Electronic device for enhancing image quality and method for enhancing image quality using thereof

Abstract

The disclosed embodiments relate to an electronic device for improving image quality and a method for improving image quality using the same. A method for generating a second person image of high quality by performing image processing on a first person image of low quality using an artificial intelligence model according to the disclosed electronic device comprises the steps of identifying the first person image, applying the first person image to an artificial intelligence model, and obtaining the second person image output from the artificial intelligence model. The artificial intelligence model can recognize a first face by performing face identification and face recognition on the first person image using a face recognition artificial intelligence model, and obtain a second person image by performing image processing to improve image quality on an area corresponding to the first face using an image quality improvement artificial intelligence model, and output the second person image to the electronic device.

Description

Technology Field

[0001] The disclosed embodiments relate to an electronic device for improving image quality and a method for improving image quality using the same. Background Technology

[0002] Artificial Intelligence (AI) systems are computer systems that achieve human-level intelligence. Unlike existing rule-based smart systems, they are systems that learn, make judgments, and become smarter on their own. As AI systems improve in recognition accuracy and become capable of understanding user preferences more accurately with continued use, existing rule-based smart systems are gradually being replaced by deep learning-based AI systems.

[0003] Artificial intelligence technology consists of machine learning (deep learning) and elemental technologies utilizing machine learning.

[0004] Machine learning is an algorithmic technology that classifies and learns the features of input data on its own, and elemental technology is a technology that utilizes machine learning algorithms such as deep learning, and consists of technology fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control.

[0005] The various fields where artificial intelligence technology is applied are as follows.

[0006] Linguistic understanding is a technology that recognizes, applies, and processes human language and text, and includes Natural Language Processing, Machine Translation, Dialog Systems, Question Answering, and Speech Recognition / Synthesis.

[0007] Visual understanding is a technology that perceives and processes objects like human vision, and includes object recognition, object tracking, image retrieval, human recognition, scene recognition, 3D reconstruction / localization, and image enhancement.

[0008] Inference prediction is a technology that logically reasones and predicts by evaluating information, and includes knowledge-based reasoning, optimization prediction, preference-based planning, and recommendation.

[0009] Knowledge representation is a technology that automatically processes human experiential information into knowledge data, and includes knowledge construction (data generation / classification) and knowledge management (data utilization). Motion control is a technology that controls the autonomous driving of vehicles and the movement of robots, and includes motion control (navigation, collision, driving) and manipulation control (behavior control).

[0010] In addition, artificial intelligence technology can also be used to acquire images, such as photos or videos, and improve their quality. The problem to be solved

[0011] The disclosed embodiments aim to provide an electronic device for improving image quality and a method for improving image quality using the same.

[0012] One disclosed embodiment aims to provide a method for generating a high-quality person image by performing image processing on a low-quality person image using an artificial intelligence model.

[0013] One disclosed embodiment aims to provide a method for training an artificial intelligence model that generates a high-quality person image from a low-quality person image.

[0014] Meanwhile, the technical problems that the disclosed embodiments aim to solve are not limited to the technical problems described above. means of solving the problem

[0015] A method for generating a second person image of high quality by performing image processing on a first person image of low quality using an artificial intelligence model, disclosed as a technical means for achieving the aforementioned technical task, comprises the steps of identifying the first person image, applying the first person image to an artificial intelligence model, and obtaining the second person image output from the artificial intelligence model, wherein the artificial intelligence model recognizes the first face by performing face identification and face recognition on the first person image using a face recognition artificial intelligence model, and obtains the second person image by performing image processing to improve the image quality on an area corresponding to the first face using an image quality improvement artificial intelligence model, and outputs the second person image to the electronic device.

[0016] An electronic device for generating a second person image of high quality by performing image processing on a first person image of low quality using an artificial intelligence model disclosed as a technical means for achieving the above-described technical task includes a memory storing at least one instruction and a processor executing said at least one instruction, wherein the processor identifies the first person image by executing said at least one instruction, applies the first person image to an artificial intelligence model, and obtains the second person image output from said artificial intelligence model, and the artificial intelligence model recognizes the first face by performing face identification and face recognition on the first person image using a face recognition artificial intelligence model, and obtains the second person image by performing image processing to improve the image quality on an area corresponding to the first face using an image quality improvement artificial intelligence model, and outputs the second person image to said electronic device.

[0017] As a technical means for achieving the technical problem described above, a computer-readable recording medium may have a program recorded thereon for executing at least one of the embodiments of the disclosed method on a computer.

[0018] As a technical means for achieving the above-described technical task, an application stored on a recording medium may be for executing at least one function among the embodiments of the disclosed method. Brief explanation of the drawing

[0019] FIG. 1 is a diagram showing an electronic device acquiring a high-quality person image from a low-quality person image according to one embodiment. FIG. 2 is a flowchart of a method for an electronic device to obtain a high-quality person image from a low-quality person image according to one embodiment. FIG. 3 is a diagram showing an electronic device applying human images as training data to an artificial intelligence model according to one embodiment. FIG. 4 is a diagram showing an electronic device applying human images as training data to an artificial intelligence model according to one embodiment. FIG. 5 is a diagram showing an electronic device acquiring facial features from a person image using an artificial intelligence model according to one embodiment. FIG. 6 is a diagram showing that, according to one embodiment, an electronic device performs image quality improvement on an image quality improvement area received from a user. FIG. 7 is a flowchart of a method for improving image quality for an image quality improvement area received from a user by an electronic device according to one embodiment. FIG. 8 is a diagram showing that an electronic device improves image quality according to an image quality improvement direction received from a user, according to one embodiment. FIG. 9 is a flowchart of a method for improving image quality according to an image quality improvement direction received from a user by an electronic device, according to one embodiment. FIG. 10 is a block diagram of an electronic device according to one embodiment. FIG. 11 is a block diagram of a software module stored in the memory of an electronic device according to one embodiment. FIG. 12 is a block diagram of a server according to one embodiment. FIG. 13 is a block diagram of a software module stored in the memory of a server according to one embodiment. Specific details for implementing the invention

[0020] This specification clarifies the scope of the present invention and explains the principles of the present invention and discloses embodiments so that a person skilled in the art can practice the present invention. The disclosed embodiments may be implemented in various forms. The disclosed embodiments may be implemented alone or at least two or more embodiments may be combined.

[0021] Throughout the specification, the same reference numerals refer to the same components. This specification does not describe all elements of the embodiments, and general content in the art to which the invention pertains or content that overlaps between embodiments is omitted. The term "part" or "portion" used in this specification may refer to a hardware component, such as a processor or circuit, and / or a software component executed by a hardware component, such as a processor. Depending on the embodiments, it is possible for a plurality of "parts" to be implemented as a single unit or for a single "part" to include a plurality of elements. The operating principles and embodiments of the present invention will be described below with reference to the attached drawings.

[0022] Some embodiments of the present disclosure may be represented by functional block configurations and various processing steps. Some or all of these functional blocks may be implemented by various numbers of hardware and / or software configurations that execute specific functions. For example, the functional blocks of the present disclosure may be implemented by one or more microprocessors or by circuit configurations for a specific function. Additionally, for example, the functional blocks of the present disclosure may be implemented in various programming or scripting languages. The functional blocks may be implemented as algorithms executed on one or more processors. Furthermore, the present disclosure may employ prior art for electronic configuration, signal processing, and / or data processing, etc. Terms such as “mechanism,” “element,” “means,” and “configuration” may be used broadly and are not limited to mechanical and physical configurations.

[0023] Throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "electrically connected" with other components interposed between them. Furthermore, when a part is described as "including" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0024] Furthermore, the connecting lines or connecting members between the components depicted in the drawings are merely illustrative of functional connections and / or physical or circuit connections. In the actual device, connections between components may be represented by various alternative or added functional connections, physical connections, or circuit connections.

