Method and device for generating past images using generative AI-based recall data of elderly people

Abstract

The present invention relates to a method and apparatus for generating past images using recollection data of an elderly person based on generative AI. The method for generating past images using recollection data of an elderly person based on generative AI, performed by a server, comprises: receiving user information including speech data spoken by an elderly user while recalling the past; analyzing the speech data to extract object information and time information for generating past images; and generating the user's past images based on the object information and time information. The past recalled by the user can be generated as an image using the speech data spoken by the user.

Description

Technology Field

[0001] The present invention relates to a method and apparatus for generating past images using recollection data of an elderly person based on generative AI, and more specifically, to a method and apparatus for generating past images using recollection data of an elderly person based on generative AI that can generate past images using speech data spoken by a user while reminiscing about the past. Background Technology

[0002] The population of the elderly aged 60 and over is rapidly increasing, and the number of people living alone in loneliness or suffering from the worsening of various diseases and dementia is on the rise. In particular, dementia in the elderly is a disease that causes suffering not only to the individual but also to their spouse and children, so there is a need for measures to verify the memory of the elderly and diagnose dementia in advance to delay the worsening of symptoms.

[0003] Furthermore, no matter how much local or central governments manage a region's history and culture, there are bound to be aspects that are overlooked, and this issue will only intensify with each passing generation. For instance, there is a problem in that useful materials—such as those passed down orally or based on the experiences of the elderly—are not being properly preserved, even though they are not recorded in history books or cultural archives. Therefore, there is a need to ensure the further development and continuation of local history and culture by documenting the experiences and memories of the elderly. Prior art literature

[0004] (Patent Document 0001) KR 10-2019-0093953 A The problem to be solved

[0005] The present invention provides a method and apparatus for generating past images using recollection data of an elderly person based on generative AI, which can generate images of the past recalled by the elderly person using speech data spoken by the elderly user.

[0006] The present invention provides a method and apparatus for generating past images using recollection data of the elderly based on generative AI, which can generate public content that can be utilized to supplement and correct content needed by the elderly or the history and culture of the region using past images.

[0007] The problems that the present invention aims to solve are not limited to those mentioned above, and other unmentioned problems will be clearly understood by a person skilled in the art from the description below. means of solving the problem

[0008] A method and apparatus for generating past images using recollection data of an elderly person based on generative AI according to an embodiment of the present invention may include a method for generating past images using recollection data of an elderly person based on generative AI performed by a server, the method comprising: receiving user information including speech data spoken by an elderly user while reminiscing about the past; analyzing the speech data to extract object information and time information for generating past images; and generating past images of the user based on the object information and the time information.

[0009] The above speech data includes voice data, and the process of receiving the user information may further include the process of receiving at least one of video data and text data that allows the user to recall the past.

[0010] The above object information may include at least one of place information, person information, building information, animal information, plant information, and object information.

[0011] The process of generating the past image above may include: a process of generating a base image using the object information; and a process of correcting the base image using the time information.

[0012] The process further includes collecting auxiliary information for correcting the above-mentioned basic image, and the process of collecting the auxiliary information may collect at least one of public image information, crawling information, and user feedback information related to at least one of the object information and the time information.

[0013] The process of correcting the above basic image may include a process of supplementing the above basic image using the above auxiliary information corresponding to the above timing information.

[0014] The process of collecting the above auxiliary information includes collecting reference auxiliary information corresponding to the above time information, previous auxiliary information corresponding to an earlier time based on the above time information, and subsequent auxiliary information corresponding to a later time. In the case where the above reference auxiliary information is not available, the process of supplementing the above basic image may supplement the above basic image by using the auxiliary information of the time closest to the above time information among the previous auxiliary information and the subsequent auxiliary information.

[0015] A past image generation device utilizing recollection data of an elderly person based on generative AI according to an embodiment of the present invention comprises: a user terminal capable of acquiring user information including speech data spoken by an elderly person while reminiscing about the past; and a management server capable of receiving said user information and generating said past images of said user based on said speech data. The management server analyzes said speech data to extract object information and time information for generating past images, and can generate said past images of said user based on said object information and said time information.

[0016] A method for generating past images using recollection data of an elderly person based on generative AI according to an embodiment of the present invention can be performed by a computer program stored on a recording medium readable by a computer, combined with a computer which is hardware.

