KR102665707B1 - Face image conversion method using diffusion model - Google Patents

Abstract

According to one embodiment of the present disclosure, a method for training a neural network model for transforming an image is disclosed, which is performed by one or more processors of a computing device. The method includes receiving a target image and an original image; Using a neural network model, removing part or all of first image information containing noise and obtaining second image information; Comparing the second image information and the target image to calculate a feature loss function; Comparing the second image information and the original image to calculate a structure loss function; and training the neural network model based on at least one of the feature loss function or the structure loss function.

Description

Face image conversion method using diffusion model {FACE IMAGE CONVERSION METHOD USING DIFFUSION MODEL}

This disclosure relates to a facial image conversion method using a diffusion model, and more specifically, by exchanging identity characteristics between two people using a diffusion model, so that the characteristics of one person's face are transferred to a photo of another person. It relates to a method of providing a face image.

In the case of GAN (Generative Adversarial Network)-based models, which were widely used as existing image transformation models, model learning was difficult and sensitive to external parameters because learning was performed based on conflicting loss functions.

On the other hand, in the case of the Diffusion Model, learning is carried out in the direction of maximizing the similarity of the data, so it is stable and provides guidance for the desired image using the learned model under specific conditions. It is now possible to conditionally create images that satisfy .

However, in the case of face image generation using the existing diffusion model, it generates a simple image (unconditional generation) or only reflects the characteristics of the image of a famous person with abundant data, making it difficult to apply to the general public with insufficient learning data. It existed.

In addition, since the synthesis process is different for each image, there was a problem in that synthesis was difficult when there was a difference in pose between the two objects for which synthesis was desired.

Accordingly, there is an emerging need for a method to naturally transform facial images by applying desired identity characteristics to the image, synthesizing the facial features of the person being synthesized at the correct location, and processing the person's gaze correctly.

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

The problem of this disclosure is to provide a face image transformed so that the characteristics of one person's face are transferred to a photo of another person by exchanging identity characteristics between two people using a diffusion model.

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

A method performed by a computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The method includes receiving a target image and an original image; Using a neural network model, removing part or all of first image information containing noise and obtaining second image information; Comparing the second image information and the target image to calculate a feature loss function; Comparing the second image information and the original image to calculate a structure loss function; and training the neural network model based on at least one of the feature loss function or the structure loss function.

Alternatively, the neural network model includes a neural network model learned to obtain an image from which part or all of the noise has been removed by removing part or all of the noise for Gaussian distributed noise, A neural network model learned to obtain the second image information from which part or all of the noise is removed based on the original image and the mask condition when the original image and the mask condition are input to the neural network model. may include.

Alternatively, the mask condition may be determined by creating a masking area for a portion of the original image where conversion is desired and based on the created masking area.

Alternatively, calculating a feature loss function by comparing the second image information and the target image may include extracting a first identity feature vector based on the second image information, and extracting a first identity feature vector based on the target image. It may include extracting a target identity feature vector and calculating the feature loss function based on the extracted first identity feature vector and the target identity feature vector.

Alternatively, the structure loss function includes a region loss function or a gaze loss function, and calculating the structure loss function by comparing the second image information and the original image comprises: Comparing information and the original image to calculate a region loss function; And it may include calculating a gaze loss function by comparing the second image information and the original image.

Alternatively, the step of calculating a region loss function by comparing the second image information and the original image includes extracting first segmentation region information based on the second image information, and extracting first segmentation region information based on the original image. It may include extracting original segmentation area information and calculating the area loss function based on the extracted first segmentation area information and the original segmentation area information.

Alternatively, the step of calculating a gaze loss function by comparing the second image information and the original image may include extracting first gaze direction information based on the second image information, and It may include extracting original gaze direction information based on the original image, and calculating the gaze direction loss function based on the extracted first gaze direction information and the original gaze direction information.

Alternatively, the structure loss function includes a region loss function or a gaze loss function, and training the neural network model based on at least one of the feature loss function or the structure loss function includes: calculating a transformation loss function based on the feature loss function and the structure loss function; and training the neural network model based on the transformation loss function, wherein the transformation loss function is a weighted sum of the feature loss function, the region loss function, and the gaze loss function. It can be calculated as

According to an embodiment of the present disclosure for realizing the above-described object, a computer program stored in a computer-readable storage medium is disclosed. When executed on one or more processors, the computer program causes the one or more processors to perform operations for learning a neural network model for converting an image, the operations comprising: receiving a target image and an original image;

Using a neural network model, removing part or all of first image information containing noise and obtaining second image information; Comparing the second image information and the target image to calculate a feature loss function; Comparing the second image information and the original image to calculate a structure loss function; and training the neural network model based on at least one of the feature loss function or the structure loss function.

A computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The device includes at least one processor; and a memory, wherein the processor receives a target image and an original image; Using a neural network model, remove part or all of first image information containing noise and obtain second image information; Compare the second image information and the target image to calculate a feature loss function; Compare the second image information and the original image to calculate a structure loss function; And may be configured to learn the neural network model based on at least one of the feature loss function or the structure loss function.

The present disclosure can achieve the effect of naturally converting a face image by applying desired identity characteristics to the image, synthesizing the facial features of the person being synthesized at the correct position, and processing the person's gaze correctly.

Meanwhile, the effects of the present disclosure are not limited to the effects mentioned above, and various effects may be included within the range apparent to those skilled in the art from the contents described below.

1 is a block diagram of a computing device for converting a face image using a diffusion 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.
Figure 3 is a flowchart showing a method for learning a neural network model for converting an image according to an embodiment of the present disclosure.
Figure 4 shows a process in which a neural network model is learned to obtain an image from which part or all of the noise has been removed by removing part or all of the Gaussian distributed noise according to an embodiment of the present disclosure. This is a schematic diagram to explain.
5 is a diagram in which part or all of the noise is removed based on the original image, target image, and mask condition when the original image, target image, and mask condition are input to the neural network model according to an embodiment of the present disclosure. 2 This is a schematic diagram showing the process by which a neural network model is learned to acquire image information.
FIG. 6A is a schematic diagram illustrating a process for calculating a feature loss function by comparing second image information and a target image according to an embodiment of the present disclosure.
FIG. 6B is a schematic diagram illustrating a process of calculating a region loss function by comparing second image information and an original image according to an embodiment of the present disclosure.
FIG. 6C is a schematic diagram illustrating a process of calculating a gaze direction loss function by comparing second image information and an original image 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 may transmit signals, for example, with one or more data packets (e.g., data and/or signals from one component interacting with other components in a local system, a distributed system, to other systems and over a network such as the Internet). Depending on the data being transmitted, they may communicate through local and/or remote processes.

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

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

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

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

The description of the presented embodiments is provided to enable anyone skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Therefore, the present invention is not limited to the embodiments presented herein. The present invention is to be interpreted in the broadest scope consistent with the principles and novel features presented herein.

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

1 is a block diagram of a computing device for converting a face image using a diffusion 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 components for performing the computing environment of the computing device 100, and only some of the disclosed components may configure the computing device 100.

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

The processor 110 may be composed of one or more cores, and may include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU) of a computing device. unit) may include a processor for data analysis and deep learning. The processor 110 may read a computer program stored in the memory 130 and perform data processing for machine learning according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor 110 may perform an operation for learning a neural network 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), and may also 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 may 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 the relationship between nodes based on links within a neural network, it may refer to nodes that do not have other input nodes connected by links. Similarly, the final output node may refer to one or more nodes that do not have an output node in their relationship with other nodes among the nodes in the neural network. Additionally, hidden nodes may refer to nodes constituting a neural network other than the first input node and the last output node.

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

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

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

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

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

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

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

The computing device 100 according to an embodiment of the present disclosure may directly acquire an “image for conversion using a diffusion model” or receive it from an external system. The external system may be a server, database, etc. that stores and manages images for post-processing. The computing device 100 may use an image directly acquired or received from an external system as “input data for converting a facial image using a diffusion model.”

The computing device 100 may receive the target image and the original image (S110). For example, the original image may include an image that is the target of conversion, and the target image may include an image with characteristics to be reflected in the original image. Specifically, when converting a face included in an image, the original image may include an image in which only the identity of the face is subject to conversion while maintaining the posture, placement of facial features, gaze direction, etc., and the target image may be may include an image containing the identity characteristics of the face to be reflected in the original image. Meanwhile, although a face image is used as an example, the target image and original image are not limited to face images and may include various images.

The computing device 100 may use a neural network model to remove part or all of the noise from first image information containing noise and obtain second image information (S120). At this time, the terms first image information and second image information are only used to distinguish an image from which noise is partially or completely removed from an image containing noise, and are not limited to referring to a specific image. .