[0025] Additionally, terms including ordinal numbers, such as “first” or “second” as used in this specification, may be used to describe various components, but said components should not be limited by said terms. Such terms may be used for the purpose of distinguishing one component from another. For example, although first data and second data are described in this specification, they are used merely to distinguish that they are different data and should not be limited by them.

[0026] The electronic device according to the present disclosure may utilize an artificial intelligence model to generate a high-quality image from a low-quality image. Functions related to artificial intelligence according to the present disclosure are operated through a processor and memory. The processor may be composed of one or more processors. In this case, the one or more processors may be general-purpose processors such as a CPU, AP, or DSP (Digital Signal Processor), graphics-dedicated processors such as a GPU or VPU (Vision Processing Unit), or artificial intelligence-dedicated processors such as an NPU. The one or more processors control the processing of input data according to predefined operation rules or artificial intelligence models stored in memory. Alternatively, if the one or more processors are artificial intelligence-dedicated processors, the artificial intelligence-dedicated processors may be designed with a hardware structure specialized for processing a specific artificial intelligence model. The processor may perform a preprocessing process to convert data applied to the artificial intelligence model into a form suitable for application to the artificial intelligence model.

[0027] Artificial intelligence models can be created through learning. Here, being created through learning means that a basic artificial intelligence model is trained using multiple learning data by a learning algorithm, thereby creating a predefined set of behavioral rules or an artificial intelligence model configured to perform a desired characteristic (or objective). Such learning may be performed on the device itself where the artificial intelligence according to the present disclosure is executed, or it may be performed through a separate server and / or system. Examples of learning algorithms include supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but are not limited to the examples described above.

[0028] An artificial intelligence model may be composed of multiple neural network layers. Each of the multiple neural network layers has multiple weight values ​​and performs neural network operations through operations between the results of previous layers and the multiple weights. The multiple weights possessed by the multiple neural network layers can be optimized based on the learning results of the artificial intelligence model. For example, the multiple weights may be updated so that the loss value or cost value obtained from the artificial intelligence model during the learning process is reduced or minimized. Artificial neural networks may include deep neural networks (DNNs), such as Generative Adversarial Networks (GANs), Convolutional Neural Networks (CNNs), Deep Neural Networks (DNNs), Recurrent Neural Networks (RNNs), Restricted Boltzmann Machines (RBMs), Deep Belief Networks (DBNs), Bidirectional Recurrent Deep Neural Networks (BRDNNs), or Deep Q-Networks, but are not limited to the examples mentioned above.

[0029] The disclosed artificial intelligence model may be generated by learning multiple text data and image data input as training data according to predetermined criteria. The artificial intelligence model can generate result data and output result data by performing learned functions corresponding to the input data.

[0030] In addition, the disclosed artificial intelligence model may include a plurality of artificial intelligence models trained to perform at least one function.

[0032] Embodiments are described in detail below with reference to the drawings.

[0034] FIG. 1 is a diagram showing an electronic device acquiring a high-quality person image from a low-quality person image according to one embodiment.

[0035] Referring to FIG. 1, the electronic device (10) can generate a high-quality second person image by performing image processing on a low-quality first person image using an artificial intelligence model. The electronic device (10) can store the generated second person image in memory (17) or output it to a display unit (12-1).

[0036] In the disclosed embodiments, a low-quality image refers to an image that is typically difficult to identify or is degraded, such as an image with a pixel count smaller than a predetermined number, an image with blurred edges due to noise included in the image, or an image with incorrectly specified color temperature and hue.

[0037] In the disclosed embodiments, a high-definition image refers to an image that is typically easy to identify, such as an image containing a number of pixels greater than a predetermined number, an image with clear boundaries included in the image, and an image with accurate color temperature and hue.

[0038] In the disclosed embodiments, enhancement of image quality means improving the degradation factors of an image. For example, enhancement of image quality includes image processing such as resolution enhancement, noise removal, artifact removal, color adjustment, and sharpness enhancement.

[0039] In the disclosed embodiment, the electronic device (10) can identify a low-quality person image. The electronic device (10) can apply the low-quality person image to an artificial intelligence model. The artificial intelligence model may be trained to generate and output a high-quality person image by performing image processing on the low-quality person image. The electronic device (10) can store the high-quality person image output from the artificial intelligence model in memory (17) or output it through a display unit (12-1).

[0040] In the disclosed embodiment, the artificial intelligence model may be built in at least one of the electronic device (10) and the server (20). Hereinafter, the artificial intelligence model built in the electronic device (10) is described as an example, but is not limited thereto. The artificial intelligence model built in the electronic device (10) described below may be applied by analogy to the artificial intelligence model built in the server (20).

[0041] According to one embodiment, the electronic device (10) may include a computing device such as a mobile device (e.g., smartphone, tablet PC, etc.) and a general-purpose computer (PC, Personal Computer) capable of transmitting and receiving data to and from a server (20) through a network. Additionally, the electronic device (10) may include an Internet of Things (IoT) device, a home hub device (e.g., router, conversational AI speaker, etc.) connected to various IoT devices and the server (20). Furthermore, the electronic device (10) may include a computing device such as a mobile device (e.g., smartphone, tablet PC, etc.), a general-purpose computer (PC, Personal Computer), and a server on which an artificial intelligence model (19) is built.

[0042] According to one embodiment, the electronic device (10) can perform certain operations using an artificial intelligence model (19). For example, the electronic device (10) can perform operations such as identifying input data, classifying it, and outputting data corresponding to the input data using the artificial intelligence model (19).

[0043] According to one embodiment, the server (20) can transmit and receive data with the electronic device (10). For example, the server (20) can apply the data received from the electronic device (10) to an artificial intelligence model (29) and transmit the data output from the artificial intelligence model (29) to the electronic device (10). As another example, the server (20) can transmit data used to update the artificial intelligence model (19) built in the electronic device (10) to the electronic device (10).

[0044] According to one embodiment, the artificial intelligence model may be composed of a plurality of artificial intelligence models trained to perform a predetermined function. For example, the artificial intelligence model may include, but is not limited to, a preprocessing artificial intelligence model that performs preprocessing into a type applicable to the artificial intelligence model, an image quality classification artificial intelligence model that classifies the image quality of a person applied to the artificial intelligence model, a face detection artificial intelligence model that detects at least one face within a person image, a face recognition artificial intelligence model that identifies a person corresponding to the face identified from the person image, and an image quality improvement artificial intelligence model that generates a high-quality person image by performing image processing on a low-quality person image. There may be a plurality of artificial intelligence models that perform the same function, and one artificial intelligence model may perform at least one of the functions of the disclosed embodiments.

[0045] According to one embodiment, an artificial intelligence model can perform training using training data. For example, the artificial intelligence model can perform training for a face identification AI model, a face recognition AI model, and an image quality enhancement AI model by using multiple person images as training data. As another example, the artificial intelligence model can use a low-quality person image and a high-quality person image obtained by converting a high-quality person image as a pair of training data. As another example, the artificial intelligence model can use multiple person images classified by person as training data. As another example, the artificial intelligence model can use training data regarding an image quality enhancement area selected by a user. As another example, the artificial intelligence model can use training data regarding an image quality enhancement direction selected by a user.

[0046] According to one embodiment, the electronic device (10) can perform image quality improvement based on an input in which a user selects an image. For example, the electronic device (10) can perform image quality improvement by applying an image specified by the user for image quality improvement to an artificial intelligence model.

[0047] According to one embodiment, the electronic device (10) can perform image quality improvement of an image stored in the electronic device (10) during the time when the user is not using the electronic device (10). For example, the electronic device (10) can identify whether an image stored in the electronic device (10) is a low-quality image during idle time and perform image quality improvement.

[0048] According to the disclosed embodiment, the electronic device (10) can obtain a high-quality image by identifying a low-quality image and performing image quality improvement.