[0017] Other specific details of the present invention are included in the detailed description and drawings. Effects of the invention

[0018] According to the present invention, images of the past recalled by an elderly user can be generated using speech data spoken by the user. Furthermore, the generated images of the past can be used to create user-customized content, such as albums or memoirs, and can be used to create public content to supplement and correct the history and culture of the region. In particular, the images of the past and the user-customized content generated using them can be utilized to improve the user's cognitive function and prevent dementia, and can help in remembering the user after their death.

[0019] In addition, users can provide historical images as resources to supplement and preserve historical materials by uniqueizing them, thereby generating revenue.

[0020] The effects of the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by a person skilled in the art from the description below. Brief explanation of the drawing

[0021] FIG. 1 is a conceptual diagram of a past image generation device utilizing recollection data of an elderly person based on generative AI according to an embodiment of the present invention. FIG. 2 is a block diagram illustrating the detailed configuration of the management server shown in FIG. 1. FIG. 3 is a block diagram schematically showing a method for generating past images using recollection data of an elderly person based on generative AI according to an embodiment of the present invention. FIG. 4 is a diagram showing an example of generating a past image using a generative AI-based past image generation method utilizing recall data of an elderly person according to an embodiment of the present invention. Specific details for implementing the invention

[0022] 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 may 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 present invention, and the present invention is defined only by the scope of the claims.

[0023] The terms used in this specification are for describing embodiments and are not intended to limit the invention. In this specification, the singular form includes the plural form unless specifically stated otherwise in the text. The terms "comprises" and / or "comprising" used in this specification do not exclude the presence or addition of one or more other components in addition to the components mentioned. Throughout the specification, the same reference numerals refer to the same components, and "and / or" includes each of the mentioned components and all combinations of one or more. Although terms such as "first," "second," etc., are used to describe various components, these components are not limited by these terms. These terms are used merely to distinguish one component from another. Therefore, the first component mentioned below may be the second component within the technical scope of the invention.

[0024] Unless otherwise defined, all terms used herein (including technical and scientific terms) may be used in a meaning commonly understood by those skilled in the art to which the present invention pertains. Additionally, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise.

[0025] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

[0026] FIG. 1 is a conceptual diagram of a past image generation device utilizing recollection data of an elderly person based on generative AI according to an embodiment of the present invention, and FIG. 2 is a block diagram for explaining the detailed configuration of a management server shown in FIG. 1.

[0027] As illustrated in FIGS. 1 and 2, a past image generation device utilizing recollection data of an elderly person based on generative AI according to an embodiment of the present invention may include a user terminal (10), a management server (20), and an administrator terminal (30). In this case, the administrator terminal (30) may be omitted.

[0028] Here, the user terminal (10), the management server (20), and the administrator terminal (30) can synchronize and transmit and receive data in real time using a wireless communication network. Wireless communication networks can support various long-distance communication methods, for example, Wireless LAN (WLAN), DLNA (Digital Living Network Alliance), WiBro (Wireless Broadband), WiMAX (World Interoperability for Microwave Access), GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), CDMA2000 (Code Division Multi Access 2000), EV-DO (Enhanced Voice-Data Optimized or Enhanced Voice-Data Only), WCDMA (Wideband CDMA), HSDPA (High Speed ​​Downlink Packet Access), HSUPA (High Speed ​​Uplink Packet Access), IEEE 802.16, Long Term Evolution (LTE), LTEA (Long Term Evolution-Advanced), Wireless Mobile Broadband Service (WMBS), BLE (Bluetooth Low Energy), Zigbee, RF (Radio Frequency), Various communication methods such as LoRa (Long Range) may be applied, but are not limited thereto; various widely known wireless or mobile communication methods may also be applied.

[0029] First, the past image generation device utilizing recollection data of an elderly person based on generative AI according to the present embodiment can generate a past image that the user is recalling based on speech data that the user speaks while recalling the past.

[0030] To this end, the user terminal (10) can receive speech data spoken by the user and transmit the received speech data to the management server (20) using a communication network. Here, the speech data may include voice data. Additionally, the user terminal (10) may be connected to the management server (20) in the form of an app or web, such as a smartphone or computer, and can receive additional information such as text data and video data possessed by the user, as well as personal information such as the user's gender, age, and health status, and transmit it to the management server (20). Hereinafter, information that is input or provided by the user, such as speech data and additional information, is referred to as user information.

