KR-102961948-B1 - SYSTEM FOR CONTROLLING INTERACTION OF CONTEXT AWARE VIRTUAL HUMAN

KR102961948B1KR 102961948 B1KR102961948 B1KR 102961948B1KR-102961948-B1

Abstract

The present invention relates to an interaction control system for a context-aware virtual human, and specifically comprises: a user detection unit that detects a user determined to use the kiosk from the front of the kiosk through a camera provided in the kiosk; a virtual human activation unit that activates a virtual human, which is a conversational artificial intelligence model implemented in the kiosk, when a user is detected from the front of the kiosk by the user detection unit; an identity verification unit that scans the face of the detected user using a camera provided in the kiosk to generate a face image set composed of face image data, and verifies the identity by comparing the generated face image set with face information previously registered in a database; and an interaction content provision unit that provides personalized interaction content using the virtual human according to the identity verification result from the identity verification unit.

Inventors

김성현
박재원

Assignees

주식회사 솔트룩스

Dates

Publication Date: 20260507
Application Date: 20241030

Claims (20)

In a context-aware virtual human interaction control system implemented as a computing device comprising one or more processors and one or more memories storing instructions executable on said processors, A user detection unit that detects a user determined to use the kiosk from the front of the kiosk through a camera provided in the kiosk; A virtual human activation control unit that controls whether to activate a virtual human, which is an interactive artificial intelligence model implemented in the kiosk, depending on whether a user is detected in front of the kiosk; An identity verification unit that uses a camera provided in the above kiosk to scan the detected user's face to generate a face image set composed of face image data, and verifies the identity by comparing the generated face image set with face information previously registered in a database; and, It includes an interaction content providing unit that provides personalized interaction content using the virtual human according to the identity verification result of the identity verification unit above, and The above user detection unit is, Based on the installation point of the above kiosk, if the distance from the point where the user is located is within a first threshold distance, the user is detected as an active user directly interacting with the kiosk, and The above user detection unit is, Based on the installation point of the kiosk, if the distance from the point where the user is located exceeds the first threshold distance while simultaneously being within the second threshold distance, the user is detected as a potential user who may use the kiosk. The above user detection unit is, A context-aware virtual human interaction control system characterized by outputting a message that induces access through the virtual human of the interaction content provider when the above user is detected as the above potential user.
delete
delete
delete
In paragraph 1, The above-mentioned identity verification unit is, The user's face area detected by the user detection unit is displayed as a bounding box, and the user's face is scanned a threshold number of times through a camera provided in the kiosk starting from a preset specific point in time to generate a face image set composed of multiple face image data. A context-aware virtual human interaction control system characterized by verifying the identity of a user by determining whether there exists a user with an identity corresponding to the face image set in the face information previously registered in the database using a deep learning-based face recognition model.
In paragraph 5, The specific point in time previously set above is, This is the point in time when the width of the bounding box for the detected user face region exceeds a preset threshold, and A context-aware virtual human interaction control system characterized by performing the user's face scan at least 5 times.
In paragraph 5, The above-mentioned identity verification unit is, In the above database, employee face information and regular visitor information are pre-registered as face registration information, and the identity of the detected user is classified as one of an employee, a regular visitor, or an unidentified user. A queuing unit that stores user face image data included in a face image set for verifying the identity of the detected user in a queue; and, A context-aware virtual human interaction control system further comprising a reranking unit that performs reranking by applying face image data stored in the above queue to a heuristic algorithm, and is characterized by finally determining the identity of the detected user from the result of the reranking unit.
In Paragraph 7, The above-mentioned identity verification unit is, If the above active user is determined to be an unknown user, A context-aware virtual human interaction control system characterized by providing a query for identity verification to the active user through the interaction content provider, and enabling the active user to collect a response to the provided query and perform identity verification using an input means provided in the kiosk.
In Paragraph 7, The above heuristic algorithm is, As a result of comparing the face image data included in the above face image set with the feature vectors of the face registration information registered in the above database, A context-aware virtual human interaction control system characterized by determining that the active user and the identified user are the same person when one identity is confirmed in a set of face images and the similarity score calculated by the face recognition model is less than a preset cut score.
In Paragraph 7, The above heuristic algorithm is, As a result of comparing the face image data selected with the aforementioned preset number with the feature vectors of the face registration information registered in the aforementioned database, An interaction control system for a context-aware virtual human characterized by, when two or more identities are identified in a set of face images, calculating the average of similarity scores for each identity and comparing it with a preset cut score to determine if they are the same person.
In Paragraph 10, The above heuristic algorithm is, A context-aware virtual human interaction control system characterized by identifying two or more identities in a set of face images, and when multiple identities are derived with the average of the similarity scores being less than a preset cut score, comparing the average similarity scores for each identity and determining the user of the identity having the lowest average similarity score as the same person as the active user.
In Paragraph 10, The above heuristic algorithm is, A context-aware virtual human interaction control system characterized by, when two or more identities are identified in a set of face images and two or more identities are derived such that the average of the similarity scores is less than a preset cut score, additionally performing a face image scan of the active user using a camera equipped in the kiosk.
In Paragraph 10, The above heuristic algorithm is, A context-aware virtual human interaction control system characterized by, when there are no identities with a score lower than a preset cut score as a result of averaging the similarity scores above, applying a boosting weight according to the frequency of appearance of identities to finally select a user corresponding to an identity judged to have the highest similarity to the active user based on the boosted score, and determining them to be the same person.
In Paragraph 13, The above boosting weight is, An interaction control system for a context-aware virtual human characterized by assigning a relatively higher weight as the frequency of appearance of an identity in a set of face images increases.
In paragraph 1, If the user detected by the user detection unit is determined to be an active user, A context-aware virtual human interaction control system characterized by the interaction content provider calculating the azimuth of the active user and adjusting the gaze angle of the virtual human to correspond to the calculated azimuth.
In paragraph 1, If the user detected by the user detection unit is determined to be an active user, A context-aware virtual human interaction control system characterized by the interaction content providing unit calculating the azimuth angle of the active user and, when a change in the position of the active user is detected during the provision of the interaction content, adjusting the rotation angle of the neck portion of the virtual human to correspond to the changed position.
In paragraph 1, The above interaction content providing unit is, A context-aware virtual human interaction control system characterized by loading a user profile corresponding to the identity verification result of the active user and providing differentiated interactions of the virtual human according to the information included in the user profile.
In paragraph 1, The above interaction content providing unit is, A context-aware virtual human interaction control system characterized by providing differentiated context menus executable through the kiosk according to the identity of the active user.
In paragraph 1, The virtual human activation control unit above is, A context-aware virtual human interaction control system characterized by the user detection unit deactivating the virtual human and switching the kiosk to standby mode when no user is detected in front of the kiosk.
In paragraph 1, When the identity of the active user using the kiosk is verified by the identity verification unit above, It further includes an interaction management unit that stores and manages interaction data between an active user whose identity has been verified and the virtual human; A context-aware virtual human interaction control system characterized in that the interaction content providing unit provides continuous interaction content based on the interaction data history stored in the interaction management unit when providing the next interaction content.