[0050] FIG. 2 is a flowchart of a method for an electronic device to obtain a high-quality person image from a low-quality person image according to one embodiment.

[0051] Referring to step S210, the electronic device (10) can identify a first person image of low quality. For example, it can identify a first person image stored in the memory (17) of the electronic device (10). As another example, the electronic device (10) can identify a first person image acquired using the camera of the electronic device (10). As another example, the electronic device (10) can identify a first person image shared on the web. As another example, the electronic device (10) can identify a first person image shared through an application.

[0052] According to one embodiment, the electronic device (10) can identify a first person image based on an input in which a user selects a low-quality person image. For example, the electronic device (10) can identify a first person image based on an input in which a user selects an image for which they want to improve the quality among a plurality of person images.

[0053] According to one embodiment, the electronic device (10) can identify a first person image of low quality based on a predetermined standard. For example, the electronic device (10) can identify an image containing a pixel count less than a predetermined number, or an image in which the frequency of the boundary line is less than or equal to a predetermined value.

[0054] According to one embodiment, the electronic device (10) can identify a first person image using an artificial intelligence model. For example, the electronic device (10) can identify a first person image of low quality from a plurality of images using a disclaimer artificial intelligence model. In this case, the disclaimer artificial intelligence model may be an artificial intelligence model trained to distinguish between a high-quality person image and a high-quality person image generated from a low-quality person image by an image quality improvement artificial intelligence model.

[0055] Referring to step S230, the electronic device (10) can apply the first person image to the artificial intelligence model. The electronic device (10) can apply the first person image identified in step S210 to the artificial intelligence model.

[0056] According to one embodiment, the electronic device (10) can apply a first person image, on which preprocessing has been performed, to an artificial intelligence model. For example, the electronic device (10) can detect a face within the first person image and apply information regarding the result of the face detection to the artificial intelligence model together with the first person image. As another example, the electronic device (10) can recognize a face within the first person image and apply information regarding a person corresponding to the recognized face to the artificial intelligence model together with the first person image. As yet another example, the electronic device (10) can apply information regarding the result of classifying the first person image as a low-quality image based on a predetermined criterion to the artificial intelligence model together with the first person image. Alternatively, the artificial intelligence model may directly perform preprocessing on the input first person image without performing the preprocessing described above.

[0057] According to one embodiment, the electronic device (10) can apply information regarding an input selecting a part of the image quality of a person included in a first person image received from a user to an artificial intelligence model. For example, the electronic device (10) can apply information regarding a user input selecting to enhance the detail of the person's eyes to an artificial intelligence model.

[0058] According to one embodiment, the electronic device (10) may apply information regarding an input selecting a direction for improving the image quality of a first person image received from a user to an artificial intelligence model. For example, the electronic device (10) may apply information regarding a user's input selecting to modify the first person image to resemble a predetermined image to an artificial intelligence model. Specifically, the electronic device (10) may apply information regarding at least one of the color tone, sharpness, and resolution of the image selected by the user to an artificial intelligence model. As another example, the electronic device (10) may apply information regarding a user's input selecting to modify a person included in the first person image to resemble a predetermined person to an artificial intelligence model.

[0059] According to one embodiment, the electronic device (10) can transmit the first person image to the server (20) in order to apply the first person image to an artificial intelligence model built on the server (20).

[0060] According to one embodiment, the artificial intelligence model may have learned a plurality of human images as training data to perform face identification and face recognition from human images and to perform image quality improvement for the recognized faces.

[0061] According to one embodiment, the artificial intelligence model may be trained in pairs of a high-quality person image and a low-quality person image to which degradation has been applied to the high-quality person image, so as to perform quality improvement on the low-quality person image.

[0062] According to one embodiment, the artificial intelligence model may use multiple person images classified by person as training data to perform personalized learning for each person. Additionally, the artificial intelligence model that has performed personalized learning may be lightweighted.

[0063] According to one embodiment, the artificial intelligence model may be trained to acquire facial features for each person by performing personalized learning. Additionally, the artificial intelligence model may be trained to identify and recognize a person from a person image based on the acquired facial features. Furthermore, the artificial intelligence model may be trained to improve image quality based on the acquired facial features. For example, the artificial intelligence model may be trained to improve image quality by performing image processing on an area on the image corresponding to an image quality improvement area selected by the user, based on the acquired facial features. As another example, the artificial intelligence model may be trained to improve image quality by modifying the acquired facial features according to the image quality improvement direction selected by the user.

[0064] Referring to step S250, the electronic device (10) can obtain a high-quality second person image output from an artificial intelligence model. The electronic device (10) can obtain the second person image obtained by the artificial intelligence model performing image quality improvement on the first person image.

[0065] According to one embodiment, the electronic device (10) can obtain a second person image generated by performing image quality improvement on a face recognized from a first person image. For example, the electronic device (10) can obtain a second person image generated by performing image processing to increase the resolution of the first person image. For another example, the electronic device (10) can obtain a second person image generated by performing image processing to remove noise from the first person image. For another example, the electronic device (10) can obtain a second person image generated by performing image processing to adjust the color of the first person image. For another example, the electronic device (10) can obtain a second person image generated by performing image processing to adjust the color of the first person image. For another example, the electronic device (10) can obtain a second person image generated by performing image processing to improve the clarity of the first person image.

[0066] According to one embodiment, the electronic device (10) can obtain a second person image from an artificial intelligence model in which image quality enhancement has been performed on an area corresponding to an image quality enhancement area selected by the user. For example, the electronic device (10) can apply information regarding user input selecting to enhance the detail of the person's eyes to the artificial intelligence model.

[0067] According to one embodiment, the electronic device (10) can obtain a second person image in which image quality improvement has been performed according to an image quality improvement direction selected by the user from an artificial intelligence model. For example, the electronic device (10) can obtain a second person image in which image quality improvement has been performed according to the input of a user who selects to modify a first person image to be similar to a predetermined image. Specifically, the electronic device (10) can obtain a second person image in which image quality improvement has been performed similarly to at least one of the color tone, sharpness, and resolution of the image selected by the user. As another example, the electronic device (10) can obtain a second person image in which a person included in the first person image has been modified to be similar to a predetermined person according to the input of the user.

[0068] According to one embodiment, the electronic device (10) can receive a second person image generated from an artificial intelligence model built on a server (20).

[0069] According to one embodiment, the electronic device (10) can display a second person image on the display unit (12-1). The electronic device (10) can display the first person image and the second person image together so that they can be compared.

[0070] According to one embodiment, the electronic device (10) may store a second person image in memory (17) in accordance with a user's confirmation input regarding the image quality improvement result. The electronic device (10) may store the second person image at the location where the first person image is stored. For example, the electronic device (10) may store the second person image together with the first person image. Alternatively, the electronic device (10) may store the first person image by replacing it with the second person image.

[0072] FIG. 3 is a diagram showing an electronic device applying human images as training data to an artificial intelligence model according to one embodiment.

[0073] Referring to FIG. 3, the electronic device (10) can apply a plurality of high-resolution images (310) and a plurality of low-resolution images (330) as training data to an artificial intelligence model. For example, the electronic device (10) can identify a plurality of images stored in the memory (17) of the electronic device (10) and apply them to the artificial intelligence model. As another example, the electronic device (10) can download a plurality of images disclosed on the web and apply them to the artificial intelligence model. As yet another example, the electronic device (10) can apply a plurality of images shared through an application to the artificial intelligence model.

[0074] According to one embodiment, the electronic device (10) can apply a plurality of degraded low-quality images (330) and a plurality of high-quality images (310) to an artificial intelligence model as training data.

[0075] For example, the electronic device (10) can apply the multiple low-quality images (330) and the multiple high-quality images (310) obtained by performing downsampling on each of the multiple high-quality images (310) to an artificial intelligence model.

