Apparatus and method for photo output using text prompt and artificial intelligence template
The electronic device integrates AI templates and neural networks to generate high-quality, customized images with preserved facial features and optimal print quality, addressing the limitations of conventional photo output devices in integrating art styles and environmental adaptability.
Patent Information
- Application Number
- KR1020260116775
- Authority / Receiving Office
- KR · KR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2026-06-25
- Publication Date
- 2026-11-06
- Estimated Expiration
- 2046-06-25
AI Technical Summary
Conventional photo output devices struggle to naturally incorporate a specific art style or detailed concept desired by the user without damaging the original photograph, and existing Generative AI technology is not integrated effectively for real-time image generation in offline instant photo printing devices to preserve unique facial features.
An electronic device comprising an input device, camera, printing device, memory, and processor that uses AI templates and image generation neural networks to generate high-quality, customized images by extracting facial features, applying AI templates, and optimizing placement and environmental conditions for print output.
Generates high-quality, customized images that preserve unique facial features and natural integration of art styles, while ensuring visual harmony and optimal print quality under varying environmental conditions.
Smart Images

Figure 112026077748713-PAT00003_ABST
Abstract
Description
Technology Field
[0001] The present invention relates to an apparatus and method for providing photo output using a text prompt and an artificial intelligence template. Background Technology
[0003] Unless otherwise indicated in this specification, the contents described in this section are not prior art for the claims of this application, and are not to be recognized as prior art simply because they are included in this section.
[0004] With the recent expansion of unmanned photo studios and instant photo printers, services that allow users to take photos themselves and print them with desired frames or filters are becoming popular. Going beyond simple photo taking, these printing devices provide users with new entertainment elements by offering various digital stickers, background effects, and color correction features.
[0005] Conventional photo output devices primarily rely on methods that overlay pre-set 2D graphic elements onto the user's captured image or apply simple color filters. Alternatively, technology that recognizes the user's face and simply composites it with the face of a specific character or virtual props is applied. However, these conventional technologies remain at the level of merely attaching graphic elements to the original image or adjusting pixel color values, making it easy for a sense of incongruity to occur between the user's actual face shape and the composite elements. Furthermore, there are technical limitations in naturally incorporating a specific art style or detailed concept desired by the user without damaging the original photograph.
[0006] With the recent advancement of Generative AI technology, various models capable of generating high-quality images based on text prompts are being researched. However, an integrated system capable of applying this in real-time to offline instant photo printing devices to generate entirely new styles of images from the latent space stage while preserving the user's unique facial features is still lacking. Consequently, there is a continuously increasing demand for new types of photo printing devices and methods that can instantly generate and print natural, high-quality customized AI images by combining a user's facial feature points with AI templates such as text prompts. Prior art literature
[0008] Korean Registered Patent No. 10-2973100 (June 1, 2026) The problem to be solved
[0009] One embodiment of the present invention provides a device and method for providing photo output using a text prompt and an artificial intelligence template.
[0010] The technical problems to be solved by the present invention are not limited to those mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art to which the present invention belongs from the description below. means of solving the problem
[0012] To achieve the above-mentioned purpose, an electronic device according to one embodiment of the present invention comprises an input device, a camera, a printing device, a memory, and a processor connected to the memory. The processor acquires information regarding a selected layout, a selected frame design, and a selected AI template selected by a user among a plurality of preset layouts, a plurality of preset frame designs, and a plurality of preset AI templates through the input device, acquires a user-captured image through the camera, generates a user-AI image based on the user-captured image and the selected AI template through an artificial intelligence module, generates a basic output image by placing the user-AI image on the selected layout, generates an output image by applying the selected frame design to the basic output image, and prints the output image through the printing device.
[0013] At this time, the above layout includes information on the number, size, aspect ratio, and location of the divided areas where the user AI image is placed, the above frame design includes graphic background elements and text information applied to the border or margin of the divided areas, the above AI template includes metadata regarding the style, color tone, and lighting effects to be applied to the user-captured image, and text prompt information, and the above selection AI template may be an AI template that is individually selected for each divided area included in the selection layout.