[0031] Additionally, the user terminal (10) can receive past images, user-customized content, etc. generated using speech data or user information from the management server (20) and visualize them so that the user can check them.

[0032] Such user terminals (10) are terminals possessed by the user and can operate using an application program (or application) in the present disclosure, and such application programs can be downloaded from an external server or management server (20) via wireless communication. For example, user terminals (10) may include various terminals such as smartphones, PDAs (Personal Digital Assistants), tablets, wearable devices (e.g., smartwatches, smart glasses, head-mounted displays, etc.), desktops, and various IoT (Internet of Things) terminals, but are not limited thereto.

[0033] The management server (20) receives user information transmitted from the user terminal (10) and can use generative AI to generate user information, particularly speech data, into a past image of the past that the user is recalling. Additionally, the management server (20) can generate user-customized content using the generated past image.

[0034] Referring to FIG. 2, the management server (20) may include a data transmission / reception unit (21), a data analysis unit (22), a database unit (23), a data collection unit (24), a matching unit (25), an image generation unit (26), and a content generation unit (76). Such a management server (20) may be implemented by a hardware circuit (e.g., a CMOS-based logic circuit), firmware, software, or a combination thereof. For example, it may be implemented by utilizing transistors, logic gates, and electronic circuits in the form of various electrical structures.

[0035] The data transmission / reception unit (21) can be connected to the user terminal (10) via a communication network and can receive user information transmitted from the user terminal (10). Additionally, the data transmission / reception unit (21) can receive an image generation request from the user terminal (10) and can transmit the generated images, such as a base image, an intermediate image, a past image, etc., to the user terminal (10). Here, the base image refers to the image initially generated using user information received from the user terminal (10), and the past image refers to an image requested by the user terminal (10), such as a final image. Furthermore, the intermediate image is an image generated during the intermediate process of generating the past image after the base image is generated, and may or may not be generated as needed.

[0036] The data analysis unit (22) can extract object information and time information necessary for generating past images from user information based on an image generation request. Here, object information may refer to people, places, objects, buildings, actions, food, animals, plants, etc., and time information may refer to a time prior to the present recalled by the user. For example, object information may include me, mom, teacher, school, beach, Seoul Station, Bulguksa Temple, landmark, park, market, funeral home, swimming, sleep, meal, bulgogi, puppy, cat, ginkgo tree, etc., and time information may include 1981, 1999, 2010, etc. Additionally, time information may include seasons such as spring, summer, autumn, and winter, and specifically may include specific years, months, and days.

[0037] This data analysis unit (22) may include a natural language analysis module (not shown). In determining image generation conditions based on an image generation request, the data analysis unit (22) may use the natural language analysis results of user information, particularly user comments included in speech data. Natural language analysis is performed by a natural language analysis module and may include lexical analysis, syntactic analysis, and semantic analysis. For example, the data analysis unit (22) may use the natural language analysis results of speech data such as "I remember going to Haeundae Beach with my wife and daughter in August 1995 and swimming" to extract "me," "wife," "daughter," "Haeundae Beach," and "swimming" as object information, and extract "August 1995" as time information.

[0038] The object information and time information extracted from the data analysis unit (22) can be stored in the database unit (23).

[0039] The data collection unit (24) can collect auxiliary information necessary for generating past images other than object information and time information. For example, the data collection unit (24) can collect auxiliary information for generating backgrounds necessary for generating past images. At this time, the auxiliary information may include public image information, crawling information, etc. For example, the public image information may include old movies and dramas related to the subject that have no copyright issues, and the crawling information may include building registers, etc.

[0040] Auxiliary information collected by the data collection unit (24) can be analyzed by the data analysis unit (22), classified according to the subject, and stored in the database unit (23). Here, the subject is defined through named entity recognition, and the auxiliary information can be classified by corresponding it with object information such as people, places, objects, buildings, etc. For example, if the collected auxiliary information is "Haeundae Beach image," it can be classified under the subject of beach.

[0041] The data analysis unit (22) may include an image analysis module (not shown) for analyzing collected auxiliary information. The image analysis module may include a color profiling method, a texture and pattern distribution, a YOLO model capable of object detection and recognition, a CNN (Convolutional Neural Network) model, etc. For example, the image analysis module can classify images according to a predetermined topic by finding features within the auxiliary information and calculating the distance of association between them.

