KR20240025440A - Electronic apparatus for displaying 3d image and operating method for the same - Google Patents

Abstract

The present disclosure includes a display, a memory storing at least one instruction, and at least one processor executing the at least one instruction stored in the memory, wherein the at least one processor acquires an input document including components; , recognizes components from the input document, classifies the type of input document based on the type of recognized component, and organizes the recognized components into one row according to the identified order based on the type of the classified input document. It includes an electronic device that sorts and obtains a corrected document, and displays the obtained corrected document on a display, and a method of operating the electronic device.

Description

Electronic device and operating method of electronic device for displaying 3D image {ELECTRONIC APPARATUS FOR DISPLAYING 3D IMAGE AND OPERATING METHOD FOR THE SAME}

This disclosure relates to an electronic device for displaying a 3D image and a method of operating the electronic device.

Thanks to the development of electronic technology, various types of electronic devices are being developed and distributed. Electronic devices, including display devices that display images, have been developing rapidly in recent years.

As display devices develop, the types of images displayed by the display devices also become more diverse. Display devices that can display not only two-dimensional (2D) images but also three-dimensional (3D) images are being developed.

Recently, in order to display 3D images, devices and methods for displaying 3D images using a volumetric display that can display objects in 3D space have been proposed. In particular, a stacked display has been proposed that includes a plurality of display panels stacked and displays an image on each of the plurality of display panels to provide a 3D image.

A stacked display displays a 3D image by displaying a plurality of images obtained based on light field images taken from different views on a plurality of stacked display panels.

The electronic device 100 according to an embodiment of the present disclosure includes a base panel 211, a layer panel 212 disposed on the base panel 211, a memory 230 that stores at least one instruction, and It may include at least one processor 220 that executes at least one instruction stored in the memory 230. At least one processor 220 may acquire a first frame image in the first frame of the input content and a second frame image in the second frame, which is the frame immediately before the first frame, by executing at least one command. . At least one processor 220 applies the first frame image and the second frame image to the image generation module, thereby creating a base image corresponding to the base panel in the first frame and a layer panel in the first frame, and A layer image can be created based on the movement between the image and the second frame image. At least one processor 220 may generate a correction-base image based on the movement between the first frame image and the second frame image by applying the base image or layer image to the image correction module. At least one processor 220 may display the correction-base image on the base panel and display the layer image on the layer panel.

Another embodiment of the present disclosure provides a method of operating an electronic device 100 including a base panel 211 and a layer panel 212 disposed on the base panel 211. The method of operating the electronic device 100 may include acquiring a first frame image in a first frame of input content and a second frame image in a second frame that is the frame immediately preceding the first frame. The operating method of the electronic device 100 is to generate a base image corresponding to the base panel 211 in the first frame and a layer panel 212 in the first frame by applying the first frame image and the second frame image to the image generation module. It may include generating a corresponding layer image based on movement between the first frame image and the second frame image. The method of operating the electronic device 100 may include generating a correction-base image based on the movement between the first frame image and the second frame image by applying the base image or the layer image to an image correction module. there is. The method of operating the electronic device 100 may include displaying a correction-base image on the base panel 211 and displaying a layer image on the layer panel 212.

In one embodiment of the present disclosure, a computer-readable recording medium on which a program for performing at least one method among the disclosed operating method embodiments is recorded on a computer may be provided.

The present disclosure may be understood in combination with the following detailed description and accompanying drawings, where reference numerals refer to structural elements.
1 is a diagram for explaining an electronic device according to an embodiment of the present disclosure.
FIG. 2 is a diagram for explaining the operation of an electronic device according to an embodiment of the present disclosure.
Figure 3 is a block diagram for explaining the configuration of an electronic device according to an embodiment of the present disclosure.
FIG. 4 is a flowchart for explaining the operation of an electronic device according to an embodiment of the present disclosure.
FIG. 5 is a flowchart for explaining the operation of an electronic device including an optical layer according to an embodiment of the present disclosure.
FIG. 6 is a diagram for explaining the operation of an electronic device that generates a layer image and a correction-base image according to an embodiment of the present disclosure.
FIG. 7 is a diagram for explaining movement between a first frame image and a second frame image according to an embodiment of the present disclosure.
FIG. 8 is a diagram for explaining the operation of an electronic device that generates a correction-base image including a layer image and a plurality of sub-correction-base images according to an embodiment of the present disclosure.
FIG. 9 is a diagram for explaining the operation of an electronic device that generates a correction-base image through an image correction module including a second artificial intelligence model, according to an embodiment of the present disclosure.
FIG. 10 is a diagram for explaining shift values according to an embodiment of the present disclosure.
FIG. 11 is a diagram illustrating a training process of a first artificial intelligence model included in an image generation module and a second artificial intelligence model included in an image correction module, according to an embodiment of the present disclosure.
FIG. 12A is a diagram illustrating a first loss function for training a first artificial intelligence model according to an embodiment of the present disclosure.
FIG. 12B is a diagram illustrating a second loss function for training a first artificial intelligence model according to an embodiment of the present disclosure.
FIG. 12C is a diagram illustrating a third loss function for training a second artificial intelligence model according to an embodiment of the present disclosure.

Terms used in the present disclosure will be briefly described, and an embodiment of the present disclosure will be described in detail.

The terms used in the present disclosure have selected general terms that are currently widely used as much as possible while considering the function in an embodiment of the present disclosure, but this may vary depending on the intention or precedent of a person working in the art, the emergence of new technology, etc. there is. In addition, in certain cases, there are terms arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the description of the corresponding embodiment of the present disclosure. Therefore, the terms used in this disclosure should be defined based on the meaning of the term and the overall content of this disclosure, rather than simply the name of the term.

Singular expressions may include plural expressions, unless the context clearly dictates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as generally understood by a person of ordinary skill in the technical field described herein.

Throughout the present disclosure, when a part “includes” a certain element, this means that it may further include other elements rather than excluding other elements, unless specifically stated to the contrary. In addition, terms such as "... unit" and "module" described in the present disclosure refer to a unit that processes at least one function or operation, which may be implemented as hardware or software, or as a combination of hardware and software. there is.

The expression “configured to” used in the present disclosure means, for example, “suitable for,” “having the capacity to,” depending on the situation. ”, “designed to”, “adapted to”, “made to”, or “capable of”. The term “configured (or set to)” may not necessarily mean “specifically designed to” in hardware. Instead, in some situations, the expression “system configured to” may mean that the system is “able to” work with other devices or components. For example, the phrase “processor configured (or set) to perform A, B, and C” refers to a processor dedicated to performing the operations (e.g., an embedded processor), or by executing one or more software programs stored in memory. It may refer to a general-purpose processor (e.g., CPU or application processor) that can perform the corresponding operations.

In addition, in the present disclosure, when a component is referred to as “connected” or “connected” to another component, the component may be directly connected or directly connected to the other component, but in particular, the contrary It should be understood that unless a base material exists, it may be connected or connected through another component in the middle.

Below, with reference to the attached drawings, embodiments of the present disclosure will be described in detail so that those skilled in the art can easily practice them. However, an embodiment of the present disclosure may be implemented in various different forms and is not limited to the embodiment described herein. In order to clearly describe an embodiment of the present disclosure in the drawings, parts that are not related to the description are omitted, and similar parts are assigned similar reference numerals throughout the present disclosure.

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings.

1 is a diagram for explaining an electronic device according to an embodiment of the present disclosure.

Referring to FIG. 1, in one embodiment of the present disclosure, the electronic device 100 may include a display 130. The electronic device 100 may display an image 160 through the display 130.

In one embodiment, the image 160 that the electronic device 100 provides to the user 150 through the display 130 may vary depending on the location of the user 150. In one embodiment, the image 160 provided by the electronic device 100 may be an image that can provide a three-dimensional effect to the user 150 using the electronic device 100.

In one embodiment, the electronic device 100 reproduces light reflected from a real-world object and provided to the user 150. In one embodiment, the electronic device 100 provides the user 150 with light 140 having the same path as the light reflected from a real object and provided to the user 150. The user 150 is provided with light 140 having the same path as light reflected from a real object through the image 160 displayed on the electronic device 100. Accordingly, the user 150 can feel the three-dimensional effect of the object included in the image 160 displayed on the electronic device 100, as if viewing a real object.

In one embodiment, the electronic device 100 may provide different images 160 depending on the location of the user 150, allowing the user 150 to feel the three-dimensional effect of the object included in the image 160. For convenience of explanation, the image 160 is described as including an object having a hexahedron shape as shown in FIG. 1.

In one embodiment, when the user 150 is located in front of the electronic device 100, the electronic device 100 may provide the user 150 with an image 160 including the front of an object.

In one embodiment, when the user 150 is located in a first direction that intersects the direction perpendicular to the front, rather than in the front of the electronic device 100, the electronic device 100 sends the user 150 an object. An image 160 including a first side and a front view may be provided. In one embodiment, the first side of the object and the front of the object that the electronic device 100 provides to the user 150 may vary depending on the angle between the direction perpendicular to the front and the first direction. In one embodiment, depending on the angle between the direction perpendicular to the front and the first direction, the electronic device 100 may provide the user 150 with an image 160 including only the first side of the object.

In one embodiment, when the user 150 intersects the direction perpendicular to the front of the electronic device 100 and is located in a second direction different from the first direction, the electronic device 100 An image 160 including a second side and a front side different from the first side of the object may be provided. In one embodiment, the second side of the object and the front of the object that the electronic device 100 provides to the user 150 may vary depending on the angle between the direction perpendicular to the front and the second direction. In one embodiment, depending on the angle between the direction perpendicular to the front and the second direction, the electronic device 100 may provide the user 150 with an image 160 including only the second side of the object.

In one embodiment, the first side and front of the object may be areas of the object that can be seen in reality when the user 150 looks at the object in a first direction that intersects a direction perpendicular to the front of the object. In one embodiment, the second side and the front of the object may be areas of the object that can be seen in reality when the user 150 looks at the object in a second direction that intersects a direction perpendicular to the front of the object.

In one embodiment, the electronic device 100 may provide the user 150 with light 140 through the image 160, similar to viewing an object in reality. Accordingly, the user 150 can feel the three-dimensional effect of the object included in the image 160 displayed by the electronic device 100.

In one embodiment, the electronic device 100 provides different images 160 to the left eye and right eye of the user 150, allowing the user 150 to maintain binocular disparity. It can make you feel it. In one embodiment, the first direction may be a direction in which the user 150 looks at the electronic device 100 through the left eye. The second direction may be a direction in which the user 150 looks at the electronic device 100 through the right eye. The user 150 can feel binocular disparity because the images 160 provided to the right eye and the left eye are different, and through this, the user 150 can also feel the three-dimensional effect of an object.

In one embodiment, display 130 may include a plurality of panels 110 and 120. In one embodiment, the plurality of panels 110 and 120 may be stacked and arranged. Although the display 130 is shown in FIG. 1 as including two panels 110 and 120, the present disclosure is not limited thereto. Display 130 may include three or more panels. Additionally, the display 130 may include an optical layer 213 (see FIG. 3) disposed between a plurality of panels. The electronic device 100 may provide the image 160 to the user 150 by displaying a base image and a layer image, which will be described later, on each of the two panels 110 and 120.

FIG. 2 is a diagram for explaining the operation of an electronic device according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same components as those described in FIG. 1 will be omitted.

Referring to FIG. 2 , the electronic device 100 may include a display 210, a memory 230 that stores at least one instruction, and at least one processor 220. At least one processor 220 may control the operation of the electronic device 100 by executing at least one instruction included in the memory 230. In one embodiment, the display 210 may include a base panel 211, a layer panel 212, an optical layer 213, and a backlight 214.

In one embodiment, the input content 240 may include information about the image 260 to be provided to the user 250 through the electronic device 100. The input content 240 may include a plurality of frame images, each corresponding to a plurality of frames. The input content 240 may include information about the image 260 in which a plurality of frame images are sequentially displayed.

In one embodiment, the input content 240 may be information obtained by photographing a real-world object from a plurality of different views. The input content 240 may be information obtained by photographing a real object from a plurality of different views during a plurality of frames. However, the present disclosure is not limited thereto, and the input content 240 may include information generated to provide the user 250 with images of an object from a plurality of different views.

In one embodiment, when a real object moves during a plurality of frames, the positions of the objects included in each of the plurality of frame images included in the input content 240 may be different. For example, when a real object moves during a plurality of frames, the positions of the objects included in each plurality of frame images may be different. When a part of a real object moves during a plurality of frames, the positions of the parts corresponding to the moving part of the real object among the objects included in each plurality of frame images may be different.

In one embodiment, the at least one processor 220 processes a first frame image 240_1 in a first frame of the input content 240 and a second frame image 240_2 in a second frame that is the frame immediately preceding the first frame. ) can be obtained. When the input content 240 includes information about the moving image 260 during the first frame and the second frame, the location of the object included in the first frame image 240_1 and the second frame image 240_2 The positions of the included objects may be different.