[0014] At this time, the processor can generate the user AI image by extracting feature data including facial contours and the positions of facial features from the user-captured image through the artificial intelligence module, vectorizing text prompt information included in the selected AI template to generate template embedding data, and setting the feature data and the template embedding data as input conditions to the image generation neural network of the artificial intelligence module to perform a denoising operation to gradually remove noise in the latent space.
[0015] At this time, the image generation neural network may be an image generation model that has been pre-trained based on at least one of a Generative Adversarial Network (GAN), a Diffusion Model, and a Neural Style Transfer algorithm.
[0016] At this time, the processor detects a user's face region in the user AI image and calculates a first bias vector indicating the direction and distance where the center point of the face region is located relative to the center point of the user AI image, and a first area ratio indicating the area of the face region relative to the total area of the user AI image; for each of the plurality of segmented regions constituting the selection layout, calculates a second bias vector indicating the direction and distance where the center point of the corresponding segmented region is located relative to the total center point of the selection layout, and a second area ratio indicating the area of the corresponding segmented region relative to the total area of the selection layout; compares the first bias vector, the first area ratio, the second bias vector, and the second area ratio with each other to determine the placement position of the user AI image among the plurality of segmented regions such that the larger the area of the segmented region, the more the user AI image with a smaller face region area is matched; and inserts the user AI image into the determined segmented region to generate the basic output image.
[0017] At this time, the processor can calculate the placement suitability based on the dot product value between the first bias vector and the second bias vector and the area leveling error value based on the product of the first area ratio and the second area ratio, and determine the division area with the highest placement suitability as the placement location.
[0018] At this time, the above placement suitability is derived by the following mathematical formula,
[0019]
[0020] S_fit represents the above-mentioned placement fit, V_1 represents the above-mentioned first bias vector, V_2 represents the above-mentioned second bias vector, R_1 represents the above-mentioned first area ratio, R_2 represents the above-mentioned second area ratio, K represents a preset reference area constant for visual size equalization, and a and b may represent preset computational weights.
[0021] At this time, the electronic device further includes a temperature and humidity sensor, and the processor acquires temperature and humidity data of the space where the electronic device is located through the temperature and humidity sensor, corrects the template embedding data by reflecting the temperature and humidity data as weights in a vector value of a preset dimension that controls the color temperature and saturation among the template embedding data, performs the denoising operation based on the corrected template embedding data to generate the output image, and can print the output image by dynamically correcting the ink ejection amount of the printing device or the heating temperature of the thermal transfer head based on the humidity data. Effects of the invention
[0023] As such, according to one embodiment of the present invention, a device and method for providing photo output using a text prompt and an artificial intelligence template can be provided.
[0024] The effects obtainable from the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below. Brief explanation of the drawing
[0026] Other aspects, features, and benefits of specific preferred embodiments of the present invention, as described above, will become more apparent from the following description in conjunction with the accompanying drawings. FIG. 1 is a conceptual diagram of a photo output providing device using a text prompt and an artificial intelligence template according to one embodiment of the present invention. FIG. 2 is a block diagram of an electronic device according to one embodiment of the present invention. FIG. 3 is a diagram showing the generation of a user AI image according to an embodiment of the present invention. FIG. 4 is a diagram showing a first bias vector and a second bias vector according to an embodiment of the present invention. FIG. 5 is a drawing showing a temperature and humidity sensor according to one embodiment of the present invention. FIG. 6 is a flowchart of a method for providing photo output using a text prompt and an artificial intelligence template according to an embodiment of the present invention. It should be noted that in the drawings above, similar reference numbers are used to illustrate identical or similar elements, features, and structures. Specific details for implementing the invention
[0027] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.
[0028] In describing the embodiments, technical details that are well known in the technical field to which the present invention belongs and are not directly related to the present invention are omitted. This is intended to convey the essence of the present invention more clearly without obscuring it by omitting unnecessary explanations.
[0029] For the same reason, some components in the attached drawings have been exaggerated, omitted, or schematically depicted. Additionally, the size of each component does not entirely reflect its actual dimensions. Identical or corresponding components in each drawing have been assigned the same reference numbers.