[0042] The database section (23) may include a user information DB (231), an auxiliary information DB (232), a basic image DB (233), and a content DB (234).

[0043] The user information DB (231) can store user information received from the user terminal (10), object information extracted from the data analysis unit (22), and time information by matching them with each other. In addition, the user information DB (231) can store images generated using auxiliary information as well as user information, such as basic images, intermediate images, and past images.

[0044] The auxiliary information DB (232) can store auxiliary information, such as images, that is collected through the data collection unit (24), analyzed and classified by the data analysis unit (22). At this time, the auxiliary information DB (232) can store the auxiliary information classified by topic.

[0045] The basic image DB (233) is a set of images prepared in advance to generate images based on object information and time information, and may include images that are publicly available and not subject to copyright. This basic image DB (233) can classify and store images, such as basic images, according to themes such as people, places, and seasons.

[0046] The content DB (234) can store user-customized content, historical data content, etc., generated using past images created using user information and auxiliary information. For example, user-customized content may include albums, memoirs, etc., and historical data content may include the process of change of local cultural heritage, etc. In this case, the historical data content may be used as public content.

[0047] The matching unit (25) can match data stored in the database unit (23) according to object information and time information for generating past images. For example, the matching unit (25) can select and match basic images stored in the basic image DB (233) and auxiliary information stored in the auxiliary information DB (232) according to object information and time information.

[0048] The image generation unit (26) can generate a past image using the basic image selected and matched by the matching unit (25) and auxiliary information. The image generation unit (26) can generate a basic image based on object information and time information, and generate a past image by creating a background on the basic image using auxiliary information. At this time, the past image may be the final image or an intermediate image. That is, the image generation unit (26) can generate a basic image and a background, and transmit the generated image to the user terminal (10) through the data transmission and reception unit (21). It can also receive feedback information from the user terminal (10) through the data transmission and reception unit (21). If there is a user's request for modification in the feedback information, the image may become an intermediate image. Subsequently, if the image generation unit (26) modifies and supplements the intermediate image using the feedback information, the image may become the past image, which is the final image. On the other hand, if there is a user's request for approval in the feedback information, the image may become the past image, which is the final image.

[0049] This image generation unit (26) may include an image generation module (not shown) capable of generating an image according to image generation conditions such as object information, time information, and auxiliary information. In this case, the image generation module may include a generative AI model. For example, the image generation module may generate a base image according to object information and time information by utilizing various techniques such as vector drawing techniques and machine learning techniques. Additionally, the image generation module may generate a past image using the base image and auxiliary information by utilizing a StyleGAN model.

[0050] The content creation unit (76) can create user-customized content, such as albums and memoirs, or public content, such as historical materials, using past images created by the image creation unit (26) when there is a request from a user. Various contents created by the content creation unit (76) can be stored in the user information DB (231).

[0051] The administrator terminal (30) is a terminal possessed by a separate administrator and can transmit and receive data by being synchronized in real time with the user terminal (10) and the management server (20) using a wireless communication network. At this time, the administrator terminal (30) can transmit and receive data using an application program.

[0052] Such an administrator terminal (30) may include various portable electronic communication devices that support communication with a user terminal (10) and a management server (20). For example, as a separate smart device, it may include various terminals such as a smartphone, PDA (Personal Digital Assistant), tablet, wearable device (e.g., smartwatch, smart glass, HMD (Head Mounted Display), etc.) and various IoT (Internet of Things) terminals, but is not limited thereto.

[0054] Hereinafter, a method for generating past images using recollection data of an elderly person based on generative AI according to an embodiment of the present invention is described. The method for generating past images using recollection data of an elderly person based on generative AI according to an embodiment of the present invention is the aforementioned method for generating past images using recollection data of an elderly person based on generative AI It can be performed using a device.

[0055] FIG. 3 is a block diagram schematically showing a method for generating past images using recollection data of an elderly person based on generative AI according to an embodiment of the present invention, and FIG. 4 is a diagram showing an example of generating past images using a method for generating past images using recollection data of an elderly person based on generative AI according to an embodiment of the present invention.

