. However, the discrimination networks 40-1, 40-2, and 40-3 are not limited to the eye, nose, and mouth regions and may include examples of various regions and are not limited to three. Additionally, the computing device 100 calculates an average of the calculated plurality of loss functions, calculates the average of the calculated plurality of loss functions as a final loss function, and models the neural network based on the calculated final loss function. can be learned. For example, computing device 100 calculates the average of the truth loss functions 41-1, 41-2, and 41-3 as a final truth loss function, and based on the calculated final truth loss function, the computing device 100 calculates the average of the truth loss functions 41-1, 41-2, and 41-3. A neural network model can be trained, the average of the false loss functions 51-1, 51-2, and 51-3 is calculated as a final false loss function, and the neural network model is formed based on the calculated final false loss function. can be learned. Through this, the neural network model can be trained to distinguish well whether the input image information is image information converted through the generation network of FIG. 4 or target image information. However, when calculating the final loss function, various calculation methods other than the average of a plurality of loss functions may be used. Meanwhile, in a generative adversarial network, a generative network (generator) generates data, and a discriminator network (discriminator) determines whether the input data is generated data (fake) or real data (real). is learned so that the discriminator can determine that the data it generates is real (i.e., it can deceive the discriminator network), and the discriminator determines whether the input data is generated data (fake). It is trained to accurately determine whether it is real data, and the generator and discriminator are trained adversarially, resulting in the generator generating data that is difficult to distinguish from real data. This is a model to create. Therefore, according to various embodiments of the present disclosure, the partial image information of each region input to the discriminant networks 40-1, 40-2, and 40-3 included in the neural network model for image conversion of the present disclosure. It can be learned to distinguish for each face region whether it is converted image information or target image information (through the generation network of FIG. 4), and correspondingly, the generation network included in the neural network model for image transformation ( generator) allows the discriminator (40-1, 40-2, 40-3) to determine that the data it generates is real (i.e., the discriminator (40-1, 40-2, 40 -3) It can be learned to generate converted image information so that it can be deceived. Through this, the discrimination networks (40-1, 40-2, 40-3) of the neural network model are learned to individually and accurately discriminate facial components such as eyes, nose, mouth, etc., and the generation network of the neural network model is the target. When converting image information and obtaining the converted image information, the neural network model can be learned to produce more realistic results for each facial region.

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 minimally using 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.

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

Although the present disclosure has generally been described above as being capable of being implemented by a computing device, those skilled in the art will understand that the present 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 in combination with hardware and software. It will be well known that it can be implemented as a combination.

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 are applicable to 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 each of these may be implemented in other computer system configurations, including those capable of operating in conjunction with one or more associated devices.

The described embodiments of the 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 skilled in the art will also recognize removable optical media such as zip drives, magnetic cassettes, flash memory cards, cartridges, etc. It will be appreciated that other types of computer-readable media, such as the like, 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 (10)

A method of transforming an image utilizing a neural network model, performed by one or more processors of a computing device, comprising:
Obtaining target image information and image information converted through a generator included in the neural network model;
extracting a plurality of target partial image information based on the target image information;
extracting a plurality of converted partial image information based on the converted image information;
calculating a plurality of loss functions based on the plurality of target partial image information and the plurality of transformed partial image information through a plurality of discriminator networks; and
Learning the neural network model based on the calculated plurality of loss functions.
Including,
The step of training the neural network model based on the calculated plurality of loss functions includes:
calculating a final loss function based on an average of the calculated plurality of loss functions; and
training a generator included in the neural network model based on the calculated final loss function;
Including,
method.
delete According to claim 1,
The step of extracting a plurality of target partial image information based on the target image information includes:
extracting first segmentation information for the target image information;
Receiving first condition information for the extracted first division information; and
extracting first target partial image information based on the first condition information and the extracted first segmentation information;
Including,
The step of extracting a plurality of converted partial image information based on the converted image information includes:
extracting second segmentation information for the converted image information;
receiving the first condition information for the extracted second division information; and
extracting first converted partial image information based on the first condition information and the extracted second division information;
Including,
method.
According to claim 3,
The step of extracting first target partial image information based on the first condition information and the extracted first segmentation information includes:
generating a first target masking area based on the first condition information and the extracted first division information; and
extracting the first target partial image information using the generated first target masking area for the target image information;
Including,
The step of extracting first converted partial image information based on the first condition information and the extracted second segmentation information includes:
generating a first converted masking area based on the first condition information and the extracted second division information; and
extracting the first converted partial image information using the first converted masking area generated for the converted image information;
Including,
method.
According to claim 1,
Calculating a plurality of loss functions based on the plurality of target partial image information and the plurality of transformed partial image information through the plurality of discriminators includes:
A first loss based on first target partial image information extracted based on first condition information and the target image information and first converted partial image information extracted based on the first condition information and the converted image information calculating a function; and
A second loss based on second target partial image information extracted based on second condition information and the target image information and second converted partial image information extracted based on the second condition information and the converted image information. calculating a function;
Including,
method.
According to claim 5,
A first loss based on first target partial image information extracted based on first condition information and the target image information and first converted partial image information extracted based on the first condition information and the converted image information The steps to calculate the function are:
Inputting the first target portion image information into a first discriminator network to calculate a first truth loss function; and
Inputting the first converted partial image information into a first discriminator network to calculate a first fake loss function;
Including,
A second loss based on second target partial image information extracted based on second condition information and the target image information and second converted partial image information extracted based on the second condition information and the converted image information. The steps to calculate the function are:
Inputting the second target portion image information into a second discriminator network to calculate a second truth loss function; and
calculating a second fake loss function by inputting the second converted partial image information into a second discriminator network;
Including,
method.
According to claim 6,
The neural network model is,
Comprising one or more of the first decision network or the second decision network,
The step of training the neural network model based on the calculated plurality of loss functions includes:
training the first discriminant network based on one or more of the first real loss function or the first fake loss function; or
training the second discriminant network based on one or more of the second real loss function or the second fake loss function;
Including,
method.
delete 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 converting an image using a neural network model, the operations comprising:
Obtaining target image information and image information converted through a generator included in the neural network model;
extracting a plurality of target partial image information based on the target image information;
extracting a plurality of converted partial image information based on the converted image information;
calculating a plurality of loss functions based on the plurality of target partial image information and the plurality of transformed partial image information through a plurality of discriminator networks; and
An operation of training the neural network model based on the calculated plurality of loss functions.
Including,
The operation of training the neural network model based on the calculated plurality of loss functions includes:
calculating a final loss function based on an average of the calculated plurality of loss functions; and
An operation of training a generator included in the neural network model based on the calculated final 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,
Obtain converted image information through a generator included in the target image information and neural network model;
extracting a plurality of target partial image information based on the target image information;
extracting a plurality of converted partial image information based on the converted image information;
calculate a plurality of loss functions based on the plurality of target partial image information and the plurality of transformed partial image information through a plurality of discriminator networks; and
Configured to learn a neural network model based on the calculated plurality of loss functions,
The process of learning the neural network model based on the calculated plurality of loss functions is,
calculate a final loss function based on an average of the calculated plurality of loss functions; and
Including the process of training one generative network (generator) included in the neural network model based on the calculated final loss function,
Computing device.