[0030] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but can be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Throughout the specification, the same reference numerals refer to the same components.
[0031] At this point, it will be understood that each block of the process flow diagrams and combinations of the flow diagrams can be executed by computer program instructions. Since these computer program instructions can be loaded into the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment, the instructions executed through the processor of the computer or other programmable data processing equipment create means to perform the functions described in the flow diagram block(s). Since these computer program instructions can also be stored in computer-available or computer-readable memory that can be directed toward the computer or other programmable data processing equipment to implement the function in a specific way, the instructions stored in computer-available or computer-readable memory can also produce a manufactured item containing instruction means to perform the function described in the flow diagram block(s). Since computer program instructions can be loaded onto a computer or other programmable data processing equipment, instructions that perform a series of operation steps on the computer or other programmable data processing equipment to create a process executed by the computer can also provide steps for executing the functions described in the flowchart block(s).
[0032] Additionally, each block may represent a module, segment, or part of code containing one or more executable instructions for executing a specified logical function(s). It should also be noted that in some alternative execution examples, the functions mentioned in the blocks may occur out of order. For instance, two blocks described in succession may actually be executed substantially simultaneously, or the blocks may be executed in reverse order according to their corresponding functions.
[0033] In this embodiment, the term "part" refers to a software or hardware component such as a field-programmable gate array (FPGA) or an application-specific integrated circuit (ASIC), and the "part" performs certain roles. However, the meaning of "part" is not limited to software or hardware. The "part" may be configured to reside in an addressable storage medium or configured to run one or more processors. Accordingly, as an example, the "part" includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functions provided within the components and "parts" may be combined into a smaller number of components and "parts" or further separated into additional components and "parts." In addition, the components and '~parts' may be implemented to play one or more CPUs within the device or secure multimedia card.
[0034] In describing the embodiments of the present invention in detail, the primary focus will be on examples of specific systems, but the main point claimed in this specification is applicable to other communication systems and services having a similar technical background without significantly departing from the scope disclosed in this specification, and this will be possible at the judgment of a person with skilled technical knowledge in the relevant technical field.
[0035] In addition, the user terminal described below may include a communication-capable desktop computer, laptop computer, notebook, smartphone, tablet PC, mobile phone, smart watch, smart glass, e-book reader, PMP (portable multimedia player), portable game console, navigation device, digital camera, DMB (digital multimedia broadcasting) player, digital audio recorder, digital audio player, digital video recorder, digital video player, PDA (Personal Digital Assistant), etc.
[0037] FIG. 1 is a conceptual diagram of a photo output providing device using a text prompt and an artificial intelligence template according to one embodiment of the present invention, and FIG. 2 is a block diagram of an electronic device (100) according to one embodiment of the present invention.
[0038] An electronic device (100) according to one embodiment includes an input device, a camera, a printing device, a processor (110), and a memory (120). The processor (110) can perform at least one of the methods described above. The memory (120) can store information related to the method described above or store a program in which the method described above is implemented. The memory (120) may be a volatile memory or a non-volatile memory. The memory (120) may be referred to as a 'database', 'storage unit', etc.
[0039] The processor (110) can execute a program and control the electronic device (100). The code of the program executed by the processor (110) can be stored in memory (120). The device (100) can be connected to an external device (e.g., a personal computer or a network) through an input / output device (not shown) and exchange data.
[0040] At this time, the processor (110) can acquire information about a selected layout, a selected frame design, and a selected AI template selected by the user among a plurality of preset layouts, a plurality of preset frame designs, and a plurality of preset AI templates through the input device, acquire a user-captured image through the camera, generate a user AI image based on the user-captured image and the selected AI template through an artificial intelligence module, generate a basic output image by placing the user AI image on the selected layout, generate an output image by applying the selected frame design to the basic output image, and print the output image through the printing device.
[0041] At this time, the above layout includes information on the number, size, aspect ratio, and location of the divided areas where the user AI image is placed, the above frame design includes graphic background elements and text information applied to the border or margin of the divided areas, the above AI template includes metadata regarding the style, color tone, and lighting effects to be applied to the user-captured image, and text prompt information, and the above selection AI template may be an AI template that is individually selected for each divided area included in the selection layout.