[0056] Referring to FIG. 3, a method for generating a past image using recollection data of an elderly person based on generative AI according to an embodiment of the present invention may include, in a method for generating a past image based on recollection performed by a server, a process of receiving user information including speech data that the user speaks while reminiscing about the past, a process of analyzing the speech data to extract object information and time information for generating a past image, and a process of generating a past image of the user based on the object information and the time information.

[0057] Here, the age of the user may vary, but is described as an elderly person or senior citizen aged 60 or older.

[0058] When a request for the creation of a past image is received from a user terminal (10), the management server (20) may request the transmission of user information to the user terminal (10) through a data transmission / reception unit (21). At this time, the user information may include personal information such as the user's age, gender, and address, and speech data spoken while recalling the past. For example, the speech data may be a sentence such as, "I remember going to Haeundae Beach with my wife and daughter in August 1995 and swimming." At this time, the user is currently recalling the summer of 30 years ago, for example, in July 2024. Additionally, the user information may further include various data such as text data and video data possessed by the user. At this time, the video data may include the user's past photos / videos and current photos / videos.

[0059] When user information is received from a user terminal (10), the data analysis unit (22) of the management server (20) can analyze the speech data to extract object information and time information. At this time, the object information may be "me," "wife," "daughter," "Haeundae Beach," or "beach," and the time information may be "August 1995."

[0060] In this way, when object information and time information are extracted, the extracted information can be stored in the user information DB (231) of the database unit (23).

[0061] Additionally, the data collection unit (24) can collect auxiliary information necessary for generating images. For example, the data collection unit (24) can collect images of Haeundae Beach from about 5 to 10 years prior to and 5 to 10 years after August 1995. The auxiliary information may include reference auxiliary information corresponding to past time information recalled by the user, previous auxiliary information corresponding to a time prior to the time information, and subsequent auxiliary information corresponding to a time subsequent.

[0062] When auxiliary information is collected, the data analysis unit (22) can analyze the collected auxiliary information, such as images of Haeundae Beach, through the image analysis module, select images suitable for object information and time information, and store them in the auxiliary information DB (232).

[0063] Afterwards, the matching unit (25) can match object information, time information and auxiliary information necessary to generate a past image.

[0064] And the image generation unit (26) can generate a basic image using object information and generate a past image using auxiliary information that matches the time information. In this case, if there is no standard auxiliary information that matches the time information, the auxiliary information of the time closest to the time information among the previous auxiliary information and the subsequent auxiliary information can be used. For example, if there is no image of Haeundae Beach in August 1995, but there are images of Haeundae Beach in August 1994, July 1995, and October 1995, the image of Haeundae Beach in July 1995, which is the time closest to August 1995, can be used as auxiliary information.

[0065] Referring to FIG. 4, the image generation unit (26) can generate a basic image of a past recall period recalled by the user using object information. Additionally, among the collected auxiliary information, at least one of the image prior to the past recall period and the image after the past recall period can be merged to generate an intermediate image. At this time, at least one of the image prior to the past recall period and the image after the past recall period can be used to generate a background of the basic image or to supplement the generated background.

[0066] In this way, when an intermediate image is generated, the intermediate image can be transmitted to a user terminal (10) through a data transmission / reception unit (21). Then, when a signal acknowledging the image is transmitted from the user terminal (10), the image, such as the intermediate image, can be set as the final image, which is the past image.

[0067] On the other hand, if a signal (feedback) is transmitted from the user terminal (10) requesting modification or supplementation without approving the image, the image can be modified or supplemented using auxiliary information and user feedback information as an intermediate image to create a past image, which is the final image. For example, there may be a request in the feedback information to add an image of a specific building that the user is recalling, as the intermediate image does not contain that specific building. In this case, auxiliary information including building register information for the specific building requested by the user, such as the time of the user's past recall, can be additionally collected, and an image of the building can be additionally collected through street view search, etc. Once an image of the building is collected, the collected image can be stored in the auxiliary information DB (232). Then, the past image, which is the final image, can be created by adding the image of the building to the intermediate image through the image generation unit (26).

[0068] This method can generate more accurate historical images by utilizing various data, thereby increasing user satisfaction.

[0069] In this way, when a past image is generated, the generated image can be stored in the user information DB (231) and transmitted to the user terminal (10) through the data transmission and reception unit (21).