For example, the first image information containing noise may include isotropic Gaussian distributed noise, and the computing device 100 may partially remove noise from the Gaussian distributed noise using the neural network model. The image can be acquired as second image information. Additionally, the computing device 100 may obtain an image from which the noise has been completely removed by repeating the process of removing the noise using the neural network model, and in this case, the image from which the noise has been completely removed will be included in the second image information. An image containing noise immediately before an image from which noise has been completely removed may be included in the first image information. In this regard, the specific process in which part or all of the noise is removed from the first image information containing noise by using a neural network model and the second image information is acquired will be described later with reference to FIG. 5. Meanwhile, the neural network model includes a neural network model learned to obtain an image from which part or all of the noise has been removed by removing part or all of the noise with respect to Gaussian distributed noise, and the neural network model When the original image and the mask condition are input, it may include a neural network model learned to obtain the second image information from which part or all of the noise is removed based on the original image and the mask condition. You can. The specific process by which the neural network model is trained to remove part or all of the Gaussian distributed noise to obtain an image from which part or all of the noise has been removed will be described later with reference to FIG. 4. Meanwhile, the mask condition can be determined by creating a masking area for the part of the original image received through step S110 for which conversion is desired and based on the created masking area. The specific process for this is shown in FIG. 5 below. It is described later.

The computing device 100 may calculate a feature loss function by comparing the target image received through step S110 and the second image information obtained through step S120 (S130-1). Specifically, the computing device 100 extracts a target identity feature vector based on the target image received through step S110, and extracts a first identity feature vector based on the second image information obtained through step S120. A feature vector may be extracted, and the feature loss function may be calculated based on the extracted first identity feature vector and the target identity feature vector. The calculated feature loss function can be used in the process of learning a neural network model for image conversion, and the specific process will be described later with reference to FIGS. 5 and 6A. Meanwhile, the feature loss function may be obtained through comparison of the identity feature vectors (first and target), but is not limited to the above examples and may be obtained through various methods.

The computing device 100 may calculate a structure loss function by comparing the original image received through step S110 and the second image information obtained through step S120 (S130-2). At this time, the structure loss function may include a region loss function or a gaze loss function, and the structure loss function is calculated by comparing the original image received through step S110 and the second image information obtained through step S120. The calculating step includes calculating a region loss function by comparing the second image information and the original image; And it may include calculating a gaze loss function by comparing the second image information and the original image. Specifically, the computing device 100 extracts first segmentation area information based on the second image information, extracts original segmentation area information based on the original image, and extracts the extracted first segmentation area information. 1 The region loss function can be calculated based on the division region information and the original division region information. Additionally, the computing device 100 extracts first gaze direction information based on the second image information, extracts original gaze direction information based on the original image, and extracts the extracted first gaze direction information. And the gaze direction loss function may be calculated based on the original gaze direction information. The calculated structure loss function can be used in the process of learning a neural network model for image conversion, and the specific process will be described later with reference to FIGS. 5, 6B, and 6C. Meanwhile, the structure loss function may be obtained through comparison of the segmentation area information (first and original) and comparison of the gaze direction information (first and original), but in the above example It is not limited to these and can be obtained through various methods.

The computing device 100 may train the neural network model based on at least one of the feature loss function calculated through step S130-1 or the structure loss function calculated through step S130-2 (S140).

For example, the computing device 100 may calculate a transformation loss function based on the feature loss function and the structure loss function, and train the neural network model based on the transformation loss function. Specifically, the transformation loss function may be calculated as a weighted sum of the feature loss function, the region loss function, and the gaze loss function. At this time, the computing device 100 calculates a transformation loss function by adjusting the weights of the feature loss function, the region loss function, and the gaze direction loss function, and learns the neural network model based on the transformation loss function, When performing image conversion using the neural network model, a technical effect can be obtained that can focus on reflecting desired characteristics. Meanwhile, the specific process in which the transformation loss function is calculated and the neural network model is learned based on it will be described later with reference to FIG. 5.

Figure 4 shows a process in which a neural network model is learned to obtain an image from which part or all of the noise has been removed by removing part or all of the Gaussian distributed noise according to an embodiment of the present disclosure. This is a schematic diagram to explain.

Referring to FIG. 4, the computing device 100 according to an embodiment of the present disclosure removes part or all of the isotropic Gaussian distributed noise 10-4, thereby reducing the noise to A neural network model can be trained to obtain an image (10-1, 10-2, or 10-3) from which part or all of the image has been removed. At this time, the neural network model may include a conditional noise prediction model. In addition, the conditional noise prediction model may include a U-Net structure where the input and output have the same size, and the noise-containing image x(t) and the diffusion time step t are used as input. And, diffusion noise included in the noise-containing image x(t) can be predicted and output.