In one embodiment, the input content 240 may include multiple view images obtained by photographing an object from multiple different views. In one embodiment, each of the plurality of view images may be an image acquired through a plurality of cameras disposed at different locations and photographing objects from different views. Additionally, in one embodiment, each of the plurality of view images may be an image acquired from different views through a camera including a micro lens array. Hereinafter, for convenience of explanation, multiple view images are defined as images acquired through multiple cameras that photograph objects from different views.

In one embodiment, each of the first frame image 240_1 and the second frame image 240_2 included in the input content 240 may include images having a plurality of different views.

Figure 2 shows input content 240 including a human face. The input content 240 may include information obtained by photographing a person's face whose expression changes during the first frame and the second frame. The facial expression of the person included in the first frame image 240_1 may be different from the facial expression of the person included in the second frame image 240_2. However, the illustration in FIG. 2 is for convenience of explanation, and the object is not limited to the shape of a human face. Input content 240 may include information obtained by photographing various types of objects.

In one embodiment, the electronic device 100 is provided with input content 240. In one embodiment, at least one processor 220 included in the electronic device 100 may receive a first frame image 240_1 and a second frame image 240_2.

At least one processor 220 executes at least one instruction included in the memory 230 to create the first frame image 240_1, which is an image in the first frame of the input content 240, and the input content ( A correction-base image and a layer image to be displayed on the display 210 in the first frame may be generated based on the second frame image 240_2, which is the image in the second frame of 240).

At least one processor 220 displays a correction-base image on the base panel 211 and a layer image on the layer panel 212 to provide the user 250 with an image 260 that reproduces the captured object. can do.

Hereinafter, the process by which at least one processor 220 generates a correction-base image and a layer image will be described later with reference to FIGS. 4 to 12B.

In one embodiment, the display 210 is shown in FIG. 2 as including a back light 214 and an optical layer 213. In one embodiment, backlight 214 may generate light and provide it to user 250. In one embodiment, the light generated by the backlight 214 may be white light. However, the present disclosure is not limited to this, and the light provided by the backlight 214 may have a color other than white.

In one embodiment, the base panel 211 and the layer panel 212 may each include a plurality of pixels. In one embodiment, when the base panel 211 and the layer panel 212 are liquid crystal displays, each of the base panel 211 and the layer panel 212 includes a plurality of color filters. It may be a filter layer that does. In one embodiment, each of the plurality of pixels may correspond to a plurality of color filters.

In one embodiment, at least one of the base panel 211 or the layer panel 212 may include a plurality of red, green, and blue pixels. In one embodiment, at least one of the base panel 211 or the layer panel 212 may include a plurality of red, green, and blue color pixels and openings that do not filter light. In one embodiment, at least one of the base panel 211 or the layer panel 212 may include a plurality of yellow or blue pixels.

However, the present disclosure is not limited to this. In one embodiment, the base panel 211 and the layer panel 212 are divided according to the wavelength of light provided by the backlight 214 and the combination of colors for displaying the image 260 using the light provided by the backlight 214. The colors of a plurality of pixels included in may vary.

In one embodiment, the optical layer 213 may be disposed between the base panel 211 and the layer panel 212. The light 270 that has transmitted through the base panel 211 by the optical layer 213 may be refracted, reflected, dispersed, etc., and then provided to the layer panel 212. In FIG. 2, the optical layer 213 is shown as disposed between the base panel 211 and the layer panel 212, but the present disclosure is not limited thereto. In one embodiment, when the display 210 includes three or more panels, the optical layer 213 may be disposed somewhere other than between the base panel 211 and the layer panel 212. Additionally, in one embodiment, the optical layer 213 may be omitted. Hereinafter, for convenience of explanation, the light 270 passing through the optical layer 213 will be described as being refracted.

In one embodiment, the light 270 provided from the backlight 214 may pass through the base panel 211 and the layer panel 212, respectively, and be provided to the user 250.

In one embodiment, light 270 generated by the backlight 214 is provided to the user 250 through one pixel included in the base panel 211 and one pixel included in the layer panel 212. It can be. Light 270 provided to the user 250 according to at least one of the color or transmittance of each pixel of the base panel 211 and the layer panel 212 through which the light 270 generated by the backlight 214 passes. Wavelength and intensity are determined. The wavelength and intensity of the light 270 provided to the user 250 are determined according to the combination of the pixels of the base panel 211 and the pixels of the layer panel 212 through which the light 270 generated by the backlight 214 passes. .

At this time, depending on the location of the user 250, the pixels of the base panel 211 and the pixels of the layer panel 212 through which the light 270 generated by the backlight 214 passes to reach the user 250 change. Specifically, depending on the location of the user 250, each pixel of the base panel 211 and the pixel of the layer panel 212 through which the light 270 generated by the backlight 214 passes to reach the user 250. Color or transmittance may vary. In one embodiment, the pixels of the base panel 211 and the pixels of the layer panel 212 through which the light 270 generated by the backlight 214 passes to reach the user 250 according to the location of the user 250. Combinations may vary.

In one embodiment, when the display 210 includes an optical layer 213, light 270 provided from the backlight 214 transmits through the base panel 211 and is refracted by the optical layer 213. After passing through the layer panel 212, it may be provided to the user 250. The light whose path has been changed by being refracted by the optical layer 213 may be provided to the user 250 at a location corresponding to the changed path.

In one embodiment, optical layer 213 may include at least one lens. Although the optical layer 213 is shown in FIG. 2 as including three lenses, the present disclosure is not limited thereto. The optical layer 213 may include two or fewer lenses, or four or more lenses.

In one embodiment, the base panel 211 is formed according to the positional relationship between the pixels included in the base panel 211 and at least one lens included in the optical layer 213, and the number, shape, and refractive index of the at least one lens. The passing light 270 may be refracted and the path toward the layer panel 212 may be changed. The light whose path is refracted by the optical layer 213 may pass through the layer panel 212 and then be provided to the user 250 who views the electronic device 100 at a position corresponding to the path. Accordingly, light passing through different pixels of the base panel 211 and pixels of the layer panel 212 is provided to the user 250 depending on the position of the user 150, and includes different sides of the object to the user 250. A video 260 can be provided.

In one embodiment, when the display 210 includes the optical layer 213, compared to the case where the display 210 does not include the optical layer 213, the path of light passing through the base panel 211 is the electronic device 100. The angle formed with the direction perpendicular to the front may increase. Accordingly, the viewing angle at which the user 250 can view the image 260 through the electronic device 100 may increase. Additionally, the number, shape, and refractive index of at least one lens included in the optical layer 213 may be adjusted to vary the number of views provided to the user 250 and the spacing between adjacent views.

At this time, 'view' may correspond to a position where the user 250 can see different sides of the object included in the image 260 displayed through the electronic device 100. As the location of the user 250 corresponds to one of the plurality of views, the electronic device 100 provides the user 250 with images 260 including sides of different objects to It can create a three-dimensional feeling.

Hereinafter, for convenience of explanation, the light passing through the base panel 211 will be referred to as the base image displayed on the base panel 211, and the light passing through the layer panel 212 will be referred to as the base image displayed on the layer panel 212. It is called a layer image. At this time, the base image may include information such as the transmittance and color of a plurality of pixels corresponding to the light passing through the base panel 211. The layer image may include information such as transmittance and color of a plurality of pixels corresponding to light passing through the layer panel 212.

In one embodiment, the base image included in the base panel 211 may include a plurality of sub-base images each corresponding to a plurality of different views. Light emitted from each of the plurality of sub-base images may be refracted by the optical layer 213 and then provided to the layer panel 212. Light emitted from each of the plurality of sub-base images may be refracted in different paths by the optical layer 213 and provided to the layer panel 212. Light provided to the layer panel 212 through different paths may pass through different areas of the layer image and then be provided to the user 250 in each view.

Accordingly, the electronic device 100 may provide the user 250 with an image 260 that changes in response to a change in the user 250's location. The user 250 can feel the three-dimensional effect of the object included in the image 260 based on the image 260 provided by the electronic device 100 changing in response to a change in the user's 250 location.

In one embodiment, when the user 250 is located in a first direction that intersects the direction perpendicular to the front of the electronic device 100, the image 260 provided by the electronic device 100 to the user 250 and Image 260 provided by the electronic device 100 to the user 250 when the user 250 intersects the direction perpendicular to the front of the electronic device 100 and is located in a second direction different from the first direction. may be different.

In one embodiment, the electronic device 100 may provide the user 250 located in the first direction with an image 260 that makes the user 250 feel as if he or she is looking at an object in the first direction. In one embodiment, the electronic device 100 may provide the user 250 located in the second direction with an image 260 that makes the user 250 feel as if he or she is looking at an object in the second direction. Accordingly, the user 250 can feel the three-dimensional effect of the object included in the image 260 displayed by the electronic device 100.

However, the present disclosure is not limited to this, and at least one of the base panel 211 and the layer panel 212 is each of the base panel 211 and the layer panel 212, such as an organic light emitting diode display, an inorganic light emitting diode display, etc. It could be a display that generates light on its own. If both the base panel 211 and the layer panel 212 are panels that generate light by themselves, the display 210 may not include the backlight 214. In this case, the electronic device 100 adjusts the intensity, wavelength, and transmittance of each sub-panel of the light generated from each panel to provide a different image 260 to the user 250 depending on the location of the user 250. It may be possible.

Figure 3 is a block diagram for explaining the configuration of an electronic device according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same components as those described in FIGS. 1 and 2 will be omitted.

Referring to FIGS. 2 and 3 , in one embodiment, the electronic device 100 may include a display 210, a memory 230, at least one processor 220, and a communication interface 280. However, not all components shown in FIG. 3 are essential components. The electronic device 100 may be implemented with more components than those shown in FIG. 3, or may be implemented with fewer components. The display 210, memory 230, at least one processor 220, and communication interface 280 may each be electrically and/or physically connected to each other.

In one embodiment, the display 210 is one of a liquid crystal display, a plasma display, an organic light emitting diodes display, and an inorganic light emitting diodes display. may include. However, the present disclosure is not limited thereto, and the display 210 may include other types of displays capable of providing the image 260 to the user 250.

In one embodiment, the display 210 may include a base panel 211, a layer panel 212, an optical layer 213, and a backlight 214. However, the display 210 may include more or fewer components than those shown in FIG. 2 . As an example, the display 210 may include only the base panel 211 and the layer panel 212. Additionally, the display 210 may include three or more panels in addition to the optical layer 213 and the backlight 214.

In one embodiment, the memory 230 is a flash memory type, hard disk type, multimedia card micro type, or card type memory (e.g., SD or XD). Memory, etc.), RAM (Random Access Memory), SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), Mask ROM, Flash ROM, etc.), a hard disk drive (HDD), or a solid state drive (SSD). The memory 230 may store instructions or program codes for performing functions or operations of the electronic device 100. Instructions, algorithms, data structures, program codes, and application programs stored in the memory 230 may be implemented in programming or scripting languages such as C, C++, Java, assembler, etc.

In one embodiment, various types of modules that can be used to provide the image 260 to the user 250 through the display 210 may be stored in the memory 230. The memory 230 may store an input content acquisition module 231, an image generation module 232, and an image correction module 233. However, not all modules shown in FIG. 3 are essential modules. The memory 230 may store more or fewer modules than the modules shown in FIG. 3 . In one embodiment, the memory 230 may further store a module for preprocessing the obtained input content.

A 'module' included in the memory 230 may refer to a unit that processes functions or operations performed by at least one processor 220. The 'module' included in the memory 230 may be implemented as software such as instructions, algorithms, data structures, or program code.

In one embodiment, the input content acquisition module 231 may be composed of instructions or program code related to an operation or function of acquiring the input content 240. The input content acquisition module 231 may be composed of commands or program codes related to an operation or function of receiving input content 240 from an external server or surrounding electronic devices. In one embodiment, the input content acquisition module 231 is an operation or function of acquiring the first frame image 240_1 in the first frame and the second frame image 240_2 in the second frame of the input content 240. It may consist of instructions or program code related to.

In one embodiment, the image generation module 232 generates a base image corresponding to the base panel 211 in the first frame and a layer panel 212 in the first frame based on the acquired input content 240. It may consist of instructions or program code related to the operation or function of creating a layer image. In one embodiment, the layer image may be an image displayed on the layer panel 212 in the first frame.

In one embodiment, the image generation module 232 may be composed of instructions or program code related to an operation or function of generating a base image and a layer image by performing factorization. In one embodiment, the image generation module 232 may include a first artificial intelligence model that infers the base image and the layer image by performing factorization based on the input content 240. In one embodiment, the first artificial intelligence model included in the image generation module 232 may include a machine learning or deep learning model.

In one embodiment, the first artificial intelligence model included in the layer image generation module 330 is based on the first frame image 240_1 and the second frame image 240_2, and the base corresponding to the base panel in the first frame is It may be an artificial intelligence model trained to infer a layer image corresponding to a layer panel from the image and the first frame.