[0042] In the present invention, layout refers to an overall geometric structure in which a photograph is placed, frame design refers to a two-dimensional graphic element that decorates the outer edge of the structure, and AI template refers to an artificial intelligence generation condition for transforming a subject inside the original photograph into a new form.
[0043] The preset values applied to these components are parameter data that are stored in memory in advance by the manager or service provider of the present invention or are periodically updated via an external communication network.
[0044] In this regard, specific pre-set thresholds or weight values can be determined as pre-optimized figures by analyzing seasonal themes, specific events, or accumulated user selection frequency data.
[0045] For example, for the spring season, the preset values of the AI template can be tuned to fix the saturation threshold of red pixels within the image at an increased level for a cherry blossom theme, and to include a text prompt that induces a bright atmosphere.
[0046] At this time, the present invention departs from the conventional method of simply adding stickers or applying only color filters to the acquired user-captured image, and generates an image by separating the roles of each setting value and combining them step by step.
[0047] An example is when a user selects a 4-part layout and individually assigns different AI templates, such as watercolor, cyberpunk, black and white film, or retro animation, to each partition area.
[0048] At this time, the processor generates four completely new user AI images in parallel through an artificial intelligence module, while maintaining the facial features of the original user-captured image and reflecting four different prompt conditions. Subsequently, the generated results are inserted into designated positions in the selection layout to complete the basic output image.
[0049] In addition, a user-selected frame design is finally applied to the borders and margins of the above-mentioned basic output image using an overlay method to generate a single complete output image.
[0050] Consequently, the present invention can technically resolve the visual discrepancy between the user's face and the graphic background that occurs due to simple image distortion or crude synthesis processes. By generating a high-quality customized result in which the user's unique characteristics and diverse concepts are fused into one through the deep learning computation of artificial intelligence, it has the effect of providing the user with maximized visual satisfaction and entertainment elements in an offline photo output environment.
[0052] FIG. 3 is a diagram showing the generation of a user AI image according to an embodiment of the present invention.
[0053] Referring to FIG. 3, the processor can generate the user AI image by extracting feature data including facial contours and the positions of facial features from the user-captured image through the artificial intelligence module, vectorizing text prompt information included in the selected AI template to generate template embedding data, and setting the feature data and the template embedding data as input conditions to the image generation neural network of the artificial intelligence module to perform a denoising operation to remove noise in the latent space step by step.
[0054] In this case, the feature data in the present invention refers to data obtained by mathematically extracting only geometric shape information, such as the relative positions of the eyes, nose, and mouth or the contours of the face, rather than simply copying pixels from the user's original photo.
[0055] Template embedding data refers to values obtained by converting text-based prompts into vectors, which are multi-dimensional numerical arrays, so that artificial intelligence can compute the concept selected by the user.
[0056] Furthermore, the latent space refers to a mathematical computational space where only essential information is compressed, rather than at the pixel level of the original image, and the denoising operation refers to the process of generating a clear image by progressively removing noise from a state of randomly distributed noise using feature data and template embedding data as guidelines.
[0057] Pre-set threshold values for the initial noise input amount or the number of denoising iterations applied during this computation process are optimized in advance according to the attributes of each AI template and stored in memory.
[0058] For example, the initial noise threshold can be set low for ID photo templates that must maintain the shape of the original face almost intact, and the initial noise threshold can be set high for 3D animation character templates that differ significantly from the original in shape, thereby controlling the degree of freedom of the AI's generation.
[0059] In addition, through this configuration, the present invention derives a technical effect that is clearly differentiated from conventional filter methods that simply overlay graphic stickers on existing images or change only the color of pixels in a one-dimensional manner.
[0060] For example, if a user selects a cyberpunk-style template, the present invention does not simply composite a neon sign background graphic onto a photograph of the user's face. Instead, feature data regarding the positions of the user's eyes, nose, and mouth within a latent space is combined with template embedding data regarding cyberpunk lighting to generate a completely new image of a person receiving neon lighting from scratch, pixel by pixel.