[0076] As another example, the electronic device (10) can apply the multiple low-quality images (330) and the multiple high-quality images (310) obtained by performing image processing that applies noise to each of the multiple high-quality images (310) to an artificial intelligence model.

[0077] As another example, the electronic device (10) can apply the multiple low-quality images (330) and the multiple high-quality images (310) obtained by performing image processing that applies a blur to each of the multiple high-quality images (310) to an artificial intelligence model.

[0078] As another example, the electronic device (10) can apply the multiple low-quality images (330) and the multiple high-quality images (310) obtained by performing image processing to change the color of each of the multiple high-quality images (310) to an artificial intelligence model.

[0079] As another example, the electronic device (10) can apply the multiple low-quality images (330) and the multiple high-quality images (330) obtained by performing image processing to change the color of each of the multiple high-quality images (310) to an artificial intelligence model.

[0080] According to one embodiment, the electronic device (10) can group high-quality images and low-quality images into pairs and apply them to an artificial intelligence model. For example, the electronic device (10) can group a first high-quality image (311) and a first low-quality image (313) obtained by degrading the first high-quality image (311) among a plurality of high-quality images (310) into a pair, and apply the first high-quality image (311) and the first low-quality image (313) together to an artificial intelligence model.

[0081] According to one embodiment, the electronic device (10) may apply only a plurality of high-definition images (310) to an artificial intelligence model. The artificial intelligence model (19) may obtain a plurality of low-definition images (330) by performing degradation application on each of the plurality of high-definition images (310). The artificial intelligence model (19) may perform learning using the plurality of high-definition images (310) and the plurality of low-definition images (330) as training data.

[0082] According to one embodiment, the electronic device (10) can perform preprocessing on a plurality of high-resolution images (310) and a plurality of low-resolution images (330) and apply them to an artificial intelligence model. For example, the electronic device (10) can identify faces on a plurality of high-resolution images (310) and a plurality of low-resolution images (330), extract regions corresponding to the identified faces, and apply them to an artificial intelligence model as training data.

[0083] According to one embodiment, the artificial intelligence model (19) can learn the applied first high-definition image (311) and the first low-definition image (331).

[0084] For example, the artificial intelligence model (19) can train an image quality improvement artificial intelligence model to generate a first high-quality image (311) from a first low-quality image (331). Additionally, the artificial intelligence model (19) can train a discrimination artificial intelligence model to identify differences by comparing the second high-quality image generated from the first low-quality image (331) with the first high-quality image (311).

[0085] Specifically, the AI ​​model can be trained to improve image quality by applying the objective function (Loss Function) of Equation 1.

[0086]

[0087] is the objective function (Loss Function) of the image quality improvement AI model, and is the objective function (Loss Function) of a general GAN-based artificial intelligence model, and This is the objective function (Loss Function) for the image quality improvement AI model to generate a high-quality image as similar as possible from a low-quality image.

[0088] Additionally, the artificial intelligence model (19) can train a discrimination artificial intelligence model to distinguish between the first high-definition image (311) and the second high-definition image using an objective function (Loss Function).

[0089] Although the above describes a method for training an artificial intelligence model built on an electronic device (10), it can also be applied by analogy to a method for training an artificial intelligence model built on a server (20).

[0090] According to one embodiment, an electronic device (10) may receive information regarding an artificial intelligence model learned from a server (20). For example, the server (20) may train an artificial intelligence model built on the server (20) and transmit data regarding an artificial intelligence model (19) updated by the learning to the electronic device (10). In this case, the electronic device (10) may receive information regarding updated weights among the weights of the artificial intelligence model and update the artificial intelligence model (19) built on the electronic device (10) using the received information.

[0092] FIG. 4 is a diagram showing an electronic device applying human images as training data to an artificial intelligence model according to one embodiment.

[0093] Referring to FIG. 4, the electronic device (10) can generate a database containing various images of a specific person by classifying each of the multiple images of a person in a personalized manner so that it is included in one of the multiple images of a person. That is, the electronic device (10) can classify the multiple images of a person in a personalized manner such as a first image of a person (410), a second image of a person (430), and a third image of a person (450). The electronic device (10) can perform personalized learning of an artificial intelligence model by applying the classified first image of a person (410), second image of a person (430), and third image of a person (450) to an artificial intelligence model as training data.

[0094] According to one embodiment, the electronic device (10) can personalize and classify multiple person images based on input selected by the user as the same person.

[0095] According to one embodiment, the electronic device (10) can personalize and classify multiple person images based on a path where images are stored. For example, the electronic device (10) can personalize and classify images stored in the same folder of memory (17) into images related to the same person.

[0096] According to one embodiment, the electronic device (10) can personalize and classify a plurality of person images based on the user who provided the images. For example, the electronic device (10) can use a messenger application to personalize and classify images provided by the same user into images related to the same person.

[0097] According to one embodiment, the electronic device (10) can perform personalized classification of multiple person images using an artificial intelligence model. For example, the electronic device (10) can perform face detection from multiple person images using a face detection artificial intelligence model. The electronic device (10) can identify a first face by performing face recognition on the detected face using a face recognition artificial intelligence model. Based on the face recognition result, the electronic device (10) can identify a first person corresponding to the first face among multiple people.

[0098] According to one embodiment, the electronic device (10) can apply information regarding the results of personalized classification of a plurality of person images to an artificial intelligence model together with the plurality of person images. For example, the electronic device (10) can insert tags regarding the results of personalized classification into a plurality of high-resolution person images and a plurality of low-resolution person images. The electronic device (10) can apply the plurality of high-resolution person images and the plurality of low-resolution person images with inserted tags to an artificial intelligence model. In this case, the plurality of low-resolution person images may be generated by applying degradation to the plurality of high-resolution person images.

[0099] According to one embodiment, the electronic device (10) can apply a plurality of human images for which facial identification has not been performed to an artificial intelligence model. The artificial intelligence model (19) can train a facial identification artificial intelligence model and a facial recognition artificial intelligence model by using the plurality of human images as training data.

[0100] According to one embodiment, the artificial intelligence model (19) can update the image quality improvement artificial intelligence model by performing personalized learning of the image quality improvement artificial intelligence model using a plurality of input images of people. The updated image quality improvement artificial intelligence model may be a model specialized to improve the image quality of low-quality images of a specific person. For example, the artificial intelligence model can obtain an image quality improvement artificial intelligence model specialized to improve the image quality of low-quality images of the first person, the second person, and the third person by training the image quality improvement artificial intelligence model using the first person image (410), the second person image (430), and the third person image (450).

[0101] According to one embodiment, the artificial intelligence model (19) can perform lightweighting on at least one of a face detection artificial intelligence model that performs personalized learning, a face recognition artificial intelligence model, and an image quality improvement artificial intelligence model. For example, the artificial intelligence model (19) can apply a lightweighting technique, such as filter pruning, to the image quality improvement artificial intelligence model. The artificial intelligence model (19) can obtain an image quality improvement artificial intelligence model with excellent image quality improvement performance for a specific person that has been personalized learned and with lightweight data. As another example, the artificial intelligence model (19) can obtain a face recognition artificial intelligence model with excellent face recognition performance for a specific person that has been personalized learned and with lightweight data.

[0103] FIG. 5 is a diagram showing an electronic device acquiring facial features from a person image using an artificial intelligence model according to one embodiment.

[0104] Referring to FIG. 5, the electronic device (10) can obtain facial features of a person included in the person image (510) by applying the person image (510) to an artificial intelligence model (19).

[0105] According to one embodiment, the artificial intelligence model (19) can perform face detection from a person image (510). For example, the artificial intelligence model (19) can detect at least one face from the person image (510) using a face detection artificial intelligence model.

[0106] According to one embodiment, the artificial intelligence model (19) can obtain facial features from a detected face. For example, the artificial intelligence model (19) can use a facial recognition artificial intelligence model to obtain facial features such as the contour of the face detected from a person image, the shape, size, proportion and location of landmarks of the face (e.g., eyes, nose, mouth, ears, etc.), and details of the face (e.g., eyebrows, wrinkles, hair, skin tone).