In one embodiment, the electronic device 100 may train a first artificial intelligence model included in the image generation module 232. The electronic device 100 may perform transfer learning using a pre-learning model to train the first artificial intelligence model included in the image generation module 232. However, the present disclosure is not limited to this, and the image generation module 232 performs factorization to create a base image and a layer image in the first frame based on the first frame image 240_1 and the second frame image 240_2. The first artificial intelligence model learned to infer may be received from an external server or surrounding electronic devices through the communication interface 280. Hereinafter, the operation of the image generation module 232 and the training of the first artificial intelligence model included in the image generation module 232 will be described later with reference to FIGS. 6 to 11.

In one embodiment, the image correction module 233 includes instructions or programs related to an operation or function of generating a correction-base image based on the movement between the first frame image and the second frame image, based on the base image or layer image. It can be composed of code. In one embodiment, the image correction module 233 reflects the movement between the first frame image and the second frame image based on the base image or the layer image, and produces a correction-base image displayed on the base panel in the first frame. can be created.

In one embodiment, when the display 210 includes the optical layer 213 and the base image includes a plurality of sub-base images corresponding to a plurality of different views, the image correction module 233 is configured to operate the base image or Based on the layer image, based on the difference between a plurality of sub-base images, it may be composed of instructions or program code related to an operation or function of generating a correction-base image including a plurality of sub-correction-base images. . In one embodiment, the image correction module 233 reflects differences between a plurality of sub-base images based on the base image or layer image and displays the image on the base panel in correspondence to a plurality of different views in the first frame. A correction-base image can be created.

In one embodiment, the image correction module 233 may include a second artificial intelligence model that infers a correction-base image based on the base image. In one embodiment, the second artificial intelligence model included in the image correction module 233 may include a machine learning or deep learning model.

In one embodiment, the second artificial intelligence model included in the image correction module 233 is an artificial intelligence trained to infer the correction-base image displayed on the base panel in the first frame, based on the base image. It could be a model.

In one embodiment, the electronic device 100 may train a second artificial intelligence model included in the image correction module 233. The electronic device 100 may perform transfer learning using a pre-learning model to train the second artificial intelligence model included in the image correction module 233. However, the present disclosure is not limited to this, and the image correction module 233 uses the learned second artificial intelligence model to infer the correction-base image based on the base image to an external server or surrounding area through the communication interface 280. It can also be received from electronic devices. Hereinafter, the operation of the image correction module 233 and the training of the second artificial intelligence model included in the image correction module 233 will be described later with reference to FIGS. 6 to 9, 12A, and 12B.

In one embodiment, the first artificial intelligence model included in the image generation module 232 and the second artificial intelligence model included in the image correction module 233 may include a plurality of neural network layers. Each neural network layer has a plurality of weight values, and the calculation of the current neural network layer can be performed through the calculation result of the previous neural network layer and the calculation of the plurality of weights. Examples of artificial intelligence models include Convolutional Neural Network (CNN), Deep Neural Network (DNN), Recurrent Neural Network (RNN), Restricted Boltzmann Machine (RBM), Deep Belief Network (DBN), and Bidirectional Recurrent Deep Neural Network (BRDNN). , Deep Q-Networks, GAN (Generative Adversarial Networks), VAE (Variational Auto Encoder), etc., and the first artificial intelligence model and the second artificial intelligence model in the present disclosure are limited to the examples described above. It doesn't work.

In one embodiment, the memory 230 may further store a preprocessing module consisting of instructions or program code related to an operation or function for preprocessing the acquired input content 240. The preprocessing module is a command for the operation or function of preprocessing the input content 240 by performing wrangling, transformation, integration, cleaning, reduction, discretization, etc. It may consist of fields or program code. The image generation module 232 may generate a base image and a layer image based on the input content 240 preprocessed through the preprocessing module.

In one embodiment, the communication interface 280 may perform data communication with an external server under the control of at least one processor 220. Additionally, the communication interface 280 can perform data communication not only with an external server but also with other peripheral electronic devices. The communication interface 280 includes, for example, wired LAN, wireless LAN, Wi-Fi, Bluetooth, zigbee, WFD (Wi-Fi Direct), and infrared communication (IrDA, infrared Data Association), BLE (Bluetooth Low Energy), NFC (Near Field Communication), Wibro (Wireless Broadband Internet, Wibro), WiMAX (World Interoperability for Microwave Access, WiMAX), SWAP (Shared Wireless Access Protocol), WiGig Data communication may be performed with a server or other surrounding electronic devices using at least one of data communication methods including (Wireless Gigabit Allicance, WiGig) and RF communication.

In one embodiment, at least one processor 220 may receive input content 240 from an external server or surrounding electronic devices through the communication interface 280. In one embodiment, at least one processor 220 may transmit the generated correction-base image and layer image to an external server or surrounding electronic devices through the communication interface 280.

FIG. 4 is a flowchart for explaining the operation of an electronic device according to an embodiment of the present disclosure.

Referring to FIGS. 2, 3, and 4, in one embodiment, the operating method of the electronic device 100 includes the first frame image 240_1 in the first frame of the input content 240 and the first frame. It may include acquiring the second frame image 240_2 in the second frame, which is the previous frame (S100). At least one processor 220 may execute the input content acquisition module 231 to obtain the first frame image 240_1 and the second frame image 240_2.

In one embodiment, the method of operating the electronic device 100 includes applying the first frame image 240_1 and the second frame image 240_2 to the image generation module 232 to generate the base panel 211 in the first frame. It may include generating a layer image corresponding to the layer panel 212 in the base image and the first frame and based on the movement between the first frame image and the second frame image (S200). At least one processor 220 executes the image generation module 232 to generate a base image corresponding to the base panel 211 in the first frame based on the first frame image 240_1 and the second frame image 240_2. And it corresponds to the layer panel 212 in the first frame and can generate a layer image based on the movement between the first frame image and the second frame image. Hereinafter, the base image and layer image will be described in detail with reference to FIGS. 6 to 8.

In one embodiment, the method of operating the electronic device 100 generates a correction-base image based on the movement between the first frame image and the second frame image by applying the base image or layer image to the image correction module 233. It may include a step (S300). At least one processor 220 may execute the image correction module 233 to generate a correction-base image based on the movement between the first frame image and the second frame image based on the base image or layer image.

In one embodiment, in the step of generating a correction-base image (S300), the correction-base image may be generated by applying the first frame image and the layer image to the image correction module 233. At least one processor 220 may execute the image correction module 233 to generate a correction-base image based on the first frame image and the layer image.

In one embodiment, in the step of generating the correction-base image (S300), the correction-base image may be generated based on the base image. At least one processor 220 may execute the image correction module 233 to generate a correction-base image based on the base image. Hereinafter, the correction-base image will be described in detail in FIGS. 6 to 8.

In one embodiment, the method of operating the electronic device 100 may include displaying a correction-base image on the base panel 211 and displaying a layer image on the layer panel 212 (S400). At least one processor 220 may control the base panel 211 to display a correction-base image and control the layer panel 212 to display a layer image.

FIG. 5 is a flowchart for explaining the operation of an electronic device including an optical layer according to an embodiment of the present disclosure. Hereinafter, steps that are the same as those described in FIG. 4 will be given the same reference numerals and descriptions will be omitted.

2, 3, and 5, when the electronic device 100 further includes an optical layer 213 disposed between the base panel 211 and the layer panel 212, the base image is divided into a plurality of different It may include a plurality of sub-base images, each corresponding to a view. Each of the plurality of sub-base images provided through the base panel 211 may be refracted in different directions by the optical layer 213 and provided to the layer panel 212.

In one embodiment, in the step of generating the base image and the layer image (S200a), the first frame image and the second frame image are applied to the image generation module 232 to correspond to the base panel 211 in the first frame. And, a base image and a layer image including a plurality of sub-base images respectively corresponding to a plurality of different views can be generated. At least one processor 220 executes the image generation module 232 to generate a plurality of different views corresponding to the base panel 211 in the first frame based on the first frame image and the second frame image. A layer image corresponding to the layer panel 212 may be generated from the base image and the first frame, each of which includes a plurality of corresponding sub-base images.

In one embodiment, in the step of generating a correction-base image (S300a), the base image and the layer image are applied to the image correction module 233 to determine the difference between a plurality of sub-base images according to a plurality of different views. A correction-base image that includes a plurality of sub-correction-base images can be generated. At least one processor 220 executes an image correction module 233 to perform correction based on differences between a plurality of sub-base images according to a plurality of different views based on the base image and a layer image, and perform a plurality of sub-correction- A correction-base image including base images can be generated. Hereinafter, the base image and the correction-base image when the electronic device 100 includes the optical layer 213 will be described later with reference to FIGS. 7 and 8.

FIG. 6 is a diagram for explaining the operation of an electronic device that generates a layer image and a correction-base image according to an embodiment of the present disclosure. FIG. 7 is a diagram for explaining movement between a first frame image and a second frame image according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same components as those described in FIG. 3 will be omitted.

FIG. 6 is a diagram for explaining the operation of the electronic device 100 when the electronic device 100 includes a base panel 211 and a layer panel 212. At this time, the electronic device 100 may further include a backlight 214 that provides light to the base panel 211 and the layer panel 212.

3, 4, and 6, in one embodiment, at least one processor 220 may obtain input content 630. The input content 630 includes a first frame image 630_1 in the first frame of the input content 630 and a second frame image 630_2 in the second frame, which is the frame immediately preceding the first frame of the input content 630. may include. The first frame image 630_1 may include multiple view images 631 obtained from multiple different views. The second frame image 630_2 may include multiple view images 632 obtained from multiple different views.

In one embodiment, input content 630 may include information about an object moving during a plurality of frames. However, the present disclosure is not limited thereto, and the input content 630 includes information about an object whose shape, size, or part of the object (for example, a change in the expression of a person's face, etc.) changes during a plurality of frames. It may also be included. Accordingly, among the plurality of frames included in the input content, the information included in the first frame image in the first frame and the information included in the second frame image in the second frame, which is the frame immediately preceding the first frame, may be different. You can.

Referring to FIGS. 6 and 7 , a first frame image 731_1 and a second frame image 732_1 are shown in FIG. 7 . The first frame image 731_1 and the second frame image 732_1 include information about the first objects 733_1 and 733_2 and the second objects 734_1 and 734_2, respectively.

In one embodiment, the first object 733 included in the input content 630 may be an object that moves during a plurality of frames. The second object 734 included in the input content 630 may be an object that does not move for a plurality of frames. Accordingly, the location of the first object 733_1 included in the first frame image 731_1 and the location of the first object 733_2 included in the second frame image 732_1 may be different from each other. The positions of the second object 734_1 included in the first frame image 731_1 and the second object 734_2 included in the second frame image 732_1 may be the same.

In one embodiment, the movement 735 between the first frame image 731_1 and the second frame image 732_1 may be calculated by comparing the first frame image 731_1 and the second frame image 732_1. The movement 735 between the first frame image 731_1 and the second frame image 732_1 is an optical flow calculated based on the first frame image 731_1 and the second frame image 732_2, which are consecutive frames. ) can mean.

In one embodiment, the movement 735 between the first frame image 731_1 and the second frame image 732_1 is caused by the movement of the first object 733_1 and the second frame image 732_1 included in the first frame image 731_1. ) may include information about the movement of the first object 733_2 included in . Movement 735 between the first frame image 731_1 and the second frame image 732_1 may include information about the direction and distance in which the first object 733 moved from the second frame to the first frame. . Hereinafter, a method for calculating the movement 735 between the first frame image 731_1 and the second frame image 732_1 will be described later with reference to FIG. 12A.

In one embodiment, the at least one processor 220 applies the first frame image 630_1 and the second frame image 630_2 to the image generation module 610 to generate a panel image 640 corresponding to the display 210. ) can be created. When the display 210 includes the base panel 211 and the layer panel 212, the first frame image 630_1 and the second frame image 630_2 are applied to the image generation module 610, so that the base panel ( A base image 640_2 corresponding to 211) and a layer image 640_1 corresponding to the layer panel 212 may be generated.

In one embodiment, the base image 640_2 and the layer image 640_1 are displayed on the base panel 211 and the layer panel 212, respectively, and are input content as a combination of the base image 640_2 and the layer image 640_1. This is an image that can provide a three-dimensional image (260, see FIG. 2) that reproduces the object included in 630 to the user (250, see FIG. 2).

In one embodiment, when both the base panel 211 and the layer panel 212 are non-self-emitting displays, such as a liquid crystal display, the base image 640_2 and the layer image 640_1 This combined image 260 may be provided to the user 250.

In one embodiment, when the layer panel 212 is a self-emitting display such as an organic light emitting diode display, an inorganic light emitting diode display, etc., an image obtained by multiplying the base image 640_2 and the layer image 640_1 ( 260) may be provided to the user 250.

In one embodiment, the image generation module 610 generates a first artificial image trained to infer the base image 640_2 and the layer image 640_1 based on the first frame image 630_1 and the second frame image 630_2. May include intelligence models.

In one embodiment, the first artificial intelligence model is based on the first frame image 630_1 and the second frame image 630_2, and the base image 640_2 and the first frame image 640_2 corresponding to the base panel 211 in the first frame. It may be an artificial intelligence model that corresponds to the layer panel 212 in the frame and is trained to infer the layer image 640_1 based on the movement 735 between the first frame image 630_1 and the second frame image 630_2.