[0061] As a result, it is possible to derive a high-quality user AI image in which the art style and color scheme of the selected template are deeply reflected throughout the image, while preserving the user's unique facial features and identity.
[0062] This fundamentally eliminates the visual discrepancy commonly occurring between the user's face and background graphics, and enables the stable delivery of technically high-quality customized results.
[0064] At this time, the image generation neural network may be an image generation model that has been pre-trained based on at least one of a Generative Adversarial Network (GAN), a Diffusion Model, and a Neural Style Transfer algorithm.
[0065] At this time, the generative adversarial network mentioned in the present invention refers to a deep learning structure in which a generator that generates fake data and a discriminator that distinguishes it from real data learn competitively to generate a sophisticated image close to reality.
[0066] A diffusion model refers to an algorithm that generates a new type of image through a process of gradually adding noise to original data and then removing it in reverse.
[0067] Neural network-based style transfer refers to a technology that extracts and merges only style elements, such as art style or texture, from another image while maintaining the object shapes of the original image.
[0068] Threshold values, such as preset computational weights or the number of noise removal steps, used to drive this image generation model can be pre-optimized and stored in memory according to the characteristics of the individual AI template selected by the user.
[0069] For example, when applying a neural network-based style transfer algorithm, the ratio of content weights that maintain the shape of the original photo and style weights that reflect the art style of the template can be applied as a preset value, and in the case of a watercolor template, the threshold of the style weight can be set relatively high to induce a stronger paint bleeding effect.
[0070] In addition, the present invention can derive differentiated technical effects beyond simple pixel-level color correction or the addition of 2D graphic stickers by selectively or in combination utilizing at least one of these various artificial intelligence algorithms.
[0071] Even if the user selects an AI template of a different nature for each divided area of the selection layout, the present invention can complete the user AI image by individually applying an image generation model that best matches the characteristics of each template.
[0072] For example, when a user wants to convert their own face into a high-definition 3D character, a generative adversarial network can be activated to naturally adjust the proportions of the person's features, and when the user wants to completely replace the background with outer space, a diffusion model can be activated to generate new light and textures that are not present in the original.
[0073] As a result, it is possible to generate perfectly customized results without compromising the user's unique physical characteristics, thereby maximizing the quality of the final printed image and its value in offline photo printing environments.
[0075] FIG. 4 is a diagram showing a first bias vector and a second bias vector according to an embodiment of the present invention.
[0076] Referring to FIG. 4, the processor detects a user's face area in the user AI image and calculates a first bias vector indicating the direction and distance where the center point of the face area is located relative to the center point of the user AI image and a first area ratio indicating the area of the face area relative to the total area of the user AI image; for each of the plurality of segmented areas constituting the selection layout, calculates a second bias vector indicating the direction and distance where the center point of the segmented area is located relative to the total center point of the selection layout and a second area ratio indicating the area of the segmented area relative to the total area of the selection layout; compares the first bias vector, the first area ratio, the second bias vector, and the second area ratio with each other to determine the placement position of the user AI image among the plurality of segmented areas so that the larger the area of the segmented area, the smaller the area of the face area of the user AI image is matched; and inserts the user AI image into the determined segmented area to generate the basic output image.
[0077] In this case, the first bias vector in the present invention refers to an indicator that geometrically represents the direction and extent to which the user's actual face is deviated from the exact center of the generated user AI image.
[0078] The first area ratio represents the proportion of the face in the total area of the image. Likewise, the second bias vector and the second area ratio represent the direction in which a specific segmented area is separated from the center on the overall photo paper layout and the proportion of the size of that segmented area relative to the overall layout, respectively.
[0079] Pre-set weights or standard constants for visual standardization that serve as criteria when determining placement by comparing these geometric data can be derived in advance as values that provide the most aesthetic stability, based on the results of artificial intelligence learning multiple visual design data or the results of blind preference tests targeting users, and set in memory.