Description

Context-Aware Virtual Human Interaction Control System The present invention relates to an interaction control system for a context-aware virtual human, and specifically, to a technology that interacts with a kiosk user through a virtual human implemented in a kiosk, and in particular provides personalized interaction content to the user whose identity has been verified, thereby enhancing the user's experience and immersion in using the kiosk. Kiosks, recently installed throughout society, are rapidly becoming widespread due to advantages such as reduced labor costs and improved service efficiency. These kiosks are characterized by providing information or various services to users through non-face-to-face interaction, whereas traditional kiosks generally operate based on a touchscreen where the user inputs commands and the system responds. However, it has been pointed out that the interaction of traditional kiosks is not only extremely static and one-sided, but also that different user interfaces are applied depending on the installation location, requiring users to take time to familiarize themselves with the interface to use the kiosk, and providing an unnatural experience that lacks immersion. Accordingly, although Korean Public Patent No. 10-2023-0144422 (Kiosk User Interface Optimization System and Method, and Optimization Device thereof) discloses a kiosk user interface optimization technology capable of providing optimized product information via a kiosk according to the store or customer, there were limitations in that users still needed time to familiarize themselves with the kiosk interface, and the lack of dynamic interaction resulted in only unnatural interactions, thus failing to resolve the problem of reduced immersion associated with kiosk usage. FIG. 1 is a configuration diagram of a context-aware virtual human interaction control system according to an embodiment of the present invention. FIG. 2 is a conceptual diagram of distinguishing active users and potential users based on the distance between a kiosk and a user according to an embodiment of the present invention. FIG. 3 is an example of outputting an access induction message to a user identified as a potential user through a kiosk according to an embodiment of the present invention. FIG. 4 is a conceptual diagram of capturing face image data for user identity verification based on the distance between a kiosk and a user according to an embodiment of the present invention. FIG. 5 is a conceptual diagram of generating a face image set using face image data captured for user identity verification according to an embodiment of the present invention. FIGS. 6 to 8 are examples of determining the same person between an active user and a registered user according to an embodiment of the present invention. FIG. 9 is an example in which motion changes of a virtual human are performed according to the detection location of an active user according to an embodiment of the present invention. FIGS. 10 and 11 are examples of providing interaction content using a virtual human in a differentiated manner according to the identity verification result of an active user, in accordance with an embodiment of the present invention. FIG. 12 is an example of the internal configuration of a computing device according to an embodiment of the present invention. Hereinafter, various embodiments and/or aspects are disclosed with reference to the drawings. For illustrative purposes, numerous specific details are disclosed in the following description to aid in a general understanding of one or more aspects. However, it will also be recognized by those skilled in the art that these aspects may be practiced without such specific details. The following description and the accompanying drawings describe specific exemplary aspects of one or more aspects in detail. However, these aspects are exemplary, and some of the various methods in the principles of the various aspects may be used, and the description is intended to include all such aspects and their equivalents. As used herein, terms such as "examples," "examples," "aspects," "examples," etc., may not be interpreted as implying that any aspect or design described is better or more advantageous than other aspects or designs. Additionally, the terms “comprising” and/or “comprising” should be understood to mean that the relevant feature and/or component is present, but not to exclude the presence or addition of one or more other features, components and/or groups thereof. Additionally, terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but said components are not limited by said terms. Such terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be named the second component, and similarly, the second component may be named the first compone