In one embodiment, the first artificial intelligence model generates movement 735 between the first frame image 630_1 and the second frame image 630_2 based on the first frame image 630_1 and the second frame image 630_2. It may be a model trained to infer the layer image 640_1 that reflects .

In one embodiment, the object included in the first frame image 630_1 and the object included in the second frame image 630_2 are each reproduced as an image by combining a base image and a layer image. When the position of the object included in the first frame image 630_1 and the position of the object included in the second frame image 630_2 are different from each other, in the first frame generated based on the first frame image 630_1 The base image in the second frame generated based on the base image and the second frame image 630_2 may also be different.

At this time, the base image in the first frame and the base image in the second frame do not reflect the movement from the second frame to the first frame, and only appear in the first frame image 630_1 and the second frame image 630_2, respectively. Since it is generated based on the base image in the first frame and the base image in the second frame, not only the position information of the object but also information such as luminance, gradation, and color of the object may be different from each other.

In addition, the layer image in the first frame generated based on the first frame image 630_1 and the layer image in the second frame generated based on the second frame image 630_2 are also the base images in the first frame, respectively. and may be generated to correspond to the base image in the second frame. Accordingly, not only the position information of the object included in the layer image in the first frame and the layer image in the second frame, but also information such as luminance, grayscale, and color of the object may be different.

In one embodiment, at least one information of the luminance, grayscale, or color of an object included in the layer image and the base image in the first frame and the luminance, grayscale, or color of the object included in the layer image and the base image in the second frame When information on at least one of the colors is different, in addition to the movement of the object included in the image 260 from the second frame to the first frame, at least one of the luminance, gradation, or color of the object changes, causing flicker (flicker) to the user 250. flicker) may be recognized. Accordingly, the visibility of the user 250 and the quality of the image 260 may be reduced.

The image generation module 610 of the present disclosure is trained to infer the layer image 640_1 in the first frame by reflecting the movement 735 between the first frame image 630_1 and the second frame image 630_2. Due to the inclusion of an artificial intelligence model, when the position of the object included in the first frame image 630_1 and the position of the object included in the second frame image 630_2 are different from each other, comparison is made with the layer image in the second frame. Thus, it is possible to generate a layer image in the first frame in which only the location information of the object is different, and information such as luminance, gradation, and color of the object has little difference from the layer image in the second frame. Hereinafter, the training method of the first artificial intelligence model will be described later with reference to FIGS. 11 to 12B.

In one embodiment, the image correction module 620 includes instructions or program code to perform an operation or function of generating a correction-base image 650 according to the first frame image 630_1 and the layer image 640_1. can do.

In one embodiment, the image correction module 620 corrects the base image (base image) according to the first frame image 630_1 and the layer image 640_1, based on the display types of the base panel 211 and the layer panel 212. 650) may include instructions or program code to perform an operation or function.

In one embodiment, the first frame image 630_1 is a combination of the correction-base image 650 displayed on the base panel 211 and the layer image 640_1 displayed on the layer panel 212 by the electronic device 100. It may be an image corresponding to the image 260 generated by and intended to be provided to the user 250.

In one embodiment, when both the base panel 211 and the layer panel 212 are non-self-emitting displays such as a liquid crystal display, the image correction module 620 stores the first frame image. It may include instructions or program code to perform an operation or function of generating the correction-base image 650 according to the difference between the layer image 630_1 and the layer image 640_1.

In one embodiment, when the layer panel 212 is a self-emitting display such as an organic light-emitting diode display, an inorganic light-emitting diode display, etc., the image correction module 620 selects the layer in the first frame image 630_1. It may include commands or program codes to perform an operation or function of dividing the image 640_1 and generating the base image 650.

In one embodiment, the at least one processor 220 applies the layer image 640_1 and the first frame image 630_1 to the image correction module 620, thereby creating the first frame image 630_1 and the second frame image ( A correction-base image 650 based on the movement of the liver (630_2) can be generated. The correction-base image 650 may be an image corresponding to the base panel 211 in the first frame.

In one embodiment, the at least one processor 220 displays the correction-base image 650 on the base panel 211 and the layer image 640_1 on the layer panel 212, thereby Even if input content 630 including a moving object for one frame is provided, only the position information of the object is different in the first frame compared to the second frame, and the luminance, gradation, color, etc. of the object are the same in the image 260 is provided to the user 250, the visibility of the user 250 and the quality of the image 260 can be improved.

FIG. 8 is a diagram for explaining the operation of an electronic device that generates a correction-base image including a layer image and a plurality of sub-correction-base images according to an embodiment of the present disclosure. FIG. 9 is a diagram for explaining the operation of an electronic device that generates a correction-base image through an image correction module including a second artificial intelligence model, according to an embodiment of the present disclosure. FIG. 10 is a diagram for explaining shift values according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same components as those described in FIGS. 3 and 6 will be omitted.

Referring to FIGS. 2 and 8 , FIG. 8 shows that the electronic device 100 includes a base panel 211, a layer panel 212, and an optical layer 213 disposed between the base panel 211 and the layer panel 212. This is a diagram for explaining the operation of the electronic device 100 when it includes. At this time, the electronic device 100 may further include a backlight 214 that provides light to the base panel 211 and the layer panel 212.

3, 4, 6, and 8, in one embodiment, at least one processor 220 may obtain input content 830. The input content 830 may include a first frame image 830_1 in the first frame of the input content 830 and a second frame image 830_2 in the second frame of the input content 830. The first frame image 830_1 may include multiple view images 831 obtained from multiple different views. The second frame image 830_2 may include multiple view images 832 obtained from multiple different views.

In one embodiment, the at least one processor 220 applies the first frame image 830_1 and the second frame image 830_2 to the image generation module 810 to generate a panel image 840 corresponding to the display 210. ) can be created. At least one processor 220 applies the first frame image 830_1 and the second frame image 830_2 to the image generation module 810, thereby creating a base image 840_2 and a layer panel corresponding to the base panel 211. A layer image 840_1 corresponding to 212 can be generated.

In one embodiment, the at least one processor 220 further applies information about the base panel 211, layer panel 212, and optical layer 213 included in the display 210 to the image generation module 810. You may. In this case, the at least one processor 220 refers to the information about the base panel 211, the layer panel 212, and the optical layer 213, and creates the first frame image 830_1 and the second frame image 830_2. Based on this, the base image 840_2 and the layer image 840_1 can be generated.

In one embodiment, the information about the base panel 211 may include information about at least one of the resolution, display type, or brightness of the base panel 211. The information about the layer panel 212 may include information about at least one of the resolution, display type, or brightness of the layer panel 212. The information about the optical layer 213 may include information about at least one of the number of at least one lens included in the optical layer 213, the size of at least one lens, or the refractive index of at least one lens.

In one embodiment, the base image 840_2 may include a plurality of sub-base images 841, each corresponding to a plurality of different views. Each of the plurality of sub-base images 841 is refracted by the optical layer 213 and then passes through different areas of the layer image 840_1 displayed on the layer panel 212 to produce a view corresponding to the refracted path. It may be provided to the user 250 located in . In one embodiment, each of the plurality of sub-base images 841 is a base image for providing an image 260 including different sides of an object to the user 250 from different views.

In one embodiment, each of the plurality of sub-base images 841 is a base image for providing an image 260 including different sides of an object from different views, so each of the plurality of sub-base images 841 The location information of the object included may be different. Additionally, information such as luminance, grayscale, and color of an object included in each of the plurality of sub-base images 841 may be different from each other. Therefore, when the base image 840_2 including a plurality of sub-base images 841 is displayed on the base panel 211, at least one of the luminance, gradation, or color of the object changes according to the change in view, thereby causing the user 250 Flicker may be noticeable. Accordingly, the visibility of the user 250 and the quality of the image 260 may be reduced.

In one embodiment, the image generation module 232 generates a base image 840_2 and a layer including a plurality of sub-base images 841 based on the first frame image 830_1 and the second frame image 830_2. It may include a first artificial intelligence model trained to infer the image 840_1.

In one embodiment, the first artificial intelligence model corresponds to the base panel 211 in the first frame based on the first frame image 830_1 and the second frame image 830_2 and includes a plurality of sub-base images ( 841), which corresponds to the layer panel 212 in the first frame and the base image 840_2, and is based on the movement 735 (see FIG. 7) between the first frame image 830_1 and the second frame image 830_2. It may be an artificial intelligence model trained to infer the layer image (840_1).

In one embodiment, the image correction module 620 generates a plurality of shift layer images by shifting the layer image 840_1 using each of a plurality of shift values 860, and generates a first frame image 830_1. It may include instructions or program codes for performing an operation or function of generating a correction-base image 850 including a plurality of sub-correction-base images 851 according to the plurality of shift layer images.

In one embodiment, the image correction module 620 includes a plurality of view images 831 included in the first frame image 830_1 and a plurality of shift layers shifted to correspond to each of the plurality of view images 831. It may include instructions or program codes for performing an operation or function of generating a correction-base image 850 including a plurality of sub-correction-base images 851 according to the images.

In one embodiment, the image correction module 820 corrects-base according to a plurality of view images 831 and a plurality of shift layer images based on the display type of the base panel 211 and the layer panel 212. It may include instructions or program code to perform an operation or function to generate the image 850.

In one embodiment, when both the base panel 211 and the layer panel 212 are non-self-emitting displays, such as a liquid crystal display, the image correction module 820 may provide a plurality of view images. A command or program code for performing an operation or function of generating a correction-base image 850 including a plurality of sub correction-base images 851 according to the difference between the shift layer images corresponding to each of the fields 831. may include.

In one embodiment, when the layer panel 212 is a self-emitting display such as an organic light emitting diode display, an inorganic light emitting diode display, etc., the image correction module 820 displays each of the plurality of view images 831. and a command or program code for performing an operation or function of dividing the corresponding shift layer image to generate a correction-base image 850 including a plurality of sub-correction-base images 851.

Referring to FIG. 10, a base panel 211, an optical layer 213, and a layer panel 212 included in the display 210 are shown. In one embodiment, optical layer 213 is disposed on base panel 211. The layer panel 212 is disposed on the optical layer 213. In one embodiment, the display 210 may further include a backlight 214 disposed at the bottom of the base panel 211.

In one embodiment, the base panel 211 and the layer panel 212 may be included in the display 210 with different intervals between them. In one embodiment, the layer panel 212 is positioned spaced apart from the base panel 211, and the distance between the layer panel 212 and the base panel 211 is defined as the reference distance 1010. In one embodiment, the reference distance 1010 may be the distance between the upper surface of the base panel 211 and the lower surface of the layer panel 212.

In one embodiment, the base panel 211 and the layer panel 212 may each include a plurality of pixels. In one embodiment, light provided from a backlight (not shown) illuminates one pixel among a plurality of pixels included in the base panel 211 and one pixel among a plurality of pixels included in the layer panel 212. It may be transmitted and provided to the user 250 (see FIG. 2).

In one embodiment, light transmitted through the base panel 211 is transmitted by the optical layer 213 based on at least one of the size of the at least one lens included in the optical layer 213, the shape of the lens, or the refractive index of the lens. The degree of refraction can be determined. The light passing through the base panel 211 is refracted by the optical layer 213, changes its path, and then passes through the layer panel 212 to be provided to the user 250.

In one embodiment, the base image 1051 displayed on the base panel 211 may include a first sub-base image 1051_1, a second sub-base image 1051_2, and a third sub-base image 1051_3. .

In one embodiment, each of the first to third sub-base images 1051_1, 1051_2, and 1051_3 may be refracted by the optical layer 213 and provided to users 250 located in different views. Each of the first to third sub-base images 1051_1, 1051_2, and 1051_3 may be an image for providing different sides of an object to the user 250 located in different views.

In one embodiment, the second sub-base image 1051_2 and the third sub-base image 1051_3 are images adjacent to the first sub-base image 1051_1. The view in which the user is provided with the image 260 of the first sub-base image 1051_1 may be adjacent to the view in which the user is provided with the image 260 of the second sub-base image 1051_2. The view in which the user is provided with the image 260 of the first sub-base image 1051_1 may be adjacent to the view in which the user is provided with the image 260 of the third sub-base image 1051_3.

In FIG. 10, the base image 1051 is shown as including three sub-base images, but the present disclosure is not limited thereto. Based on at least one of the resolution of the base panel 211, the size of at least one lens included in the optical layer 213, the shape of the lens, or the refractive index of the lens, the base image 1051 is divided into two sub-base images or four sub-base images. It may include more than one sub-base image.

In one embodiment, each of the first to third base images 1051_1, 1051_2, and 1051_3 is refracted by the optical layer 213 and then passes through an area in the corresponding optical path of the layer image 1040 to provide a user It can be provided to (250).