[0107] According to one embodiment, the artificial intelligence model (19) can acquire facial features of a specific person by having the facial recognition artificial intelligence model perform personalized learning. For example, the artificial intelligence model (19) can acquire facial features of a first person by learning a person image of a first person.

[0108] According to one embodiment, the artificial intelligence model (19) can update the facial recognition artificial intelligence model using acquired facial features. For example, the artificial intelligence model (19) can update the weights of the facial recognition artificial intelligence model using facial features of a specific person.

[0109] According to one embodiment, the artificial intelligence model (19) can update the image quality improvement artificial intelligence model using acquired facial features. For example, the artificial intelligence model (19) can train the image quality improvement artificial intelligence model to modify facial features such as facial contours, the shape, size, and location of facial landmarks (e.g., eyes, nose, mouth, ears, etc.), and facial details (eyebrows, wrinkles, hair).

[0111] FIG. 6 is a diagram showing that an electronic device performs image quality improvement on an image quality improvement area received from a user according to one embodiment, and FIG. 7 is a flowchart of a method for improving image quality on an image quality improvement area received from a user by an electronic device according to one embodiment.

[0112] Referring to FIG. 6, the electronic device (10) can receive input from a user for selecting an image quality improvement area for a person included in the first person image (610). Based on the input received from the user, the electronic device (10) can output a second person image (630) with improved image quality of the first person image (610).

[0113] Referring to step S710, the electronic device (10) receives input from the user (1) to select an image enhancement area and can identify the user's selected image enhancement area (Region of image enhancement).

[0114] For example, the electronic device (10) may receive input from the user (1) selecting a part of the face of a person included in the first person image (610) that requires image quality improvement. For example, with reference to FIG. 6, the electronic device (10) may receive input from the user (1) selecting the eyes of a person included in the first person image (610).

[0115] As another example, the electronic device (10) can receive input from a user selecting a part of the face of a person included in the first person image (610) that requires image quality improvement, provided as a list (preset).

[0116] FIG. 6 is illustrated such that an electronic device (10) receives input from a user selecting an image quality enhancement area based on the touch input of a user (1), but is not limited thereto. The electronic device (10) can identify an image quality enhancement area based on input from a user received through various interfaces capable of receiving input from a user.

[0117] According to one embodiment, the electronic device (10) can detect the face of the first person and features of the first person's face from the first person image (610) using a face detection artificial intelligence model. The electronic device (10) can identify an area corresponding to an image quality enhancement area selected by the user (1) using the face detection artificial intelligence model. For example, the electronic device (10) can detect features of the face using the face detection artificial intelligence model and, by performing face parsing based on the detected features of the face, identify that the area where the user input is received corresponds to the eyes of the first person.

[0118] According to one embodiment, the electronic device (10) can identify a first person corresponding to a face detected from a first person image (610) using a facial recognition artificial intelligence model.

[0119] Referring to step S730, the electronic device (10) can apply information about the image quality improvement area to the artificial intelligence model (19).

[0120] According to one embodiment, the electronic device (10) may apply information regarding an image quality improvement area to the artificial intelligence model (19) together with the first person image (610). For example, the electronic device (10) may apply the first person image (610), which has the area where user input was received marked, to the artificial intelligence model (19). For another example, the electronic device (10) may apply feature information regarding the area where user input was received to the artificial intelligence model (19) together with the first face image (610). For another example, the electronic device (10) may apply information regarding a facial area corresponding to the area where user input was received to the artificial intelligence model (19) together with the first face image (610).

[0121] According to one embodiment, the electronic device (10) can apply information regarding a first person identified from a first person image (610) using a facial recognition artificial intelligence model to the artificial intelligence model (19) together with the first person image (610). The electronic device (10) can apply a plurality of high-definition images containing the first person to the artificial intelligence model (19) together with the first person image (610).

[0122] Referring to step S750, the electronic device (10) can train an artificial intelligence model using training data for an image quality improvement area.

[0123] According to one embodiment, the artificial intelligence model (19) can identify an image quality improvement area from the first person image (610) based on information regarding an input image quality improvement area. For example, the artificial intelligence model (19) can identify an image quality improvement area from the first person image (610) based on information regarding a facial area (e.g., eyes) selected by the user (1).

[0124] According to one embodiment, the artificial intelligence model (19) can acquire training data regarding the image quality improvement area based on information regarding the input image quality improvement area. For example, the artificial intelligence model (19) can acquire a plurality of high-quality images of a facial area (e.g., eyes) selected by the user based on information regarding a facial area (e.g., eyes) selected by the user. The artificial intelligence model (19) can output data requesting the application of a plurality of high-quality images to the electronic device (10).

[0125] According to one embodiment, the artificial intelligence model (19) can train the artificial intelligence model to perform face parsing by learning the learning data input from the electronic device (10).

[0126] According to one embodiment, the artificial intelligence model (19) can set an objective function (Loss Function) of the image quality improvement artificial intelligence model based on face parsing result data output from the face detection artificial intelligence model and information regarding an image quality improvement area selected by the user. For example, the artificial intelligence model (19) can set an objective function (Loss Function) of the weighted loss method for the image quality improvement artificial intelligence model.

[0127] According to one embodiment, the artificial intelligence model (19) can train an image quality improvement artificial intelligence model using training data input from the electronic device (10). The artificial intelligence model (19) can train an image quality improvement artificial intelligence model using an objective function (Loss Function) set for the image quality improvement artificial intelligence model and training data input from the electronic device (10). For example, the artificial intelligence model (19) can train an image quality improvement artificial intelligence model using a plurality of high-definition images of a facial area (e.g., eyes) selected by the user. As another example, the artificial intelligence model (19) can train an image quality improvement artificial intelligence model that improves the image quality of a facial area (e.g., eyes) of a first person selected by the user using a plurality of high-definition images of a first person included in a first person image (610).

[0128] Referring to step S770, the electronic device (10) can obtain a second person image with improved image quality in the image quality improvement area.

[0129] According to one embodiment, the artificial intelligence model (19) can generate a second person image (630) by performing image quality improvement on the first person image (610) using the image quality improvement artificial intelligence model updated in step S750. For example, the artificial intelligence model (19) can generate a second person image (630) by performing image quality improvement on a predetermined facial area (e.g., eyes) of the first person included in the first person image (610). The artificial intelligence model (19) can output the generated second person image (630) to an electronic device.

[0130] According to one embodiment, the electronic device (10) can display a second person image on the display unit (12-1). The electronic device (10) can display the first person image and the second person image together so that they can be compared.

[0131] According to one embodiment, the electronic device (10) may store a second person image (630) output from an artificial intelligence model (19) in memory (17). For example, the electronic device (10) may store the second person image in a path where the first person image is stored. For example, the electronic device (10) may store the second person image together with the first person image. Alternatively, the electronic device (10) may store the first person image by replacing it with the second person image.

[0133] FIG. 8 is a diagram showing that an electronic device improves image quality according to an image quality improvement direction received from a user according to one embodiment, and FIG. 9 is a flowchart of a method for improving image quality according to an image quality improvement direction received from a user according to one embodiment.

[0134] Referring to FIG. 8, the electronic device (10) can obtain feature information from the third person image (820) based on the input of a user (1) who selects the third person image (820) in the direction of improving the image quality of the first person image (810). The electronic device (10) can obtain the second person image (830) by improving the image quality of the first person image (810) using the feature information obtained from the third person image (820).

[0135] For example, the electronic device (10) can obtain a second person image (830) by using an artificial intelligence model (19) to perform image processing that increases the resolution of the first person image (810) to correspond to the resolution of the third person image (820).