The computing device 100 repeats the process of gradually adding random Gaussian distribution noise over T number of type steps (time steps) to the image x(0)(10-1) that does not contain noise (from 10-2). direction of 10-3), and as a result, a forward process can be performed to obtain isotropic Gaussian distributed noise x(T) (isotropic gaussian distributed noise) (10-4). The forward process according to an embodiment of the present disclosure can be performed using the following equation.

In Equation 1 above, can be used as a hyperparameter in the process of calculating the diffusion coefficient, can be set to any value, 0< < < … < It can be set to a value of <1. for example This has a value of 0.0001 can have a value of 0.02, from until The value of may increase linearly or according to a cosine function, and T, which means the total number of diffusion steps, may be set to 1000. but, It is only an example that increases linearly or according to a cosine function, and can be increased in various ways. The amount of increase can be determined. also, may mean random Gaussian distributed noise. A general representation of the noise-containing image x(t)(10-3) at time step t relative to the noise-free image x(0)(10-1) and the diffusion noise included. The equation can be expressed as follows.

According to an embodiment of the present disclosure, equations (1) and (2) in Equation 2 are the diffusion coefficient It is a way to express the specific meaning of . Specifically, in equation (1) of Equation 2 above, the diffusion coefficient at a specific time step t The hyperparameters above are 1 It can be calculated by subtracting the diffusion coefficient in equation (2) of Equation 2 above. may mean the diffusion coefficient sequentially accumulated from time step 1 to t.

In addition, equation (3) in Equation 2 is the diffusion coefficient of the image x(t)(10-3) containing the noise and the image x(0)(10-1) containing the noise at time step t. ( ) and random Gaussian distributed noise ( ) is expressed as a formula for .

Through this, the computing device 100 repeats the process of gradually adding random Gaussian distribution noise over T type steps (time steps) to the image x(0)(10-1) that does not contain noise (10 direction from -2 to 10-3), and as a result, a forward process can be performed to obtain isotropic Gaussian distributed noise x(T) (10-4). However, the forward process is not limited to the example of Equation 2 above, and various processes that add noise to data may be included in the forward process.

In addition, the computing device 100 generates a random Gaussian distribution little by little over T type steps (time steps) for the isotropic Gaussian distributed noise x(T) (10-4) in the opposite direction to the forward process. Learn the neural network model to repeat the process of removing noise (in the direction from 10-3 to 10-2) and perform the reverse process to obtain an image x(0)(10-1) that does not contain noise as a result. You can do it. In this regard, the formula representing the reverse execution process can be expressed as follows.

In Equation 3, equation (1) is the noise prediction result predicted by the conditional noise prediction model for “the image containing the noise x(t)(10-3)” ( ) is a formula that represents the reverse process of obtaining “the previous image x(t-1)(10-2) from which some of the noise has been removed.” In addition, equation (2) in Equation 3 above is the diffusion coefficient at the current time step t means, and in equation (3) in equation 3 above, means the dispersion parameter, diffusion coefficient It can be calculated based on . However, the reverse process is not limited to Equation 3 above, and various processes for removing noise from data containing noise may be included in the reverse process.

Specifically, the computing device 100 inputs an image x(t)(10-3) containing noise and a time step t to the conditional noise prediction model, and matches the noise prediction result predicted by the conditional noise prediction model with the actual noise prediction result. The noise prediction loss can be calculated by comparing the included diffusion noise, and the conditional noise prediction model can be trained by performing gradient down according to the noise prediction loss.

For example, the noise prediction loss calculated by comparing the noise prediction result predicted by the conditional noise prediction model with the diffusion noise actually included can be expressed by the following formula.

In Equation 4 above, the noise prediction loss is the diffusion noise actually included ( ) and the noise prediction result predicted by the conditional noise prediction model ( ) can be calculated by comparing. However, the noise prediction loss is not limited to the example of Equation 4, and various loss functions calculated by comparing the noise prediction result and the actually included diffusion noise may be included in the noise prediction loss.

Additionally, the neural network model predicts the diffuse noise included in the noise-containing image x(t)(10-3) using the learned conditional noise prediction model, and partially removes the predicted diffuse noise. It can be learned to obtain “image x(t-1)(10-2) with some of the noise removed.” In addition, the neural network model predicts the diffusion noise included in the “image x(t)(10-3) containing the noise” and repeats the process of removing the predicted diffusion noise one or more times, “the noise By removing all the diffusion noise included in the image x(t)(10-3), it can be learned to obtain an image x(0)(10-1) from which all the noise has been removed. Meanwhile, by using an additional method in the process of removing diffusion noise from an image containing noise using the neural network model, the neural network model can be trained to generate a converted image. The specific process for this is shown in the figure below. This is explained later in Section 5.