In one embodiment, the first sub-base image 1051_1, which is incident perpendicularly to the lens included in the optical layer 213 among the layer images 1040 and is not refracted and is directed to the front of the layer panel 212, is transmitted. The area can be defined as the first layer area (1040_1). After the second sub-base image 1051_2 adjacent to the first sub-base image 1051_1 of the layer image 1040 is refracted by the optical layer 213, the transmitted area will be defined as the second layer area 1040_2. You can. After the third sub-base image 1051_3 adjacent to the first sub-base image 1051_1 of the layer image 1040 is refracted by the optical layer 213, the transmitted area will be defined as the third layer area 1040_3. You can.

In one embodiment, the adjacent first sub-base image 1051_1 and the second sub-base image 1051_2 are respectively refracted by the optical layer 213, and then the pixels of the first layer area 1040_1 and the second layer transmit. The interval between pixels in the area 1040_2 can be defined as the reference shift value 1030.

In one embodiment, the reference shift value 1030 is the interval between adjacent views among a plurality of views through which the user 250 can see different sides of an object included in the image 260 displayed through the electronic device 100. It may be determined based on at least one of , the reference distance 1010, and the resolution of the layer panel 212. In one embodiment, as the reference distance 1010 increases, the reference shift value 1030 may increase. As the resolution of the layer panel 212 increases, the reference shift value 1030 may increase. As the spacing between adjacent views increases, the reference shift value 1030 may increase.

개라고 할 때, 사용자(250)의 위치가 전자 장치(100)의 정면에서 영상(260)을 볼 수 있는 뷰에 있는 것을 기준으로 전자 장치(100)의 좌측면으로

개의 뷰, 전자 장치(100)의 우측면으로

개의 뷰가 포함될 수 있다.In one embodiment, a plurality of shift values 860 (see FIG. 8) may be determined based on the reference shift value 1030 and a plurality of different views. A view that allows the user 250 to see different sides of an object included in the image 260 displayed through the electronic device 100

When referring to the dog, the position of the user 250 is on the left side of the electronic device 100 based on the view where the image 260 can be seen from the front of the electronic device 100.

View of the left side of the electronic device 100

Views may be included.

라고 할 때, 서로 다른 복수의 뷰에 각각 대응되는 복수의 쉬프트 값들(860)은

내지

사이의 값 및 0의 값을 가질 수 있다. 이때,

은 자연수이고, 0은 사용자(250)의 위치가 전자 장치(100)의 정면에서 영상(260)을 볼 수 있는 뷰에 있을 때의 쉬프트 값일 수 있다.In one embodiment, the reference shift value (1030) is

When doing so, a plurality of shift values 860 corresponding to a plurality of different views are

inside

It can have values between and 0. At this time,

is a natural number, and 0 may be a shift value when the user 250 is in a view where the image 260 can be viewed from the front of the electronic device 100.

In one embodiment, each of the plurality of shift values 860 may include information about a plurality of different views. A plurality of shift layer images generated by shifting the layer image 840-1 using respective shift values 860 include information about the layer image provided to the user 250 in a plurality of different views. can do.

Referring again to FIG. 8 , in one embodiment, at least one processor 220 applies the layer image 840_1, the first frame image 830_1, and a plurality of shift values 860 to the image correction module 820. By doing so, a correction-base image 850 based on the movement between the first frame image 830_1 and the second frame image 830_2 can be generated. The correction-base image 850 may be an image corresponding to the base panel 211 in the first frame.

In one embodiment, the at least one processor 220 processes the plurality of view images 831 among the plurality of view images 831 and the plurality of shift values 860 included in the first frame image 830_1. Based on the difference between a plurality of sub-base images 841 according to a plurality of different views, according to the shift layer image shifted according to the shift value corresponding to each view, a plurality of sub-correction-base images 851 ) can generate a correction-base image 850 including.

In one embodiment, a plurality of view images 831 including an object reproduced to the user 250 in each view and a plurality of shift layer images shifted according to a shift value corresponding to each view are obtained. Each of the plurality of sub-correction-base images 851 is different from adjacent sub-correction-base images only in the position information of the object according to different views, and the information such as luminance, gradation, and color of the object is the same. can do.

In one embodiment, the at least one processor 220 displays the correction-base image 850 on the base panel 211 and the layer image 840_1 on the layer panel 212, thereby Even if input content 830 including a moving object for 1 frame is provided, only the position information of the object is different in the first frame compared to the second frame, and the luminance, gradation, color, etc. of the object are the same in the image 260 is provided to the user 250, the visibility of the user 250 and the quality of the image 260 can be improved.

In addition, in order to provide the user 250 with images 260 including different sides of an object from different views, the correction-base image 850 including a plurality of sub-correction-base images 851 is used as a base image. Even when displayed on the panel 211, the image 260 is provided to the user 250 where only the location information of the object is different depending on the change in view and the luminance, gradation, color, etc. of the object are the same. Visibility and quality of image 260 can be improved.

Referring to FIG. 9, in FIG. 9, the image correction module 920 is a second artificial intelligence trained to infer the correction-base image 950 based on the base image 940_2 and a plurality of shift values 960. A configuration including the model is shown. Hereinafter, overlapping descriptions of the same components as those described in FIG. 8 will be omitted.

In one embodiment, the at least one processor 220 corresponds to the base panel 211 by applying the first frame image 930_1 and the second frame image 930_2 to the image generation module 910, and a plurality of A base image 940_2 including sub-base images 941 and a layer image 940_1 corresponding to the layer panel 212 may be generated.

In one embodiment, at least one processor 220 applies a base image 940_2 including a plurality of sub-base images 941 and a plurality of shift values 960 to the image correction module 920, A correction-base image 950 including sub-correction-base images 951 of can be generated.

In one embodiment, the second artificial intelligence model generates the base panel 211 in the first frame based on the base image 940_2 including a plurality of sub-base images 941 and a plurality of shift values 960. It may be an artificial intelligence model trained to infer a correction-base image 950 that corresponds to and includes a plurality of sub-correction-base images 951 respectively corresponding to a plurality of different views.

In one embodiment, the second artificial intelligence model is based on a plurality of sub-base images 941, each corresponding to a plurality of different views, and a plurality of shift values 960, respectively, corresponding to a plurality of different views. An artificial intelligence model trained to infer a correction-base image 950 including a plurality of sub-correction-base images 951 based on differences between each of a plurality of sub-base images 941 according to a plurality of different views. It can be. Hereinafter, the training method of the second artificial intelligence model will be described later with reference to FIG. 12C.

In one embodiment, at least one processor 220 may display the correction-base image 950 on the base panel 211 and display the layer image 940_1 on the layer panel 212.

FIG. 11 is a diagram illustrating a training process of a first artificial intelligence model included in an image generation module and a second artificial intelligence model included in an image correction module, according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same configurations as those described in FIGS. 6, 8, and 9 will be omitted.

Referring to FIGS. 2 and 11 , FIG. 11 shows that the electronic device 100 includes a base panel 211, a layer panel 212, and an optical layer 213 disposed between the base panel 211 and the layer panel 212. This is a diagram for explaining the operation of the electronic device 100 when it includes. At this time, the electronic device 100 may further include a backlight 214 that provides light to the base panel 211 and the layer panel 212.

In one embodiment, the image generation module 1110 included in the electronic device 100 may include a first artificial intelligence model, and the image correction module 1120 may include a second artificial intelligence model.

In one embodiment, the electronic device 100 trains the first artificial intelligence model included in the image generation module 1110 and the second artificial intelligence model included in the image correction module 1120, as shown in FIG. 11. You can. However, the present disclosure is not limited to this, and a surrounding electronic device may train an artificial intelligence model as shown in FIG. 11, and the electronic device 100 may receive the trained artificial intelligence model from the surrounding electronic device. . Additionally, the electronic device 100 may receive a pre-trained artificial intelligence model for transfer learning from a surrounding electronic device and perform transfer learning based on the one shown in FIG. 11 . Hereinafter, for convenience of explanation, a case where the electronic device 100 trains the first artificial intelligence model included in the image generation module 1110 and the second artificial intelligence model included in the image correction module 1120 will be described.

In one embodiment, the electronic device 100 may acquire input content 1130 for training. The electronic device 100 may acquire a first training frame image 1130_1 in the first frame of the training input content 1130 and a second training frame image 1130_2 in the second frame.

In one embodiment, the first artificial intelligence model included in the image generation module 1110 includes a first training frame image 1130_1 including a plurality of training view images 1131 and a plurality of training view images. Based on the second training frame image 1130_2 including 1132, the training base corresponds to the base panel 211 in the first frame and includes a plurality of training sub-base images 1141. A training frame corresponding to the layer panel 212 in the image 1140_2 and the first frame, and reflecting the movement 753 (see FIG. 7) between the first training frame image 1130_1 and the second training frame image 1130_2. It can be trained to infer the layer image 1140_1.

In one embodiment, the at least one processor 220 executes the first training module 1160 to extract the movement 753 between the first training frame image 1130_1 and the second training frame image 1130_2. And, warping the comparison layer image corresponding to the layer panel 212 in the pre-generated second frame based on the extracted motion 753 to create a training layer image with a small difference from the warped comparison layer image. The first artificial intelligence model can be trained to infer (1140_1). Hereinafter, training of the first artificial intelligence model by the first training module 1160 will be described later with reference to FIG. 12A.

In one embodiment, in order to provide the user 250 with an image 260 that reproduces the first training frame image 1130_1, at least one processor 220 executes a second training module 1170, 1 Based on the training frame image 1130_1, a training base image 1140_2 corresponding to the base panel 211 in the first frame and a training layer image 1140_1 corresponding to the layer panel 212 in the first frame. The first artificial intelligence model can be trained to infer. At least one processor 220 executes the second training module 1170 to obtain the training base image 1140_2, the training layer image 1140_1, the brightness information of the base panel 211, and the brightness of the layer panel 212. A training base image 1140_2 and a training layer to reduce the difference between the training output image obtained by simulating the image 260 reproduced by the electronic device 100 according to the information and the first training frame image 1130_1. A first artificial intelligence model may be trained to infer the image 1140_1. Hereinafter, training of the first artificial intelligence model by the second training module 1170 will be described later with reference to FIG. 12B.

In one embodiment, at least one processor 220 executes the third training module 1280, and sets each of a plurality of training sub-base images 1241 corresponding to a plurality of different views to a reference shift value 1190. ) is used to shift, and a second correction-training base image 1250 is inferred, which includes a plurality of sub-correction-training base images 1251 with small differences from a plurality of shifted sub-training base images. Artificial intelligence models can be trained. Hereinafter, training of the second artificial intelligence model by the third training module 1180 will be described later with reference to FIG. 12C.

FIG. 12A is a diagram illustrating a first loss function for training a first artificial intelligence model according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same configurations as those described in FIGS. 6, 8, 9, and 11 will be omitted.

Referring to FIG. 12A, in one embodiment, the first training module 1260 may include a motion extraction module 1261, an image warping module 1262, and a first loss function calculation module 1263.

In one embodiment, the motion extraction module 1261 creates the first training frame image 1230_1 and the second training frame image 1230_2 based on the first training frame image 1230_1 and the second training frame image 1230_2. It may be composed of instructions or program code related to an operation or function for extracting movement 735 (see FIG. 7) between images 1230_2.

In one embodiment, the motion extraction module 1261 generates a second frame of pixels included in the training frame image 1230 based on the first training frame image 1230_1 and the second training frame image 1230_2. An operation or function to obtain the vector component from the first frame and extract the movement 735 of the pixel included in the training frame image 1230 from the second frame to the first frame based on the size and direction of the vector. It may consist of instructions or program code related to. In one embodiment, pixels included in the training frame image 1230 may be included in objects included in the training frame image 1230.

In one embodiment, the motion extraction module 1261 extracts the movement 735 of a pixel included in the training frame image 1230 from the second frame to the first frame through the Lucas-Kanade algorithm, etc. It can be extracted. In addition, the motion extraction module 1261 extracts the movement of pixels included in the training frame image 1230 from the second frame to the first frame through a deep learning model such as FlowNet, LiteFlowNet, etc. (735) can also be extracted and is not limited to any one operation or function.

In one embodiment, the motion extraction module 1261 selects a center view image and a second training frame image obtained from the center view among the plurality of training view images 1231 included in the first training frame image 1230_1. Based on the center view image acquired from the center view among the plurality of training view images 1232 included in (1230_2), the movement between the first training frame image 1230_1 and the second training frame image 1230_2 It may be composed of instructions or program code related to the operation or function of extracting (735). The center view may correspond to a position where the user 250 is located in front of the electronic device 100 and can see the front of the object included in the image 260.

In one embodiment, the at least one processor 220 applies the first training frame image 1230_1 and the second training frame image 1230_2 to the motion extraction module 1261, thereby creating a first training frame image ( Movement 753 between 1230_1) and the second training frame image 1230_2 can be extracted.

In one embodiment, the at least one processor 220 selects a center view image and a second training frame obtained from the center view among the plurality of training view images 1231 included in the first training frame image 1230_1. By applying the center view image obtained from the center view among the plurality of training view images 1232 included in the image 1230_2 to the motion extraction module 1261, the first training frame image 1230_1 and the second training frame image 1230_1 are generated. Movement 753 between the frame images 1230_2 can be extracted.