[0136] As another example, the electronic device (10) can obtain a second person image (830) by using an artificial intelligence model (19) to perform image processing that adjusts the color of the first person image (810) to correspond to the color of the third person image (820).

[0137] As another example, the electronic device (10) can obtain a second person image (830) by using an artificial intelligence model (19) to perform image processing that adjusts the clarity of the first person image (810) to correspond to the clarity of the third person image (820).

[0138] As another example, the electronic device (10) can obtain a second person image (830) by using an artificial intelligence model (19) to perform image processing that adjusts the facial features of the first person included in the first person image (810) to match the facial features of the second person included in the third person image (820).

[0139] Referring to step S910, the electronic device (10) can receive input from the user regarding image quality improvement "Nuwa*".

[0140] According to one embodiment, the electronic device (10) may receive input from a user regarding a direction for improving image quality, which adjusts at least one of the resolution, clarity, color, noise removal, and removal of artifacts occurring during image compression of a first person image (810).

[0141] For example, the electronic device (10) can receive input from a user regarding the direction of image quality improvement through an interface that selects at least one of the resolution, clarity, color, noise removal, and removal of artifacts occurring during image compression of the first person image (810).

[0142] As another example, the electronic device (10) may receive input from a user selecting a third person image (820) such that the first person image (810) includes image attributes similar to those of the third person image (820). That is, the electronic device (10) may receive input from a user selecting a direction for image quality improvement such that the first person image (810) corresponds to at least one of the resolution, sharpness, color, noise, and artifacts of the third person image (820).

[0143] According to one embodiment, the electronic device (10) may receive input from a user regarding a direction for improving image quality that modifies the facial features of the first person included in the first person image (810). Specifically, the electronic device (10) may receive input regarding a direction for improving image quality that adjusts the facial features of the first person included in the first person image (810) to match the facial features of the second person included in the third person image (820).

[0144] Referring to step S930, the electronic device (10) can apply information regarding the direction of image quality improvement to an artificial intelligence model.

[0145] According to one embodiment, the electronic device (10) may apply information regarding the direction of image quality improvement to the artificial intelligence model (19) together with the first person image (810). For example, the electronic device (10) may apply information regarding at least one of the resolution, sharpness, color, noise removal, and removal of artifacts occurring in the compressed image of the first person image (810) selected by the user to the artificial intelligence model (19) together with the first person image (810). As another example, the electronic device (10) may apply the third person image (820) selected by the user to the artificial intelligence model (19) together with the first person image (810).

[0146] According to one embodiment, the electronic device (10) may apply a plurality of high-definition data related to a second person included in a third person image (820) together with a first person image (810) to an artificial intelligence model (19) as training data. For example, the electronic device (10) may detect a face from the third person image (820) and identify a second person corresponding to the detected face. The electronic device (10) may obtain a plurality of high-definition images containing the second person. The electronic device (10) may generate a plurality of low-definition images by performing degradation on each of the plurality of high-definition images containing the second person, and may apply the generated plurality of low-definition images together with the plurality of high-definition images as training data to an artificial intelligence model.

[0147] Referring to step S950, the electronic device (10) can train an artificial intelligence model using training data regarding the direction of image quality improvement.

[0148] According to one embodiment, the artificial intelligence model (19) can obtain feature vectors of the third person image (820) from the third person image (820). The artificial intelligence model (19) can train an image quality improvement artificial intelligence model using the feature vectors of the third person image (820). For example, the artificial intelligence model (19) can obtain feature vectors regarding at least one of the resolution, sharpness, color, noise, and artifacts of the third person image (820), and train an image quality improvement artificial intelligence model to improve the image quality of the first person image (810) according to the image quality improvement direction selected by the user using the obtained feature vectors.

[0149] According to one embodiment, the artificial intelligence model (19) can obtain facial features of the second person from the third person image (820). For example, the artificial intelligence model (19) can use a facial recognition artificial intelligence model to obtain facial features such as the contour of the second person's face, the shape, size, and location of facial landmarks (e.g., eyes, nose, mouth, ears, etc.), and facial details (eyebrows, wrinkles, hair, skin tone) from the third person image (820).

[0150] According to one embodiment, the artificial intelligence model (19) can obtain facial features of the second person from a plurality of images applied together with the third person image (820). For example, the artificial intelligence model (19) can detect the face of the second person from each of a plurality of high-definition images containing the second person and obtain facial features of the second person. Additionally, the artificial intelligence model (19) can detect the face of the second person from each of a plurality of low-definition images generated by applying degradation to a plurality of high-definition images containing the second person and obtain facial features of the second person.

[0151] According to one embodiment, the artificial intelligence model (19) can train an image quality improvement artificial intelligence model using the facial features of the second person obtained. For example, the artificial intelligence model (19) can set the objective function (Loss Function) of the image quality improvement artificial intelligence model such that the facial features of the second person image (830), generated by the image quality improvement artificial intelligence model using the facial features obtained from the first person image (810), are similar to the facial features obtained from the third person image (820). The artificial intelligence model (19) can train the image quality improvement artificial intelligence model using the objective function (Loss Function) set for the image quality improvement artificial intelligence model and training data input from the electronic device (10).

[0152] Referring to step S970, the electronic device (10) can acquire a second person image according to the direction of image quality improvement.

[0153] According to one embodiment, the artificial intelligence model (19) can obtain a second person image (830) by performing at least one image processing among resolution enhancement, sharpness enhancement, color adjustment, noise removal, and artifact removal from a compressed image using an image quality enhancement artificial intelligence model to correspond to the image quality enhancement direction selected by the user. For example, the artificial intelligence model (19) can perform image processing on the first person image (810) so that the first person image (810) corresponds to at least one of the resolution, sharpness, color, noise, and artifact of the third person image (820).

[0154] According to one embodiment, the artificial intelligence model (19) can obtain a second person image (830) by using an image quality improvement artificial intelligence model to adjust the facial features of the first person included in the first person image (810) to match the facial features of the second person included in the third person image (820). For example, the artificial intelligence model (19) can adjust the shape, size, and ratio of the landmarks of the first person's face to match the shape, size, and ratio of the landmarks of the second person's face (e.g., eyes, nose, mouth, ears, etc.). As another example, the artificial intelligence model (19) can adjust the details of the first person's face to match the details of the second person's face (e.g., eyebrows, wrinkles, hair, skin tone, etc.).

[0156] FIG. 10 is a block diagram of an electronic device according to one embodiment.

[0157] Referring to FIG. 10, the electronic device (10) may include a user input unit (11), an output unit (12), a processor (13), a communication unit (15), and a memory (17). However, not all components shown in FIG. 10 are essential components of the electronic device (10). The electronic device (10) may be implemented by more components than those shown in FIG. 10, or by fewer components than those shown in FIG. 10.

[0158] The user input unit (11) refers to a means for a user to input data for controlling an electronic device (10). For example, the user input unit (11) may include a touch screen, a key pad, a dome switch, a touch pad (contact capacitive type, pressure resistive type, infrared detection type, surface ultrasonic conduction type, integral tension measurement type, piezo effect type, etc.), a touch screen, a jog wheel, a jog switch, etc., but is not limited thereto.

[0159] The user input unit (11) can receive user input necessary for the electronic device (10) to perform the embodiments described with reference to FIGS. 1 to 9.

[0160] The output unit (12) outputs information processed by the electronic device (10). The output unit (12) may output information related to the embodiments described with reference to FIGS. 1 to 9. Additionally, the output unit (12) may include a display unit (12-1) that displays the result of performing an action corresponding to an object, a user interface, or user input.

[0161] The processor (13) typically controls the overall operation of the electronic device (10). For example, the processor (13) can control the user input unit (11), output unit (12), communication unit (15), memory (17), etc., in order to perform federated learning by executing at least one instruction stored in memory (17).