5 is a diagram in which part or all of the noise is removed based on the original image, target image, and mask condition when the original image, target image, and masked condition are input to the neural network model according to an embodiment of the present disclosure. 2 This is a schematic diagram showing the process by which a neural network model is learned to acquire image information. At this time, the terms of the first image information and the second image information refer to an image in which some or all of the noise has been removed from an image containing noise (for example, the image 22-3 of x(1)'). (Example, image (22-4) of 'x(0)') is used to distinguish) and is not limited to meaning one specific image. For example, the first image information containing noise may include isotropic Gaussian distributed noise 22-1, and the computing device 100 may use the neural network model to generate the isotropic Gaussian distributed noise 22-1. x(T-1)' (22-2), an image from which noise has been partially removed, can be obtained as second image information. In addition, the computing device 100 can obtain an image x(0)'(22-4) from which the noise has been completely removed by repeating the process of removing the noise using the neural network model, and in this case, the noise has been completely removed. The obtained image x(0)'(22-4) may be included in the second image information, and includes noise at a stage immediately before the image x(0)'(22-4) from which the noise has been completely removed is acquired. The image x(1)'(22-3) may be included in the first image information.

The computing device 100 according to an embodiment of the present disclosure may perform facial image conversion using a diffusion model based on the original image 11 and the target image 21 directly acquired or received from an external system. At this time, the target image 21 may be an image with identity characteristics to be input when converting the original image 11. For example, if you want to input the identity features of the face in the image of a specific person (in the example of FIG. 5, a person wearing a gray suit) when converting the original image 11, the image of the specific person from which the identity features are to be extracted is as above. It may be included in the target image 21. Additionally, the original image 11 may include an image with features desired to be preserved during image conversion. For example, referring to FIG. 5, the original image 11 has posture, facial expression, background, gaze direction, etc. excluding identity features such as the facial outline of the eyes, nose, mouth, and lower jaw that can be applied when converting the image. May include images of people. However, the person image is only an example, and various images other than the person image may be included.

In addition, the computing device 100 according to an embodiment of the present disclosure may input the original image 11 and masked conditions 12-1 and 12-2 into the neural network model, and A neural network model may be learned to obtain the second image information from which part or all of the noise has been removed based on the image 11 and the mask conditions 12-1 and 12-2. At this time, the masked condition (12-1, 12-2) is such that a masking area (12-1) is created for the part of the original image (11) where conversion is desired, and the created masking area (12-12-1) is It can be decided based on 1). For example, if you want to transform the face part of the original image 11, segmentation technology based on deep learning can be used to create a masking area 12-1 for the face part. Deep learning-based segmentation technologies such as Fully Convolutional Network (FCN), U-Net, DeepLab, and Mask RCNN can be used, but are not limited to this and can be used to create a masking area (12-1) for the part where conversion of the original image is desired. Various examples can be used in the creation process. Specifically, the computing device 100 may utilize segmentation technology based on deep learning to create a masking area 12-1 for the facial part desired to be converted in the original image 11, and the generated masking An inverted masking area 12-2 can be created based on the area 12-1. The inverted masking area 12-2 is where random Gaussian distributed noise is gradually removed over T number of type steps (time step) with respect to isotropic Gaussian distributed noise x(T) (22-1). The process is repeated (in the direction from 22-2 to 22-3), and as a result, a converted image x(0)'(22-4) containing no noise is obtained, when converting the original image 11. It can be used to maintain information with characteristics you want to preserve. (For example, referring to FIG. 5, the black portion in the inverted masking area 12-2 may include information with features desired to be preserved when converting the original image 11, such as “eyes, nose, It may include “features other than identity features such as the facial outline of the mouth and lower diaphragm.”)