Hereinafter, for convenience of explanation, the movement 753 between the first training frame image 1230_1 and the second training frame image 1230_2 refers to the plurality of training views included in the first training frame image 1230_1. A motion extraction module extracts a center view image obtained from the center view among the images 1231 and a center view image acquired from the center view among the plurality of training view images 1232 included in the second training frame image 1230_2. (1261) is applied to explain the extracted movement.

In one embodiment, the image warping module 1262 is an operation or function of warping the extracted motion 753 and the comparison layer image 1241 corresponding to the layer panel 212 in the pre-generated second frame. It may consist of instructions or program code related to.

In one embodiment, the comparison layer image 1241 may be a layer image generated by at least one processor 220 executing the image generation module 1210 in the second frame. At least one processor 220 displays a second training frame image 1230_2 in the second frame of the training input content 1230 and a third training frame image 1230_2 in the third frame that is the immediately preceding frame of the training input content 1230. By applying it to the image generation module 1210, a comparison layer image 1241 corresponding to the layer panel 212 can be generated in the second frame.

In one embodiment, the at least one processor 220 generates and generates motion 753 between the first training frame image 1230_1 and the second training frame image 1230_2 extracted through the motion extraction module 1261. By applying one comparison layer image 1241 to the image warping module 1262, the comparison layer image 1241 is adjusted by the movement 753 between the first training frame image 1230_1 and the second training frame image 1230_2. You can warp.

In one embodiment, the comparison layer image warped by the movement 753 between the first training frame image 1230_1 and the second training frame image 1230_2 is compared to the comparison layer image 1241 in the second frame. It may be an image that reflects a change in the position of an object included in the input content 1230 during the first frame.

In one embodiment, the first loss function calculation module 1263 generates a training layer image corresponding to the layer panel 212 in the first frame generated by the at least one processor 220 executing the image generation module 1210. It may be composed of instructions or program code related to an operation or function of calculating the first loss function according to (1240_1) and the warped comparison layer image.

In one embodiment, the first loss function calculation module 1263 may calculate a first loss function according to Equation 1 below according to the training layer image 1240_1 and the warped comparison layer image. there is.

Equation 1:

은 제1 손실 함수이고,

은 제1 마스킹 값이고,

는 제1 프레임에서의 훈련용 레이어 영상이고,

는 와핑된 제2 프레임에서의 비교 레이어 영상이며,

는

와

의 차이의 L1 노름(norm)이다.At this time,

is the first loss function,

is the first masking value,

is the training layer image in the first frame,

is the comparison layer image in the warped second frame,

Is

and

This is the L1 norm of the difference.

In one embodiment, the first masking value changes the comparison layer image 1241 by the amount of movement 753 between the first training frame image 1230_1 and the second training frame image 1230_2 through the image warping module 1262. In the warping operation, the comparison layer image whose warping is not equal to the movement 753 between the first training frame image 1230_1 and the second training frame image 1230_2 is prevented from being reflected in the value of the first loss function. You can. The first loss function calculation module 1263 may calculate the first masking value according to Equation 2 below.

Equation 2:

은 제1 마스킹 값이고,

는 임의의 계수이며,

는 제1 프레임에서의 훈련용 레이어 영상이고,

는 와핑된 제2 프레임에서의 비교 레이어 영상이며,

는

와

의 차이의 L2 노름(norm)이다.At this time,

is the first masking value,

is an arbitrary coefficient,

is the training layer image in the first frame,

is the comparison layer image in the warped second frame,

Is

and

This is the L2 norm of the difference.

와

의 차이가 기-설정된 차이값보다 클 경우, 비교 레이어 영상(1241)의 와핑이 제1 훈련용 프레임 영상(1230_1)과 제2 훈련용 프레임 영상(1230_2) 간의 움직임(753)만큼 되지 않았다고 판단하여, 마스킹 값이 0에 가까운 값을 갖도록 할 수 있다. 이 경우, 제1 손실 함수의 값도 0에 가까운 값이 되어, 해당 훈련 과정은 제1 인공 지능 모델의 훈련에 반영되지 않을 수 있다. In one embodiment, the at least one processor 220, through the first loss function calculation module 1263,

and

If the difference is greater than the preset difference value, it is determined that the warping of the comparison layer image 1241 is not as much as the movement 753 between the first training frame image 1230_1 and the second training frame image 1230_2. , the masking value can be set to a value close to 0. In this case, the value of the first loss function is also close to 0, so the training process may not be reflected in the training of the first artificial intelligence model.

와

의 차이가 기-설정된 차이값과 같거나 작을 경우, 비교 레이어 영상(1241)의 와핑이 제1 훈련용 프레임 영상(1230_1)과 제2 훈련용 프레임 영상(1230_2) 간의 움직임(753)만큼 되었다고 판단하여, 마스킹 값이 1에 가까운 값을 갖도록 할 수 있다. 이 경우, 해당 훈련 과정은 제1 인공 지능 모델의 훈련에 반영될 수 있다. On the other hand, at least one processor 220 operates through the first loss function calculation module 1263.

and

If the difference is equal to or smaller than the preset difference value, it is determined that the warping of the comparison layer image 1241 is equal to the movement 753 between the first training frame image 1230_1 and the second training frame image 1230_2. Thus, the masking value can be made to have a value close to 1. In this case, the training process may be reflected in the training of the first artificial intelligence model.

의 크기를 조절하여 결정할 수 있다.

의 크기를 크게 할수록 기-설정된 차이값의 크기가 작아질 수 있다.At this time, the size of the preset difference value is

It can be determined by adjusting the size of .

As the size of , the size of the preset difference value may become smaller.

However, the present disclosure is not limited thereto, and the first loss function calculation module 1260 may calculate the first loss function and the first masking value using other equations.

In one embodiment, the value of the first loss function may become smaller as the difference between the training layer image 1240_1 and the warped comparison layer image decreases. At least one processor 220 may train the first artificial intelligence model included in the image generation module 1210 through the first training module 1260 so that the value of the first loss function becomes smaller. At least one processor 220 infers a training layer image 1240_1 that can obtain a small first loss function value based on the first training frame image 1230_1 and the second training frame image 1230_2. The first artificial intelligence model can be trained to do this. Weights included in the first artificial intelligence model may be updated so that the value of the first loss function becomes smaller.

FIG. 12B is a diagram illustrating a second loss function for training a first artificial intelligence model according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same configuration as that described in FIG. 12A will be omitted.

Referring to FIG. 12B , in one embodiment, the second training module 1270 may include a simulation module 1271 and a second loss function calculation module 1272.

In one embodiment, the simulation module 1271 displays the training base image 1240_2 generated by the at least one processor 220 through the image generation module 1210 on the base panel 211, and displays the training layer image 1240_2 on the base panel 211. When (1240_1) is displayed on the layer panel 212, it relates to an operation or function of performing simulation to obtain an output image for training, which is the image 260 provided to the user 250 in the first frame. It may consist of instructions or program code.

In one embodiment, the simulation module 1271 includes a training base image 1240_2, a training layer image 1240_1, and panel information including information about the base panel 211 and information about the layer panel 212 ( 1273), it may be composed of instructions or program code related to an operation or function that performs a simulation to obtain an output image for training.

In one embodiment, the panel information 1273 includes the display type of the base panel 211, the display type of the layer panel 212, the brightness information of the base panel 211, the brightness information of the layer panel 212, and the base panel ( Information about the gap between 211) and the layer panel 212 may be included.

In one embodiment, the panel information 1273 includes information that the brightness of the layer panel 212 when displaying an image of the same gray level is m times greater than the brightness of the base panel 211 when displaying an image of the same gray level. When included, at least one processor 220 displays the training base image 1240_2 on the base panel 211 by reflecting the corresponding brightness information through the simulation module 1271, and displays the training base image 1240_2 on the layer panel 212. When the dragon layer image 1240_1 is displayed, a simulation may be performed to obtain an output image for training that is visible to the user 250. Through this, when at least one processor 220 trains the first artificial intelligence model through the second training module 1270, which will be described later, the brightness of the training base image 1240_2 is equal to that of the training layer image 1240_1. The first artificial intelligence model can be trained to infer that the brightness is m times greater.

In addition, if the panel information 1273 includes information that the display type of at least one panel of the base panel 211 or the layer panel 212 is a display made of stacked mono-color panels, at least One processor 220, through the simulation module 1271, renders each of the stacked single-color panels into a panel containing sub-color pixels such as R, G, and B, and provides training to the base panel 211. When the base image 1240_2 is displayed and the training layer image 1240_1 is displayed on the layer panel 212, a simulation may be performed to obtain an output image for training that is visible to the user 250.

In one embodiment, the second loss function calculation module 1272 generates a second loss function according to the training output image and the first training frame image 1230_1 acquired by the at least one processor 220 through the simulation module 1271. It may consist of instructions or program code related to the operation or function of calculating the loss function.

In one embodiment, the second loss function calculation module 1272 is included in the first training frame image 1230_1 and a plurality of sub-training output images, each corresponding to a plurality of different views included in the training output image. It may be composed of instructions or program code related to an operation or function of calculating a second loss function based on differences between the plurality of training view images 1231.

In one embodiment, the value of the second loss function is a simulation operation based on the training base image 1240_2, the training layer image 1240_1, and the panel information 1273 generated through the image generation module 1210. The smaller the difference between the training output image obtained through execution and the first training frame image 1230_1, the smaller it can be.

In one embodiment, at least one processor 220 may train the first artificial intelligence model included in the image generation module 1210 to reduce the value of the second loss function through the second training module 1270. . At least one processor 220 is configured to infer a training base image 1230_1 and a training layer image 1240_1 from which a small second loss function value can be obtained based on the first training frame image 1230_1. A first artificial intelligence model can be trained. Weights included in the first artificial intelligence model may be updated so that the value of the second loss function becomes smaller.

FIG. 12C is a diagram illustrating a third loss function for training a second artificial intelligence model according to an embodiment of the present disclosure. Hereinafter, overlapping descriptions of the same components as those described in FIGS. 12A and 12B will be omitted.

Referring to FIG. 12C, in one embodiment, the third training module 1280 may include an image shift module 1281 and a third loss function calculation module 1282.

In one embodiment, the image shift module 1281 is a training base image including a plurality of sub-training base images 1241, each corresponding to a plurality of different views generated through the image generation module 1210. Based on 1240_2) and the reference shift value 1290, it may be composed of commands or program codes related to an operation or function of shifting each of the plurality of sub-training base images 1241 using the reference shift value.

개일 때, 사용자(250)의 위치가 전자 장치(100)의 정면에서 영상(260)을 볼 수 있는 뷰에 있는 것을 기준으로 전자 장치(100)의 좌측면으로

개의 뷰, 전자 장치(100)의 우측면으로

개의 뷰가 포함될 수 있다. 도 10을 참조할 때, 기준 쉬프트 값(1290)은 인접한 두 개의 뷰 각각에 대응되는 레이어 영역의 픽셀 간의 간격에 대응될 수 있다.In one embodiment, a plurality of different views

When open, the user 250 is positioned on the left side of the electronic device 100 based on the view where the image 260 can be viewed from the front of the electronic device 100.

View of the left side of the electronic device 100

Views may be included. Referring to FIG. 10, the reference shift value 1290 may correspond to the gap between pixels in the layer area corresponding to each of two adjacent views.

In one embodiment, the image shift module 1281 performs an operation or function of shifting each of the plurality of sub-training images 1241 corresponding to a plurality of different views using a reference shift value to correspond to an adjacent view. It may include related instructions or program code.

In one embodiment, each of the sub-training images corresponding to two adjacent views may include location information of different objects as they correspond to different views. The image shift module 1281 selects sub-training images corresponding to each of the two adjacent views so that the position information of the object in the shifted sub-training image is the same as the position information of the object in the sub-training image before the shift in the adjacent view. It may include instructions or program code related to an operation or function of shifting using a reference shift value.

In one embodiment, when one view among a plurality of different views is referred to as a first view and a view adjacent to the first view is referred to as a second view, the image shift module 1281 generates a plurality of base images for sub-training. The sub-training base image corresponding to the second view among fields 1241 may be shifted to correspond to the first view using a reference shift value. The position information of the object included in the sub-training base image corresponding to the shifted second view may be the same as the position information of the object included in the sub-training base image corresponding to the first view. In addition, information such as luminance, grayscale, color, etc. of the object included in the sub-training base image corresponding to the shifted second view includes the luminance of the object included in the sub-training base image corresponding to the second view before shifting, It may be the same as information such as gradation and color.

In one embodiment, the third loss function calculation module 1282 includes the correction-training base image 1250 and the shifted training base image generated by the at least one processor 220 executing the image correction module 1220. It may be composed of instructions or program code related to the operation or function of calculating the third loss function according to .

In one embodiment, the third loss function calculation module 1282 performs correction-sub correction corresponding to the first view among the plurality of sub-correction-training base images 1251 included in the training base image 1250- An operation or function for calculating a third loss function according to a sub-training base image corresponding to a second view among a plurality of shifted sub-training base images included in the training base image and the shifted training base image. It may consist of instructions or program code.