[0162] For example, the processor (13) can control the electronic device (10) to detect a face from an image and identify a person corresponding to the detected face by executing instructions stored in the face identification and face recognition module (17a). Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0163] As another example, the processor (13) can control the electronic device (10) to acquire facial features detected from an image by executing instructions stored in the facial feature acquisition module (17b). Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0164] As another example, the processor (13) can control the electronic device (10) to improve the image quality by executing instructions stored in the image quality improvement module (17c). Content that overlaps with the embodiments described above with reference to FIGS. 1 to 9 is omitted.

[0165] As another example, the processor (13) can control the electronic device (10) so that the artificial intelligence model learns the training data by executing instructions stored in the artificial intelligence model learning module (17d). Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0166] The processor (13) may be at least one general-purpose processor. Additionally, the processor (13) may include at least one processor manufactured to perform the functions of an artificial intelligence model. The processor (13) may execute a series of instructions to enable the artificial intelligence model to learn new training data. The processor (13) may perform the functions of the artificial intelligence model described above with reference to FIGS. 1 through 9 by executing a software module stored in memory (17).

[0167] The communication unit (15) may include one or more components that enable the electronic device (10) to communicate with another device (not shown) and a server (20). The other device (not shown) may be a computing device such as the electronic device (10), but is not limited thereto.

[0168] The memory (17) can store at least one instruction and at least one program for processing and controlling the processor (13), and can also store data that is input to or output from the electronic device (10).

[0169] The memory (17) may include at least one type of storage medium among memory that temporarily stores data, such as RAM (Random Access Memory) and SRAM (Static Random Access Memory), flash memory type, hard disk type, multimedia card micro type, card type memory (e.g., SD or XD memory, etc.), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, magnetic disk, optical disk, etc.

[0171] FIG. 11 is a block diagram of a software module stored in the memory of an electronic device according to one embodiment.

[0172] Referring to FIG. 11, the memory (17) is a software module containing instructions for the electronic device (10) to perform the embodiment described above with reference to FIG. 1 to 9, and may include a face identification and face recognition module (17a), a face feature acquisition module (17b), an image quality improvement module (17c), and an artificial intelligence model learning module (17d). However, the electronic device (10) may perform image quality improvement using more software modules than those shown in FIG. 11, and the electronic device (10) may perform image quality improvement using fewer software modules than those shown in FIG. 10.

[0173] For example, by the processor (13) executing instructions stored in the face identification and face recognition module (17a), the electronic device (10) can detect a face from an image and identify a person corresponding to the detected face. Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0174] As another example, by having the processor (13) execute instructions stored in the facial feature acquisition module (17b), the electronic device (10) can acquire facial features detected from an image. Content that overlaps with the embodiments described above with reference to FIGS. 1 to 9 is omitted.

[0175] As another example, the electronic device (10) can improve the image quality by having the processor (13) execute instructions stored in the image quality improvement module (17c). Content that overlaps with the embodiments described above with reference to FIGS. 1 to 9 is omitted.

[0176] As another example, the electronic device (10) can train an artificial intelligence model by having the processor (13) execute instructions stored in the artificial intelligence model training module (17d). Content that overlaps with the embodiments described above with reference to FIGS. 1 to 9 is omitted.

[0178] FIG. 12 is a block diagram of a server according to one embodiment.

[0179] The server (20) can perform at least one operation of the electronic device (10). Additionally, the server (20) can perform at least one operation of the artificial intelligence model (19) described above.

[0180] Referring to FIG. 12, a server (20) according to some embodiments may include a communication unit (25), memory (26), DB (27) and a processor (23).

[0181] The communication unit (25) may include one or more components that enable the server (20) to communicate with the electronic device (10).

[0182] The memory (26) can store at least one instruction and at least one program for processing and controlling the processor (23), and can also store data that is input to or output from the server (20).

[0183] DB (27) can store data received from the electronic device (10). DB (27) can store multiple training data sets to be used to train an artificial intelligence model.

[0184] The processor (23) typically controls the overall operation of the server (20). For example, the processor (23) can control the DB (27) and the communication unit (25), etc., by executing programs stored in the memory (26) of the server (20). By executing programs, the processor (23) can perform at least one of the operations of the electronic device (10) described with reference to FIGS. 1 to 9 and the operation of the server (20).

[0185] For example, the processor (23) can control the server (20) to detect a face from an image and identify a person corresponding to the detected face by executing instructions stored in the face identification and face recognition module (17a). Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0186] As another example, the processor (23) can control the server (20) to acquire facial features detected from an image by executing instructions stored in the facial feature acquisition module (17b). Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0187] As another example, the processor (23) can control the server (20) to improve the image quality by executing instructions stored in the image quality improvement module (17c). Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0188] As another example, the processor (23) can control the server (20) so that the artificial intelligence model learns the training data by executing instructions stored in the artificial intelligence model learning module (17d). Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0190] FIG. 13 is a block diagram of a software module stored in the memory of a server according to one embodiment.

[0191] Referring to FIG. 13, the memory (26) is a software module for the server (20) to perform the embodiments described above with reference to FIGS. 1 to 9, and may include a face identification and face recognition module (17a), a face feature acquisition module (17b), an image quality improvement module (17c), and an artificial intelligence model learning module (17d).

[0192] However, the server (20) can perform image quality improvement using more software modules than the software module shown in FIG. 13, and the server (20) can perform image quality improvement using fewer software modules than the software module shown in FIG. 12.

[0193] For example, by the processor (13) executing instructions stored in the face identification and face recognition module (17a), the electronic device (10) can detect a face from an image and identify a person corresponding to the detected face. Content that overlaps with the embodiment described above with reference to FIGS. 1 to 9 is omitted.

[0194] As another example, by having the processor (13) execute instructions stored in the facial feature acquisition module (17b), the electronic device (10) can acquire facial features detected from an image. Content that overlaps with the embodiments described above with reference to FIGS. 1 to 9 is omitted.

[0195] As another example, the electronic device (10) can improve the image quality by having the processor (13) execute instructions stored in the image quality improvement module (17c). Content that overlaps with the embodiments described above with reference to FIGS. 1 to 9 is omitted.

[0196] As another example, the electronic device (10) can train an artificial intelligence model by having the processor (13) execute instructions stored in the artificial intelligence model training module (17d). Content that overlaps with the embodiments described above with reference to FIGS. 1 to 9 is omitted.

[0198] Some embodiments may also be implemented in the form of a recording medium containing computer-executable instructions, such as program modules executed by a computer. A computer-readable medium may be any available medium accessible by a computer and includes both volatile and non-volatile media, and both removable and non-removable media. Additionally, a computer-readable medium may include a computer storage medium. A computer storage medium includes both volatile and non-volatile, removable and non-removable media implemented by any method or technique for storing information, such as computer-readable instructions, data structures, program modules, or other data.