Specifically, the computing device 100 extracts original background information through element-wise product between the generated masking area 12-1 and the original image information 10, and applies the extracted original background information to the neural network model. By inputting the random Gaussian distribution noise (direction from 22-2 to 22-3), the original image is obtained in the process of obtaining a converted image x(0)'(22-4) that does not contain noise. In (11), the converted image x(0)'(22-4) with the original background information maintained can be obtained. In addition, the computing device 100 performs an element-wise product between the inverted masking area 12-2 and an isotropic Gaussian distributed noise x(T) 22-1 for the part desired to be transformed. By extracting information and inputting the extracted information about the part for which transformation is desired into the neural network model, random Gaussian distribution noise is removed (direction from 22-2 to 22-3), and as a result, the transformed noise does not contain noise. In the process of acquiring image x(0)'(22-4), image x(0)'(22-4) in which the portion of the original image 11 that is desired to be converted may be converted. Therefore, when the neural network model for converting an image is learned, the mask conditions (12-1, 12-2) are input so that the neural network model is learned to perform conversion by selecting only the part to be converted from the image to be converted. It can be. Meanwhile, when extracting the original background information or information about the part for which conversion is desired, various computational methods available to those skilled in the art can be used in addition to element-wise multiplication.

Referring to FIG. 5, according to an embodiment of the present disclosure, the neural network model removes part or all of the noise from the first image information (eg, the image of 22-1) containing noise, In the process of acquiring the second image information (eg, the image of 22-2), the neural network model is learned based on the transformation loss function 31, resulting in a transformed image x(0) that does not contain noise. '(22-4) can be learned to obtain. At this time, the transformation loss function 31 may be calculated based on the feature loss function and the structure loss function, and the structure loss function may include a region loss function or a gaze loss function. Specifically, the conversion loss function 31 can be calculated using the following formula.

In Equation 5, the conversion loss function 31 is “a feature loss function for reflecting the desired identity of the target image 21 in the original image 11,” and “the converted image is the original image 11.” ) is calculated as a weighted sum of “region loss function for ensuring that the converted image has posture, facial expression, etc. excluding the identity features of )” and “gaze direction loss function for ensuring that the converted image has the gaze direction of the original image (11).” You can. In addition, the neural network model is learned based on the transformation loss function 31 to “reflect the desired identity of the target image 21 with respect to the original image 11, and the identity features of the original image 11 may be learned to obtain an image x(0)'(22-4)" having posture, facial expression, and gaze direction excluding At this time, the second image information x(T-1)' can be obtained by removing part or all of the random Gaussian distribution noise from the information x(T)(22-1). However, if it is not learned based on the noise removal process of the neural network model, the conversion loss may be performed so that the “image x(0)’(22-4)” without noise is the same as the original image 11. When the neural network model is learned based on the function 31, the noise of the neural network model is obtained so that the image x(0)'(22-4), which is not completely identical to the original image 11 and has the desired portion converted As a result, the computing device 100 may use the neural network model learned through the example (i.e., mask condition or transformation loss function) described above in FIG. 5 to create an isotropic Gaussian distribution noise x(T). (isotropic gaussian distributed noise) For (22-1), remove part or all of the random Gaussian distributed noise (direction from 22-2 to 22-3), and obtain the transformed image x(0)'( 22-4), in addition, according to an embodiment of the present disclosure, the computing device 100 can focus on reflecting the desired characteristics of the target image by performing original image transformation using the neural network model. , it is possible to achieve the technical effect of selecting and converting only the part to be converted while maintaining the features that are desired to be preserved in the original image, while calculating the feature loss function, area loss function, and gaze direction loss function. The specific process will be described later with reference to FIGS. 6A to 6C.

Figures 6A to 6C show the process of calculating three loss functions used to calculate the conversion loss function according to an embodiment of the present disclosure. This is a schematic diagram. The term second image information 23, which will be described later in FIGS. 6A to 6C of the present disclosure, refers to a portion or part of an image containing noise (eg, the image 22-3 of 'x(1)' in FIG. 5). It is only used to distinguish images from which all noise has been removed (eg, image 22-4 of x(0)' in FIG. 5), and is not limited to meaning a specific image.

First, FIG. 6A is a schematic diagram showing a process for calculating a feature loss function by comparing second image information and a target image according to an embodiment of the present disclosure.