[0080] In addition, the present invention provides the effect of synchronizing the center of gravity of a subject within an image with the position of the frame by comparing bias vectors.
[0081] For example, if the face in the user AI image is skewed to the upper right and the first bias vector is directed toward the upper right, the present invention can automatically determine the placement position so that the image is matched to a segmented area located at the upper right in the overall layout.
[0082] Furthermore, if the sign of the preset vector matching weight is set to the opposite, a symmetrical and three-dimensional layout can be generated by placing the image skewed to the right in the left divided area in reverse, thereby concentrating the viewer's gaze on the entire photo towards the center.
[0083] Furthermore, the present invention derives a technical effect of perfectly optimizing the visual balance of the final printed material by cross-matching images based on area ratios.
[0084] For example, if a close-up image with a very large first-area ratio, rendered to fill the screen, is placed in the largest segment area, the person's face becomes excessively large in the printout, causing the overall harmony to collapse.
[0085] Accordingly, the present invention controls the position so that images with a large proportion of the face are assigned to a relatively small segmented area, and upper body or full-body images generated from a distance with a small proportion of the face are assigned to a large segmented area.
[0086] As a result, when AI images generated randomly with different compositions and proportions are combined into a single layout, the actual areas of individual faces printed on physical photographic paper are leveled to be similar to one another, thereby generating a visually very organized and high-quality base output image.
[0088] Additionally, the processor can calculate placement suitability based on the dot product value between the first bias vector and the second bias vector and the area leveling error value based on the product of the first area ratio and the second area ratio, and determine the division area with the highest placement suitability as the placement location.
[0089] To examine this more specifically, the above placement fit can be derived by the following mathematical formula 1.
[0090] [Mathematical Formula 1]
[0091]
[0092] At this time, S_fit represents the placement fit, V_1 represents the first bias vector, V_2 represents the second bias vector, R_1 represents the first area ratio, R_2 represents the second area ratio, K represents a preset reference area constant for visual size equalization, and a and b may represent preset computational weights.
[0093] Specifically, the mathematical formula presented in this invention is a structure that perfectly combines two independent objectives—visual geometric balance and physical leveling of face size—in the form of a single multivariable function.
[0094] The inner product operation included in the first operation term derives a positive maximum value as the directions of the first bias vector V_1 and the second bias vector V_2 align, thereby synchronizing the direction in which the subject is skewed within the image with the direction in which the segmented area is located within the overall layout.
[0095] Furthermore, the second operation is a penalty logic designed to induce the ratio of the physical area of the face finally printed on the photo paper to converge to a reference area constant K, which provides the most visual stability. The value obtained by multiplying the first area ratio R_1, representing the proportion of the face within the image, and the second area ratio R_2, representing the proportion of the divided area within the overall layout, exactly matches algebraically the final physical area ratio occupied by the face relative to the total area of the actual output.
[0096] Therefore, the present invention can enforce optimal matching by designing it so that as the product of the two ratios deviates from the K value, a large deduction occurs in the overall batch fit S_fit.
[0097] For example, one can assume a situation where the reference area constant K, which is the most harmonious final face area ratio, is set to 0.1, and the weights a and b, which control the weight of direction and area leveling, are set to 10 and 100, respectively.
[0098] If a user AI image with a very large face and a first area ratio R_1 of 0.5 is placed in a segmented area with a small size and a second area ratio R_2 of 0.2, the product of the two ratios becomes 0.1.
[0099] In this case, it perfectly matches the target value of K, which is 0.1, so the penalty due to area leveling error becomes 0. Therefore, if only the direction of the vector matches, it obtains the highest level of placement fit S_fit, which is 10 points, and is determined as the final placement location.
[0100] Conversely, one can assume a case where the corresponding user AI image with a large face is incorrectly placed in a very large segmented area with a second area ratio R_2 of 0.5.
[0101] In this case, the product of the two ratios becomes 0.25, resulting in a massive deduction of 15 points by multiplying the difference of 0.15 from the K value of 0.1 by the weight b of 100. Consequently, even if the vector direction is perfectly aligned, the batch fit S_fit drops sharply to minus 5 points, and the batch is automatically excluded by the system.