Claims

Claim 1 A method for generating a second high-quality person image by performing image processing on a first low-quality person image using an artificial intelligence model, wherein a discrimination artificial intelligence model trained to distinguish between a high-quality person image and a high-quality person image generated from a low-quality person image by an image quality improvement artificial intelligence model is used to select a low-quality image among a plurality of images that is subject to image quality improvement. A method comprising: a step of identifying a first person image; a step of applying the first person image to an artificial intelligence model; and a step of obtaining a second person image output from the artificial intelligence model, wherein the artificial intelligence model identifies a first face in the first person image using a face recognition artificial intelligence model, and obtains a second person image by performing image processing to improve the image quality of an area corresponding to the first face using an image quality improvement artificial intelligence model, and outputs the second person image to the electronic device, and wherein the image quality improvement artificial intelligence model is trained based on the output result of the discrimination artificial intelligence model trained to distinguish a high-quality person image generated from a high-quality person image and a low-quality person image. Claim 2 A method according to claim 1, wherein the artificial intelligence model updates the image quality improvement artificial intelligence model by learning the plurality of low-quality images, each of which is converted from the plurality of high-quality images, and the plurality of high-quality images as learning data, and obtains the second image from the first image using the updated image quality improvement artificial intelligence model. Claim 3 A method according to claim 2, wherein the artificial intelligence model updates the image quality improvement artificial intelligence model to generate the first high-quality person image from the first low-quality person image by learning a pair of a first high-quality person image and a first low-quality person image generated by applying image degradation to the first high-quality person image, and obtains the second person image from the first person image using the updated image quality improvement artificial intelligence model. Claim 4 A method according to claim 1, wherein the artificial intelligence model updates a face recognition artificial intelligence model and an image quality improvement artificial intelligence model by performing personalized learning using a plurality of person images classified by person, identifies the first face and the first person corresponding to the first face from the first person image using the updated face recognition artificial intelligence model, and obtains the second person image from the first person image using the image quality improvement artificial intelligence model updated for the first person. Claim 5 A method according to claim 4, wherein the artificial intelligence model performs lightweighting on the face recognition artificial intelligence model and the image quality improvement artificial intelligence model updated by the personalized learning, identifies the first person from the first person image using the lightweighted face recognition artificial intelligence model, and obtains the second person image with improved image quality for the first person from the first person image using the lightweighted image quality improvement artificial intelligence model. Claim 6 In claim 4, the artificial intelligence model acquires facial features of the first person by learning a plurality of person images of the first person as training data, identifies the first person from the first person image based on the facial features of the first person, and acquires the second person image with improved image quality of the first person from the first person image based on the facial features of the first person. Claim 7 In claim 6, the method further comprises the step of receiving input from a user selecting an image quality improvement area of ​​the first person; and the step of applying information regarding the image quality improvement area selected by the user to the artificial intelligence model, wherein the artificial intelligence model updates the image quality improvement artificial intelligence model by additionally learning training data regarding the image quality improvement area of ​​the first person, and using the updated image quality improvement artificial intelligence model, obtains the second person image in which the image quality of the area corresponding to the image quality improvement area selected by the user among the areas corresponding to the first face of the first person is improved. Claim 8 In claim 6, the method further comprises the step of receiving input from a user selecting a direction for image quality improvement for the first person; and the step of applying information regarding the image quality improvement direction selected by the user to the artificial intelligence model, wherein the artificial intelligence model updates the image quality improvement artificial intelligence model by additionally learning training data regarding the image quality improvement direction, and obtains the second person image by modifying the facial features of the first person according to the image quality improvement direction using the updated image quality improvement artificial intelligence model. Claim 9 In claim 8, the method comprises the steps of: receiving user input designating a second person; obtaining data for the second person; applying the data for the second person to an artificial intelligence model as training data; wherein the artificial intelligence model obtains facial features of the second person from the data for the second person, and obtains an image of the second person by modifying the facial features of the first person based on the facial features of the second person. Claim 10 In a recording medium storing at least one instruction readable by an electronic device that generates a second high-quality person image by performing image processing on a first low-quality person image using an artificial intelligence model, the recording medium comprises, by the electronic device executing the at least one instruction, a discrimination artificial intelligence model learned to distinguish between a high-quality person image and a high-quality person image generated from a low-quality person image by an image quality improvement artificial intelligence model, wherein among a plurality of images, a low-quality image that is subject to image quality improvement A recording medium that identifies the first person image, applies the first person image to an artificial intelligence model, obtains the second person image output from the artificial intelligence model, wherein the artificial intelligence model identifies a first face in the first person image using a face recognition artificial intelligence model, and obtains the second person image by performing image processing to improve the image quality for an area corresponding to the first face using an image quality improvement artificial intelligence model, outputs the second person image to the electronic device, and wherein the image quality improvement artificial intelligence model is trained based on the output result of the discrimination artificial intelligence model trained to distinguish between a high-quality person image and a high-quality person image generated from a low-quality person image. Claim 11 An electronic device for generating a second high-quality person image by performing image processing on a first low-quality person image using an artificial intelligence model, comprising: a memory storing at least one instruction; and a processor for executing the at least one instruction, wherein the processor, by executing the at least one instruction, uses a discrimination artificial intelligence model trained to distinguish between a high-quality person image and a high-quality person image generated from a low-quality person image by an image quality improvement artificial intelligence model, among a plurality of images, a low-quality image that is subject to image quality improvement An electronic device that identifies the first person image, applies the first person image to an artificial intelligence model, obtains the second person image output from the artificial intelligence model, wherein the artificial intelligence model uses a face recognition artificial intelligence model to identify a first face in the first person image, and uses an image quality improvement artificial intelligence model to perform image processing to improve the image quality of an area corresponding to the first face, thereby obtaining the second person image, outputs the second person image to the electronic device, and wherein the image quality improvement artificial intelligence model is trained based on the output result of the discrimination artificial intelligence model trained to distinguish between a high-quality person image and a high-quality person image generated from a low-quality person image. Claim 12 An electronic device according to claim 11, wherein the artificial intelligence model updates the image quality improvement artificial intelligence model by learning a plurality of low-quality image, each of which is converted from a plurality of high-quality image, and the plurality of high-quality image as learning data, and obtains the second image from the first image using the updated image quality improvement artificial intelligence model. Claim 13 An electronic device according to claim 12, wherein the artificial intelligence model updates the image quality improvement artificial intelligence model to generate the first high-quality person image from the first low-quality person image by learning a pair of a first high-quality person image and a first low-quality person image generated by applying image degradation to the first high-quality person image, and obtains the second person image from the first person image using the updated image quality improvement artificial intelligence model. Claim 14 An electronic device according to claim 11, wherein the artificial intelligence model updates a face recognition artificial intelligence model and an image quality improvement artificial intelligence model by performing personalized learning using a plurality of person images classified by person, identifies the first face and the first person corresponding to the first face from the first person image using the updated face recognition artificial intelligence model, and obtains the second person image from the first person image using the image quality improvement artificial intelligence model updated for the first person. Claim 15 An electronic device according to claim 14, wherein the artificial intelligence model performs lightweighting on the facial recognition artificial intelligence model and the image quality improvement artificial intelligence model updated by the personalized learning, identifies the first person from the first person image using the lightweighted facial recognition artificial intelligence model, and obtains the second person image with improved image quality for the first person from the first person image using the lightweighted image quality improvement artificial intelligence model. Claim 16 An electronic device according to claim 14, wherein the artificial intelligence model acquires facial features of the first person by learning a plurality of person images of the first person as learning data, identifies the first person from the first person image based on the facial features of the first person, and acquires the second person image with improved image quality of the first person from the first person image based on the facial features of the first person. Claim 17 In claim 16, the electronic device further includes an input interface for receiving input from a user for selecting an image quality improvement area of ​​the first person, and the processor applies information regarding the image quality improvement area selected by the user to the artificial intelligence model, and the artificial intelligence model updates the image quality improvement artificial intelligence model by additionally learning learning data regarding the image quality improvement area of ​​the first person, and using the updated image quality improvement artificial intelligence model, obtains the second person image in which the image quality of the area corresponding to the image quality improvement area selected by the user is improved among the areas corresponding to the first face of the first person. Claim 18 In claim 16, the electronic device further comprises an input interface for receiving input from a user selecting a direction for image quality improvement for the first person, and the processor applies information regarding the direction for image quality improvement selected by the user to the artificial intelligence model, and the artificial intelligence model updates the image quality improvement artificial intelligence model by additionally learning learning data regarding the direction for image quality improvement, and acquires the second person image by modifying the facial features of the first person according to the direction for image quality improvement using the updated image quality improvement artificial intelligence model. Claim 19 An electronic device according to claim 18, wherein the input interface receives user input specifying a second person, the processor acquires data regarding the second person and applies the data regarding the second person as training data to an artificial intelligence model, the artificial intelligence model acquires facial features of the second person from the data regarding the second person, and acquires an image of the second person by modifying the facial features of the first person based on the facial features of the second person.

Images