Referring to FIG. 6A, the computing device 100 extracts a target identity feature vector 32-1 based on the target image 21 and extracts a first identity feature vector 32-1 based on the second image information 23. (identity) feature vector (32-2) can be extracted. For example, the computing device 100 inputs the target image 21 and the second image information 23 into an ID embedding network (e.g. ArcFace, ShphereFace, Smooth-Swap, Partial-fc, etc.) to create an “image.” “The target identity feature vector (32-1)” and the first identity feature vector (32-1), which includes identity features such as eyes, nose, mouth, and facial outline of the lower body that can be applied during conversion. 2) can be extracted respectively. In addition, the computing device 100 according to an embodiment of the present disclosure compares the extracted target identity feature vector 32-1 and the first identity feature vector 32-2 to determine the feature. A loss function can be calculated (33), and the neural network model can be trained based on the calculated feature loss function. For example, a feature loss function can be calculated by performing a cosine similarity operation between the extracted target identity feature vector 32-1 and the first identity feature vector 32-2 (33) ), the neural network model can be trained based on the feature loss function by using the calculated feature loss function as a loss term when learning the neural network model. However, it is not limited to the cosine similarity operation, and various operations may be used in the process of calculating the feature loss function. In addition, by learning the neural network model based on the feature loss function, a technical effect can be obtained that allows the identity features of the target image to be well reflected in the part to be converted when converting the image.

Second, FIG. 6B is a schematic diagram showing a process of calculating a region loss function by comparing second image information and the original image according to an embodiment of the present disclosure.

Referring to FIG. 6B, the computing device 100 extracts original segmentation area information 42-1 based on the original image 11 and performs first segmentation based on the second image information 23. (Segmentation) Area information (42-2) can be extracted. For example, the computing device 100 inputs the original image 11 and the second image information 23 into a deep learning-based segmentation network (e.g. Face Parser, etc.) to create “the original image 11.” ) “The original segmentation area information 42-1” and the first segmentation area information 42, which includes information with features that are desired to be preserved during conversion (e.g., posture, arrangement of facial features) -2) can be extracted respectively. In addition, the computing device 100 according to an embodiment of the present disclosure compares the extracted original segmentation area information 42-1 and the first segmentation area information 42-2 to determine the area. A loss function can be calculated (43), and the neural network model can be trained based on the calculated region loss function. For example, the region loss function can be calculated by calculating L1 loss between the extracted original segmentation region information 42-1 and the first segmentation region information 42-2 (43) ), the neural network model can be trained based on the region loss function by using the calculated region loss function as a loss term when learning the neural network model. However, it is not limited to the L1 loss calculation, and various operations may be used in the process of calculating the region loss function. In addition, by learning the neural network model based on the region loss function, a technical effect can be obtained that allows the features of the original image to be well maintained for the part that is desired to be preserved when converting the image.

Lastly, Figure 6C is a schematic diagram showing the process of calculating a gaze direction loss function by comparing second image information and the original image according to an embodiment of the present disclosure.

Referring to FIG. 6C, the computing device 100 extracts original gaze direction (gaze) information 52-1 based on the original image 11, and extracts first gaze information 52-1 based on the second image information 23. Direction (gaze) information (52-2) can be extracted (51). For example, the computing device 100 inputs the original image 11 and the second image information 23 into a gaze direction detection network based on deep learning (e.g. GazeNet, HRNetGaze, keypoint estimator, etc.) Extract “the original gaze direction information 52-1, which includes the gaze direction information to be preserved when converting the original image 11” and the first gaze direction information 52-2, respectively. can do. In addition, the computing device 100 according to an embodiment of the present disclosure compares the extracted original gaze direction (gaze) information 52-1 and the first gaze direction (gaze) information 52-2 to determine the gaze direction. A loss function can be calculated (53), and the neural network model can be trained based on the calculated gaze direction loss function. For example, the gaze direction loss function can be calculated by calculating L1 loss between the extracted original gaze direction (gaze) information 52-1 and the first gaze direction (gaze) information 52-2 (53) ), the neural network model can be trained based on the gaze direction loss function by using the calculated gaze direction loss function as a loss term when learning the neural network model. However, it is not limited to the L1 loss calculation, and various operations may be used in the process of calculating the gaze direction loss function. In addition, by learning the neural network model based on the gaze direction loss function, a technical effect can be obtained that allows the gaze direction of the original image to be well maintained with respect to the gaze direction desired to be preserved when converting the image.

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

Data structures can be divided into linear data structures and non-linear data structures depending on the type of data structure. A linear data structure may be a structure in which only one piece of data is connected to another piece of data. Linear data structures may include List, Stack, Queue, and Deque. A list can refer to a set of data that has an internal order. The list may include a linked list. A linked list may be a data structure in which data is connected in such a way that each data is connected in a single line with a pointer. In a linked list, a pointer can contain connection information to the next or previous data. A linked list can be expressed as a singly linked list, doubly linked list, or circular linked list depending on its form. 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 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 disclosure can be implemented in combination with computer-executable instructions and/or other program modules that can be executed on one or more computers and/or hardware. It will be well known that it can be implemented as a combination of software.