개의 뷰 중, 어느 하나의 제1 뷰에 대응되는 서브 보정-훈련용 베이스 영상과 제2 뷰에 대응되는 서브 훈련용 베이스 영상에 따라 제1 뷰에 대응되는 손실 값을 구하고, 해당 과정을

개의 뷰에 대하여 반복하는 아래의 수학식 3에 따라 제3 손실 함수를 계산할 수 있다.In one embodiment, the second loss function calculation module 1263 calculates Equation 3 below according to the sub-correction-training base image corresponding to the first view and the sub-training base image corresponding to the second view. Accordingly, the third loss function can be calculated. The second loss function calculation module 1263 is,

Among the views, the loss value corresponding to the first view is obtained according to the sub-correction-training base image corresponding to the first view and the sub-training base image corresponding to the second view, and the corresponding process is performed.

The third loss function can be calculated according to Equation 3 below, which is repeated for the views.

Equation 3:

은 제3 손실 함수이고,

은 제2 마스킹 값이고,

는 n번째 뷰에 대응되는 서브 보정-훈련용 베이스 영상이고,

는 쉬프트된 n-1번째 뷰에서의 서브 훈련용 베이스 영상이며,

는

와

의 차이의 L1 노름(norm)이다.At this time,

is the third loss function,

is the second masking value,

is the sub-correction-training base image corresponding to the nth view,

is the base image for sub-training in the shifted n-1th view,

Is

and

This is the L1 norm of the difference.

In one embodiment, the second masking value is an operation of shifting the sub-training base image corresponding to the n-1th view to correspond to the nth view using the reference shift value 1290 through the image shift module 1281. In this case, the sub-training base image whose shift is not equal to the reference shift value 1290 can be prevented from being reflected in the value of the third loss function. The third loss function calculation module 1282 may calculate the second masking value according to Equation 4 below.

Equation 4:

은 제2 마스킹 값이고,

는 임의의 계수이며,

는 n번째 뷰에 대응되는 서브 보정-훈련용 베이스 영상이고,

는 쉬프트된 n-1번째 뷰에서의 서브 훈련용 베이스 영상이며,

는

와

의 차이의 L2 노름(norm)이다.At this time,

is the second masking value,

is an arbitrary coefficient,

is the sub-correction-training base image corresponding to the nth view,

is the base image for sub-training in the shifted n-1th view,

Is

and

This is the L2 norm of the difference.

와

의 차이가 기-설정된 차이값보다 클 경우, 영상 쉬프트 모듈(1281)의 와핑이 기준 쉬프트 값(1290)만큼 되지 않았다고 판단하여,

개의 뷰 중 n번째 뷰에 대한 마스킹 값이 0에 가까운 값을 갖도록 할 수 있다. 이 경우, 제3 손실 함수의 값 중 n번째 뷰의 성분이 0에 가까운 값이 되어, 해당 훈련 과정은 제2 인공 지능 모델의 훈련에 반영되지 않을 수 있다. In one embodiment, the at least one processor 220, through the third loss function calculation module 1282,

and

If the difference is greater than the preset difference value, it is determined that the warping of the image shift module 1281 is not as high as the reference shift value 1290,

The masking value for the nth view among the views can be set to a value close to 0. In this case, the component of the nth view among the values of the third loss function becomes a value close to 0, so the training process may not be reflected in the training of the second artificial intelligence model.

와

의 차이가 기-설정된 차이값과 같거나 작을 경우, 영상 쉬프트 모듈(1281)에 의한 쉬프트 동작이 기준 쉬프트 값(1290)만큼 되었다고 판단하여,

개의 뷰 중 n번째 뷰에 대한 마스킹 값이 1에 가까운 값을 갖도록 할 수 있다. 이 경우, n번째 뷰에 의한 훈련 과정은 제2 인공 지능 모델의 훈련에 반영될 수 있다. On the other hand, at least one processor 220 operates through the third loss function calculation module 1282.

and

If the difference is equal to or smaller than the preset difference value, it is determined that the shift operation by the image shift module 1281 has reached the reference shift value 1290,

The masking value for the nth view among the views can be set to a value close to 1. In this case, the training process using the nth view may be reflected in the training of the second artificial intelligence model.

의 크기를 조절하여 결정할 수 있다.

의 크기를 크게 할수록 기-설정된 차이값의 크기가 작아질 수 있다.At this time, the size of the preset difference value is

It can be determined by adjusting the size of .

As the size of , the size of the preset difference value may become smaller.

However, the present disclosure is not limited thereto, and the third loss function calculation module 1282 may calculate the third loss function and the second masking value using other equations.

In one embodiment, the value of the third loss function may become smaller as the difference between a plurality of sub-correction-training base images corresponding to a plurality of different views and a plurality of shifted sub-training base images decreases. . At least one processor 220 may train the third artificial intelligence model included in the image correction module 1220 to reduce the value of the third loss function through the third training module 1280. At least one processor 220 is configured to obtain a small third loss function value based on a training base image 1240_2 including a plurality of sub-training base images 1241 for sub-correction-training. A third artificial intelligence model may be trained to infer the correction-training base image 1250 including the base images 1251. The weights included in the third artificial intelligence model may be updated so that the value of the third loss function becomes smaller.

In order to solve the above-described technical problem, in one embodiment, the electronic device 100 includes a base panel 211, a layer panel 212 disposed on the base panel 211, and stores at least one instruction. It may include a memory 230 that executes and at least one processor 220 that executes at least one instruction stored in the memory 230. At least one processor 220 may acquire a first frame image in the first frame of the input content and a second frame image in the second frame, which is the frame immediately before the first frame, by executing at least one command. . At least one processor 220 applies the first frame image and the second frame image to the image generation module, thereby creating a base image corresponding to the base panel in the first frame and a layer panel in the first frame, and A layer image can be created based on the movement between the image and the second frame image. At least one processor 220 may generate a correction-base image based on the movement between the first frame image and the second frame image by applying the base image or layer image to the image correction module. At least one processor 220 may display the correction-base image on the base panel and display the layer image on the layer panel.

In one embodiment, the electronic device 100 may further include an optical layer 213 disposed between the base panel 211 and the layer panel 212. The base image may include a plurality of sub-base images, each corresponding to a plurality of different views. At least one processor 220 applies the base image or layer image to the image correction module, based on the difference between a plurality of sub-base images according to a plurality of different views, and including a plurality of sub-correction-base images. A correction-base image can be created.

In one embodiment, the image generation module may include a first artificial intelligence model trained to infer a base image and a layer image based on the first frame image and the second frame image. The first artificial intelligence model may acquire a first training frame image in the first frame of the training input content and a second training frame image in the second frame, which is the frame immediately before the first frame. The first artificial intelligence model can extract movement between the first training frame image and the second training frame image. The first artificial intelligence model may use the extracted motion to warp the comparison layer image corresponding to the layer panel in the pre-generated second frame. The first artificial intelligence model is an artificial intelligence model trained based on a first loss function according to the training layer image corresponding to the layer panel and the warped comparison layer image in the first frame generated through the first artificial intelligence model. You can.

In one embodiment, each of the first training frame image and the second training frame image may include a plurality of training view images obtained from a plurality of different views. The movement between the first training frame image and the second training frame image is determined by the center view image and the second training frame image obtained from the center view among the plurality of training view images included in the first training frame image. It can be extracted from a center view image obtained from a center view among the included plurality of training view images.

In one embodiment, the first loss function may be calculated based on the difference between the training layer image and the warped comparison layer image. The first artificial intelligence model may be trained to reduce the difference between the training layer image and the warped comparison layer image based on the first loss function.

In one embodiment, the first artificial intelligence model applies the first training frame image and the second training frame image to the first artificial intelligence model, thereby generating a training base image corresponding to the base panel in the first frame, and a second training frame image. You can create a training layer image corresponding to the layer panel in 1 frame. The first artificial intelligence model may obtain an output image for training in the first frame based on the training base image, training layer image, brightness information of the base panel, and brightness information of the layer panel. The first artificial intelligence model may be an artificial intelligence model trained based on a second loss function according to the difference between the first training frame image and the output image for training.

In one embodiment, the at least one processor 220 may generate a correction-base image by executing at least one instruction and applying the first frame image and the layer image to the image correction module.

In one embodiment, the first frame image may include multiple view images obtained from multiple different views. At least one processor 220 executes at least one instruction to apply a plurality of shift values corresponding to each of the first frame image, the layer image, and the plurality of views to the image correction module to generate a correction-base image. You can. The image correction module may include a command that performs an operation or function of shifting the layer image using each of a plurality of shift values. The image correction module includes instructions for performing an operation or function of generating a correction-base image using the differences between each of the plurality of view images and the plurality of shifted layer images corresponding to the plurality of view images. can do. The plurality of shift values may be determined based on at least one of the spacing between the base panel 211 and the layer panel 212, the resolution of the layer panel 212, or a plurality of different views.

In one embodiment, the image correction module may include a second artificial intelligence model trained to infer a correction-base image based on the base image. The second artificial intelligence model may acquire a training base image and a reference shift value including a plurality of sub-training base images corresponding to a plurality of different views generated through the image generation module. The second artificial intelligence model may shift each of the plurality of sub-training base images using a reference shift value. The second artificial intelligence model may be an artificial intelligence model trained based on a third loss function according to the corrected-training base image and the shifted training base image generated through the second artificial intelligence model.

In one embodiment, the third loss function is a sub-correction-training base image corresponding to a first view among a plurality of sub-correction-training base images included in the correction-training base image and the shifted training It may be calculated according to the difference between the shifted sub-training base image corresponding to the second view adjacent to the first view among the plurality of shifted sub-training base images included in the base image. The second artificial intelligence model may be trained to reduce the difference between the sub-correction-training base image and the shifted sub-training base image based on the third loss function.

In order to solve the above-described technical problem, in one embodiment, a method of operating an electronic device 100 including a base panel 211 and a layer panel 212 disposed on the base panel 211 is provided. A method of operating the electronic device 100 according to an embodiment may include acquiring a first frame image in a first frame of input content and a second frame image in a second frame that is the frame immediately preceding the first frame. You can. The operating method of the electronic device 100 is to apply the first frame image and the second frame image to the image generation module, thereby generating the base image corresponding to the base panel 211 in the first frame and the layer panel 212 in the first frame. It corresponds to and may include generating a layer image based on movement between the first frame image and the second frame image. The method of operating the electronic device 100 may include generating a correction-base image based on movement between a first frame image and a second frame image by applying a base image or a layer image to an image correction module. The method of operating the electronic device 100 may include displaying a correction-base image on the base panel 211 and displaying a layer image on the layer panel 212.

In one embodiment, the electronic device 100 may further include an optical layer 213 disposed between the base panel 211 and the layer panel 212. The base image may include a plurality of sub-base images, each corresponding to a plurality of different views. In the step of generating a correction-base image, it is based on the difference between a plurality of sub-base images according to a plurality of different views by applying the base image or layer image to the image correction module, and includes a plurality of sub-correction-base images. It may include generating a correction-base image.

In one embodiment, the image generation module may include a first artificial intelligence model trained to infer a base image and a layer image based on the first frame image and the second frame image. The training method of the first artificial intelligence model includes acquiring a first training frame image in a first frame of input content for training and a second training frame image in a second frame that is the frame immediately before the first frame. can do. The training method of the first artificial intelligence model may include extracting movement between the first training frame image and the second training frame image. The training method of the first artificial intelligence model may include warping the comparison layer image corresponding to the layer panel 212 in the pre-generated second frame using the extracted motion. The training method of the first artificial intelligence model is based on the first loss function according to the training layer image corresponding to the layer panel 212 in the first frame generated through the first artificial intelligence model and the warped comparison layer image. , may include training a first artificial intelligence model.

In one embodiment, each of the first training frame image and the second training frame image may include a plurality of training view images obtained from a plurality of different views. In the step of extracting the movement between the first training frame image and the second training frame image, the movement between the first training frame image and the second training frame image is generated from the plurality of training frame images included in the first training frame image. It may be extracted from a center view image acquired from a center view among view images and a center view image acquired from a center view among a plurality of training view images included in the second training frame image.

In one embodiment, the training method of the first artificial intelligence model includes applying the first training frame image and the second training frame image to the first artificial intelligence model, thereby training corresponding to the base panel 211 in the first frame. It may include generating a training layer image corresponding to the layer panel 212 from the base image and the first frame. The training method of the first artificial intelligence model acquires an output image for training in the first frame based on the training base image, training layer image, brightness information of the base panel 211, and brightness information of the layer panel 212. It may include steps. The training method of the first artificial intelligence model may include training the first artificial intelligence model based on a second loss function based on the difference between the first training frame image and the output image for training.

In one embodiment, in the step of generating a correction-base image, the correction-base image may be generated by applying the first frame image and the layer image to the image correction module.