[0070] Additionally, if requested by a user, user-customized content, public content, etc., can be generated using past images through the content generation unit (76). At this time, the content generation unit (76) may generate albums, memoirs, etc., according to the user's request, and may be used as public material to supplement historical data. For example, if a user has passed away, the generated album may be used to recreate the deceased's past appearance or to produce a retrospective video in the form of an autobiography. This allows the bereaved family to receive deeper emotion and comfort through the retrospective video.

[0071] Historical images and user-customized content generated through such methods can be utilized to verify the memory of users, such as the elderly, and as tools for diagnosing dementia. Furthermore, images generated during the dementia diagnosis process can be used to provide paid diagnostic services, sell images created during diagnoses to families, or generate revenue by offering customized services linked to dementia management programs. Additionally, public content based on images derived from the memories of elderly users can be utilized as materials for producing historical films or documentaries, or as resources for history education. In this context, combining historical images with AR / VR technology during the production of public content offers the advantage of enhancing the concentration of viewers.

[0072] The steps of the method or algorithm described in connection with embodiments of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. The software module may reside in RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), Flash Memory, a hard disk, a removable disk, a CD-ROM, or any form of computer-readable recording medium well known in the art to which the present invention belongs.

[0073] Although embodiments of the present invention have been described above with reference to the attached drawings, those skilled in the art will understand that the present invention may be implemented in other specific forms without altering its technical concept or essential features. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. Explanation of the symbols

[0074] 10: User terminal 20: Management server 21: Data Transmission / Reception Unit 22: Data Analysis Unit 23: Database Department 24: Data Collection Department 25: Matching section 26: Image generation section 27: Content Creation Section

Claims

Claim 1 A method for generating past images using recollection data of an elderly person based on generative AI performed by a server, comprising: a process of receiving user information including speech data spoken by an elderly user while reminiscing about the past; a process of analyzing the speech data to extract object information and time information for generating past images; and a process of generating past images of the user based on the object information and time information; wherein the process of generating past images comprises a process of generating a base image using the object information; A method for generating past images using recollection data of an elderly person based on generative AI, comprising: a process of correcting the basic image using the above time information; further comprising a process of collecting auxiliary information for correcting the basic image, wherein the process of collecting auxiliary information collects at least one of public image information, crawling information, and user feedback information related to at least one of the object information and the above time information; wherein the process of correcting the basic image includes a process of supplementing the basic image using the auxiliary information corresponding to the above time information; wherein the process of collecting auxiliary information includes a process of collecting reference auxiliary information corresponding to the above time information, previous auxiliary information corresponding to an earlier time based on the above time information, and subsequent auxiliary information corresponding to a later time; and wherein, in the case where the reference auxiliary information is not available, the process of supplementing the basic image supplements the basic image using the auxiliary information of the time closest to the above time information among the previous auxiliary information and the subsequent auxiliary information. Claim 2 A method for generating past images using recollection data of an elderly person based on generative AI according to claim 1, wherein the utterance data includes voice data, and the process of receiving user information further includes the process of receiving at least one of image data and text data provided by the user. Claim 3 A method for generating past images using recollection data of an elderly person based on generative AI according to claim 1, wherein the object information comprises at least one of place information, person information, building information, animal information, plant information, and object information. Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 delete Claim 8 A past image generation device utilizing recollection data of an elderly person based on generative AI, comprising a user terminal capable of acquiring user information including speech data spoken by an elderly user while reminiscing about the past; A generative AI-based device for generating past images of an elderly person utilizing recall data, comprising: a management server capable of receiving the above user information and generating the user's past image based on the above utterance data; wherein the management server analyzes the above utterance data to extract object information and time information for generating past images, generates a basic image using the above object information, corrects the above basic image using the above time information to generate a past image recalled by the user, collects auxiliary information for correcting the above basic image, collects at least one of public image information, crawling information, and user feedback information related to at least one of the above object information and above time information, corrects the above basic image by supplementing the above basic image using the auxiliary information corresponding to the above time information, collects the auxiliary information by collecting reference auxiliary information corresponding to the above time information, previous auxiliary information corresponding to an earlier time based on the above time information, and subsequent auxiliary information corresponding to a later time, and if the above reference auxiliary information is not available, supplements the above basic image by using the auxiliary information of the time closest to the above time information among the previous auxiliary information and the subsequent auxiliary information. Claim 9 A computer program stored on a computer-readable recording medium that is combined with a computer, which is hardware, to perform the method of any one of claims 1 to 3.

Images