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

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

Computers typically include a variety of computer-readable media. Computer-readable media can be any medium that can be accessed by a computer, and such computer-readable media includes volatile and non-volatile media, transitory and non-transitory media, removable and non-transitory media. Includes removable media. By way of example, and not limitation, computer-readable media may include computer-readable storage media and computer-readable transmission media. Computer-readable storage media refers to volatile and non-volatile media, transient and non-transitory media, removable and non-removable, implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Includes media. Computer readable storage media may include RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage. This includes, but is not limited to, a device, or any other medium that can be accessed by a computer and used to store desired information.

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

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

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

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

These drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like. For computer 1102, drive and media correspond to storing any data in a suitable digital format. Although the description of computer-readable media above refers to removable optical media such as HDDs, removable magnetic disks, and CDs or DVDs, those of ordinary skill in the art would also recognize zip drives, magnetic cassettes, flash memory cards, and cartridges. It will be appreciated that other types of computer-readable media may also be used in the exemplary 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 the drive and RAM 1112, including an operating system 1130, one or more application programs 1132, other program modules 1134, and program data 1136. All or portions of the operating system, applications, modules and/or data may also be cached in RAM 1112. It will be appreciated that the present disclosure may be implemented 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 communications 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) enables 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 cell tower. 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, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks can operate in the unlicensed 2.4 and 5 GHz 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 (8)

1. A method for training a neural network model for transforming an image, performed by one or more processors of a computing device, comprising:
Receiving a target image and an original image;
Using a neural network model, removing part or all of first image information containing noise and obtaining second image information;
Comparing the second image information and the target image to calculate a feature loss function;
Comparing the second image information and the original image to calculate a region loss function;
Extract first gaze direction information based on the second image information, extract original gaze direction information based on the original image, and based on the extracted first gaze direction information and the original gaze direction information. calculating a gaze direction loss function; and
Learning the neural network model based on at least one of the feature loss function, the region loss function, or the gaze direction loss function;
Including,
method.

According to claim 1,
The neural network model is,
With respect to Gaussian distributed noise, it includes a neural network model learned to obtain an image from which part or all of the noise has been removed by removing part or all of the noise,
A neural network model learned to obtain the second image information from which part or all of the noise is removed based on the original image and the mask condition when the original image and the mask condition are input to the neural network model. Including,
method.
According to claim 2,
The mask conditions are,
A masking area is created for the part of the original image where conversion is desired, and is determined based on the created masking area,
method.
According to claim 1,
Comparing the second image information and the target image to calculate a feature loss function,
Extract a first identity feature vector based on the second image information, extract a target identity feature vector based on the target image, and extract the extracted first identity feature vector and the target identity feature vector. calculating the feature loss function based on
Including,
method.
1. A method for training a neural network model for transforming an image, performed by one or more processors of a computing device, comprising:
Receiving a target image and an original image;
Using a neural network model, removing part or all of first image information containing noise and obtaining second image information;
Comparing the second image information and the target image to calculate a feature loss function;
Extract first segmentation area information based on the second image information, extract original segmentation area information based on the original image, and extract the extracted first segmentation area information and the original segmentation area. calculating a region loss function based on the information;
Comparing the second image information and the original image to calculate a gaze loss function; and
Learning the neural network model based on at least one of the feature loss function, the region loss function, or the gaze direction 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 training a neural network model for converting an image, the operations being :
Receiving a target image and a source image;
Using a neural network model, removing part or all of first image information containing noise and obtaining second image information;
Comparing the second image information and the target image to calculate a feature loss function;
Comparing the second image information and the original image to calculate a region loss function;
Extract first gaze direction information based on the second image information, extract original gaze direction information based on the original image, and based on the extracted first gaze direction information and the original gaze direction information. An operation of calculating a gaze direction loss function; and
An operation of training the neural network model based on at least one of the feature loss function, the region loss function, or the gaze direction 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,
receive a target image and a source image;
Using a neural network model, remove part or all of first image information containing noise and obtain second image information;
Compare the second image information and the target image to calculate a feature loss function;
Compare the second image information and the original image to calculate a region loss function;
Extract first gaze direction information based on the second image information, extract original gaze direction information based on the original image, and based on the extracted first gaze direction information and the original gaze direction information. Calculate the gaze direction loss function; and
Configured to train the neural network model based on at least one of the feature loss function, the region loss function, or the gaze direction loss function,
Computing device.