In one embodiment, the first frame image may include multiple view images obtained from multiple different views. In the step of generating a correction-base image, the correction-base image may be generated by applying a plurality of shift values corresponding to the first frame image, the layer image, and the plurality of views to the image correction module. The image correction module may include a command that performs an operation or function of shifting the layer image using each of a plurality of shift values. The image correction module may include instructions for performing an operation or function of generating a correction-base image using each of the plurality of view images and a plurality of shifted layer images corresponding to the plurality of view images. . The plurality of shift values may be determined based on at least one of the spacing between the base panel 211 and the layer panel 212, the resolution of the layer panel 212, or a plurality of different views.

In one embodiment, the image correction module may include a second artificial intelligence model trained to infer a correction-base image based on the base image. The training method of the second artificial intelligence model includes acquiring a training base image and a reference shift value including a plurality of sub-training base images corresponding to a plurality of different views generated through an image generation module. can do. The training method of the second artificial intelligence model may include shifting each of a plurality of base images for training using a reference shift value. The training method of the second artificial intelligence model includes training the second artificial intelligence model based on a third loss function according to the corrected-training base image and the shifted training base image generated through the second artificial intelligence model. may include.

In one embodiment, the step of training the second artificial intelligence model based on the third loss function includes selecting a first view among a plurality of sub-correction-training base images included in the correction-training base image. The difference between the corresponding sub correction-training base image and the shifted sub-training base image corresponding to the second view adjacent to the first view among the plurality of shifted sub-training base images included in the shifted training base image. It may include calculating a third loss function according to . The step of training the second artificial intelligence model based on the third loss function includes training the second artificial intelligence model to reduce the difference between the sub-correction-training base image and the shifted sub-training base image based on the third loss function. It may include training.

In one embodiment of the present disclosure, a computer-readable recording medium on which a program for performing at least one method among the embodiments of the disclosed method is recorded on a computer may be provided.

Programs executed by the electronic device described in this disclosure may be implemented with hardware components, software components, and/or a combination of hardware components and software components. A program can be executed by any system that can execute computer-readable instructions.

Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device.

Software may be implemented as a computer program including instructions stored on computer-readable storage media. Computer-readable recording media include, for example, magnetic storage media (e.g., read-only memory (ROM), random-access memory (RAM), floppy disk, hard disk, etc.) and optical read media (e.g., CD-ROM). (CD-ROM), DVD (Digital Versatile Disc), etc. The computer-readable recording medium is distributed among computer systems connected to a network, so that computer-readable code can be stored and executed in a distributed manner. The recording medium can be read by a computer, stored in memory, and executed by a processor.

Computer-readable storage media may be provided in the form of non-transitory storage media. Here, 'non-transitory storage medium' only means that it is a tangible device and does not contain signals (e.g. electromagnetic waves). This term refers to cases where data is semi-permanently stored in a storage medium and temporary storage media. It does not distinguish between cases where it is stored as . For example, a 'non-transitory storage medium' may include a buffer where data is temporarily stored.

Additionally, programs according to embodiments disclosed in this specification may be included and provided in a computer program product. Computer program products are commodities and can be traded between sellers and buyers.

A computer program product may include a software program and a computer-readable storage medium on which the software program is stored. For example, a computer program product includes a product in the form of a software program (e.g., a downloadable application) that is distributed electronically by the manufacturer of the electronic device or through an electronic marketplace (e.g., Samsung Galaxy Store). can do. For electronic distribution, at least a portion of the software program may be stored on a storage medium or created temporarily. In this case, the storage medium may be a storage medium of a server of an electronic device manufacturer, a server of an electronic market, or a relay server that temporarily stores a software program.

As described above, although the embodiments have been described with limited examples and drawings, various modifications and variations can be made by those skilled in the art from the above description. For example, the described techniques may be performed in a different order than the described method, and/or components, such as a described computer system or module, may be combined or combined in a form different from the described method, or other components or equivalents may be used. Appropriate results can be achieved even if replaced or replaced by .

Claims (20)

base panel 211;
a layer panel 212 disposed on the base panel 211;
a memory 230 that stores at least one instruction; and
At least one processor 220 that executes the at least one instruction stored in the memory 230,
The at least one processor 220 executes the at least one instruction,
Obtaining a first frame image in a first frame of input content and a second frame image in a second frame that is the frame immediately preceding the first frame,
By applying the first frame image and the second frame image to the image generation module, the base image corresponding to the base panel in the first frame and the layer panel in the first frame, and the first frame image and generate a layer image based on the movement between the second frame image,
Applying the base image or the layer image to an image correction module to generate a correction-base image based on movement between the first frame image and the second frame image,
An electronic device (100) that displays the correction-base image on the base panel and displays the layer image on the layer panel. According to claim 1,
The electronic device 100 further includes an optical layer 213 disposed between the base panel 211 and the layer panel 212,
The base image includes a plurality of sub-base images each corresponding to a plurality of different views,
The at least one processor 220,
By applying the base image or the layer image to the image correction module, the correction is based on a difference between the plurality of sub-base images according to the plurality of different views and includes a plurality of sub-base images. Electronic device 100 for generating a base image. According to any one of claims 1 or 2,
The image generation module includes a first artificial intelligence model trained to infer the base image and the layer image based on the first frame image and the second frame image,
The first artificial intelligence model is,
Obtaining a first training frame image in a first frame of training input content and a second training frame image in a second frame that is the frame immediately preceding the first frame,
Extracting motion between the first training frame image and the second training frame image,
Using the extracted motion, warping a comparison layer image corresponding to the layer panel in the pre-generated second frame,
An electronic device that is an artificial intelligence model trained based on a first loss function according to the training layer image corresponding to the layer panel and the warped comparison layer image in the first frame generated through the first artificial intelligence model ( 100). According to clause 3,
Each of the first training frame image and the second training frame image includes a plurality of training view images obtained from a plurality of different views,
The movement between the first training frame image and the second training frame image includes a center view image obtained from a center view among the plurality of training view images included in the first training frame image and the second training frame image. An electronic device 100 that is extracted from a center view image acquired from a center view among the plurality of training view images included in a training frame image. According to any one of claims 3 or 4,
The first loss function is calculated according to the difference between the training layer image and the warped comparison layer image,
The first artificial intelligence model is trained to reduce the difference between the training layer image and the warped comparison layer image based on the first loss function. According to any one of claims 3 to 5,
The first artificial intelligence model is,
By applying the first training frame image and the second training frame image to the first artificial intelligence model, a training base image corresponding to the base panel in the first frame and the layer in the first frame are generated. Generate the training layer image corresponding to the panel,
Obtaining an output image for training in the first frame based on the training base image, the training layer image, brightness information of the base panel, and brightness information of the layer panel,
An electronic device 100 that is an artificial intelligence model trained based on a second loss function based on the difference between the first training frame image and the training output image. According to any one of claims 1 or 2,
The at least one processor 220 executes the at least one instruction,
An electronic device (100) that generates the correction-base image by applying the first frame image and the layer image to the image correction module. According to clause 2,
The first frame image includes a plurality of view images obtained from the plurality of different views,
The at least one processor 220 applies a plurality of shift values corresponding to the first frame image, the layer image, and the plurality of views to the image correction module by executing the at least one instruction, Generating the correction-base image,
The image correction module is,
Shifting the layer image using each of the plurality of shift values,
A command for performing an operation or function of generating the correction-base image using each of the plurality of view images and a plurality of shifted layer images corresponding to the plurality of view images, respectively,
The plurality of shift values are determined based on at least one of the interval between the base panel 211 and the layer panel 212, the resolution of the layer panel 212, or the plurality of different views. ). According to clause 2,
The image correction module includes a second artificial intelligence model trained to infer the correction-base image based on the base image and a reference shift value,
The second artificial intelligence model is,
Obtaining a training base image including a plurality of sub-training base images corresponding to the plurality of different views, which are generated through the image generation module, and the reference shift value,
Shifting each of the plurality of sub-training base images using the reference shift value,
An artificial intelligence model trained based on a correction-training base image generated through the second artificial intelligence model and a third loss function according to the shifted training base image,
The reference shift value is determined based on at least one of the distance between the base panel 211 and the layer panel 212 or the resolution of the layer panel 212. According to clause 9,
The third loss function is a sub-correction-training base image corresponding to a first view among a plurality of sub-correction-training base images included in the correction-training base image and the shifted training base image. Calculated according to the difference between a sub-training base image obtained by shifting a sub-training base image corresponding to a second view adjacent to the first view among a plurality of shifted sub-training base images included in the base image,
The second artificial intelligence model is trained to reduce the difference between the sub-correction-training base image and the shifted sub-training base image based on the third loss function. In a method of operating an electronic device 100 including a base panel 211 and a layer panel 212 disposed on the base panel 211,
Obtaining a first frame image in a first frame of input content and a second frame image in a second frame that is the frame immediately preceding the first frame;
By applying the first frame image and the second frame image to the image generation module, the base image corresponding to the base panel 211 in the first frame and the layer panel 212 in the first frame are , generating a layer image based on movement between the first frame image and the second frame image;
generating a correction-base image based on movement between the first frame image and the second frame image by applying the base image or the layer image to an image correction module; and
A method of operating an electronic device (100) including displaying the correction-base image on the base panel (211) and displaying the layer image on the layer panel (212). According to claim 11,
The electronic device 100 further includes an optical layer 213 disposed between the base panel 211 and the layer panel 212,
The base image includes a plurality of sub-base images each corresponding to a plurality of different views,
In the step of generating the correction-base image,
By applying the base image or the layer image to the image correction module, the correction is based on a difference between the plurality of sub-base images according to the plurality of different views and includes a plurality of sub-base images. Method of operating an electronic device 100 that generates a base image. According to any one of claims 11 or 12,
The image generation module includes a first artificial intelligence model trained to infer the base image and the layer image based on the first frame image and the second frame image,
The first artificial intelligence model is,
Obtaining a first training frame image in a first frame of input content for training and a second training frame image in a second frame that is the frame immediately preceding the first frame;
extracting motion between the first training frame image and the second training frame image;
warping a comparison layer image corresponding to the layer panel 212 in the pre-generated second frame using the extracted motion; and
Based on a first loss function according to the training layer image corresponding to the layer panel 212 in the first frame generated through the first artificial intelligence model and the warped comparison layer image, the first artificial intelligence model A method of operating an electronic device 100, which is an artificial intelligence model trained through the step of training an intelligence model. According to claim 13,
Each of the first training frame image and the second training frame image includes a plurality of training view images obtained from a plurality of different views,
In the step of extracting movement between the first training frame image and the second training frame image,
The movement is a center view image obtained from a center view among the plurality of training view images included in the first training frame image and a center view image among the plurality of training view images included in the second training frame image. A method of operating an electronic device 100 extracted from a center view image acquired in a center view. According to any one of claims 13 or 14,
The first artificial intelligence model is,
By applying the first training frame image and the second training frame image to the first artificial intelligence model, the training base image corresponding to the base panel 211 in the first frame and the first frame generating the training layer image corresponding to the layer panel 212;
Obtaining an output image for training in the first frame based on the training base image, the training layer image, brightness information of the base panel 211, and brightness information of the layer panel 212; and
A method of operating the electronic device 100, which is an artificial intelligence model trained through training the first artificial intelligence model based on a second loss function according to the difference between the first training frame image and the training output image. . According to any one of claims 11 or 12,
In the step of generating the correction-base image,
A method of operating the electronic device 100 for generating the correction-base image by applying the first frame image and the layer image to the image correction module. According to claim 12,
The first frame image includes a plurality of view images obtained from the plurality of different views,
In the step of generating the correction-base image, the correction-base image is generated by applying a plurality of shift values corresponding to each of the first frame image, the layer image, and the plurality of views to the image correction module; ,
The image correction module is,
Shifting the layer image using each of the plurality of shift values,
A command for performing an operation or function of generating the correction-base image using each of the plurality of view images and a plurality of shifted layer images corresponding to the plurality of view images, respectively,
The plurality of shift values are determined based on at least one of the interval between the base panel 211 and the layer panel 212, the resolution of the layer panel 212, or the plurality of different views. ) operation method. According to claim 12,
The image correction module includes a second artificial intelligence model trained to infer the correction-base image based on the base image,
The second artificial intelligence model is,
Obtaining a base image for training including a plurality of sub-training base images corresponding to each of the plurality of different views and a reference shift value generated through the image generation module;
Shifting each of the plurality of training base images using the reference shift value; and
An artificial intelligence model trained through the step of training the second artificial intelligence model based on a third loss function according to the correction-training base image and the shifted training base image generated through the second artificial intelligence model. Method of operating the electronic device 100. According to clause 18,
The step of training the second artificial intelligence model based on the third loss function is,
A sub-correction-training base image corresponding to a first view among a plurality of sub-correction-training base images included in the correction-training base image and a plurality of sub-correction-training base images included in the shifted training base image calculating the third loss function according to a difference between a shifted sub-training base image corresponding to a second view adjacent to the first view among the shifted sub-training base images; and
Based on the third loss function, training the second artificial intelligence model to reduce the difference between the sub-correction-training base image and the shifted sub-training base image. How it works. A computer-readable recording medium on which a program for performing the method of any one of claims 11 to 19 is recorded on a computer.

Images