[0102] Consequently, the present invention has the effect of providing a mathematically proven placement logic that ensures all face sizes on printed materials are uniformly standardized without any logical errors, even when AI images of various scales are randomly given to segmented regions of varying sizes.
[0104] FIG. 5 is a drawing showing a temperature and humidity sensor according to one embodiment of the present invention.
[0105] Referring to FIG. 5, the electronic device further includes a temperature and humidity sensor, and the processor acquires temperature and humidity data of the space where the electronic device is located through the temperature and humidity sensor, corrects the template embedding data by reflecting the temperature and humidity data as weights in a vector value of a preset dimension that controls the color temperature and saturation among the template embedding data, generates the output image by performing the denoising operation based on the corrected template embedding data, and can print the output image by dynamically correcting the ink ejection amount of the printing device or the heating temperature of the thermal transfer head based on the humidity data.
[0106] In this case, the temperature and humidity sensor mentioned in the present invention refers to a hardware device that numerically measures the real-time weather and air conditions inside or around an unmanned photo booth where an electronic device is installed.
[0107] Among the template embedding data, the vector values of pre-set dimensions that control color temperature and saturation refer to specific mathematical parameters that determine the emotional atmosphere, such as the overall warmth or color intensity of the photograph, when artificial intelligence renders the image.
[0108] The preset reference values for temperature and humidity or weighted threshold values applied for such environmental correction can be derived in advance as values that produce optimal results based on the results of repetitive printing tests performed under various seasonal and weather conditions and physical hygroscopicity data of the photographic paper, and can be stored in memory.
[0109] For example, a preset standard humidity at which the photo paper can express colors without distortion can be fixed at 50 percent, and the weights of the color vector and the hardware control values can be set to be corrected at a specific ratio whenever the measured humidity exceeds this standard threshold by 10 percent.
[0110] Furthermore, through this configuration, the present invention derives a unique technical effect of directly involving physical environmental data around the device in the software-based artificial intelligence computation process.
[0111] For example, when a user takes a photo on a day when it rains heavily, humidity is high, and the temperature is relatively low, the present invention immediately corrects the dimension values of the template embedding data based on the acquired temperature and humidity data.
[0112] Accordingly, when the AI module performs denoising operations, it lowers the color temperature and slightly reduces the saturation, thereby completely regenerating a new user AI image at the pixel level in which the calm and slightly cool tones characteristic of a rainy day are naturally infused. This incorporates the uncontrollable physical variable of weather into the AI's latent spatial computation process, enabling a differentiated experience that provides the user with a personalized image that faithfully captures the atmosphere of the scene on that day.
[0113] Furthermore, the present invention provides the effect of fundamentally preventing physical defects in the output by utilizing environmental data even during the hardware operation stage of printing the generated image onto actual paper.
[0114] Generally, in environments with high humidity, photographic paper absorbs moisture from the air and expands; if printing is forced with the usual settings in such a state, the ink is likely to smudge or the paper may wrinkle due to the heat from the thermal transfer head.
[0115] The present invention dynamically controls the physical operation of a printing device based on measured humidity data, and when the humidity is higher than a preset threshold, it slightly reduces the ink ejection amount by a preset ratio or lowers the heating temperature of the thermal transfer head to a certain level to safely print an output image.
[0116] Consequently, the present invention can reliably guarantee the best print quality under any weather conditions by organically synchronizing the emotional parameters of artificial intelligence software and the physical parameters of a mechanical device using a single temperature and humidity data.
[0118] FIG. 6 is a flowchart of a method for providing photo output using a text prompt and an artificial intelligence template according to an embodiment of the present invention.
[0119] Referring to FIG. 6, a method for providing photo output using a text prompt and an AI template according to an embodiment of the present invention can obtain information about a selected layout, a selected frame design, and a selected AI template selected by a user among a plurality of preset layouts, a plurality of preset frame designs, and a plurality of preset AI templates through the input device (S101).
[0120] In addition, a method for providing photo output using a text prompt and an artificial intelligence template according to one embodiment of the present invention can acquire a user-captured image through the camera (S103).
[0121] In addition, the method for providing photo output using a text prompt and an AI template according to one embodiment of the present invention can generate a user AI image based on the user-captured image and the selected AI template through an AI module (S105).
[0122] In addition, a method for providing photo output using a text prompt and an artificial intelligence template according to one embodiment of the present invention can generate a basic output image by placing the user AI image on the selection layout (S107).
[0123] In addition, the method for providing photo output using a text prompt and an artificial intelligence template according to one embodiment of the present invention can generate an output image by applying the selection frame design to the basic output image (S109).
[0124] In addition, the method for providing photo output using a text prompt and an artificial intelligence template according to one embodiment of the present invention can print the output image through the printing device (S111).
[0125] In addition, the method for providing photo output using a text prompt and an artificial intelligence template according to one embodiment of the present invention may be configured in the same way as the device for providing photo output using a text prompt and an artificial intelligence template disclosed in FIGS. 1 to 5.
[0127] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.
[0128] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.
[0129] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.
[0130] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.
[0131] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below.
Claims
Claim 1 In an electronic device, an input device; a camera; a printing device; a memory; and a processor connected to the memory; The processor comprises: acquiring information regarding a selected layout, a selected frame design, and a selected AI template selected by a user from among a plurality of preset layouts, a plurality of preset frame designs, and a plurality of preset AI templates through the input device; acquiring a user-captured image through the camera; generating a user AI image based on the user-captured image and the selected AI template through an artificial intelligence module; generating a base output image by placing the user AI image on the selected layout; generating an output image by applying the selected frame design to the base output image; and printing the output image through the printing device. The layout includes information regarding the number, size, aspect ratio, and position of the divided areas where the user AI image is placed; the frame design includes graphic background elements and text information applied to the border or margin of the divided areas; the AI template includes metadata regarding the style, color tone, and lighting effects to be applied to the user-captured image, and text prompt information; and the selected AI template is individually for each divided area included in the selected layout The selected AI template is, and the processor extracts feature data including facial contours and the positions of facial features from the user-captured image through the artificial intelligence module, vectorizes text prompt information included in the selected AI template to generate template embedding data, and generates the user AI image by setting the feature data and the template embedding data as input conditions to the image generation neural network of the artificial intelligence module and performing a denoising operation to gradually remove noise in the latent space, and the image generation neural network isThe image generation model is pre-trained based on at least one of a Generative Adversarial Network (GAN), a Diffusion Model, and a Neural Style Transfer algorithm, and the processor detects a user's face region in the user AI image and calculates a first bias vector indicating the direction and distance where the center point of the face region is located relative to the center point of the user AI image, and a first area ratio indicating the area of the face region relative to the total area of the user AI image; for each of the plurality of segmented regions constituting the selection layout, it calculates a second bias vector indicating the direction and distance where the center point of the corresponding segmented region is located relative to the total center point of the selection layout, and a second area ratio indicating the area of the corresponding segmented region relative to the total area of the selection layout; and compares the first bias vector, the first area ratio, the second bias vector, and the second area ratio with each other to determine the placement position of the user AI image among the plurality of segmented regions such that the larger the area of the segmented region, the more the user AI image with a smaller face region area is matched, and the processor in the determined segmented region A user AI image is inserted to generate the above basic output image, and the processor calculates the placement suitability based on the dot product value between the first bias vector and the second bias vector and the area leveling error value based on the product of the first area ratio and the second area ratio, determines the segmented area with the highest placement suitability as the placement location, and the placement suitability is derived by the following mathematical formula, An electronic device characterized in that S_fit represents the above-mentioned placement fit, V_1 represents the above-mentioned first bias vector, V_2 represents the above-mentioned second bias vector, R_1 represents the above-mentioned first area ratio, R_2 represents the above-mentioned second area ratio, K represents a preset reference area constant for visual size leveling, and a and b represent preset computational weights. Claim 2 delete Claim 3 delete
Citation Information
Patent Citations
Method And Apparatus for Generating Personalized Image
KR1020250162995A
Photo booth systems and methods
US20100188707A1
