Search

KR-20260062328-A - Device and method for real-time customized work recommendation and author matching based on Korean-specific emotions

KR20260062328AKR 20260062328 AKR20260062328 AKR 20260062328AKR-20260062328-A

Abstract

The present invention relates to an apparatus and method for real-time customized artwork recommendation and artist matching based on Korean-specific emotions, which enables user-customized painting artwork recommendations using artwork information obtained through user emotion analysis results and artist emotion labeling. The apparatus comprises an electronic device including at least one processor for real-time customized artwork recommendation and artist matching based on Korean-specific emotions, comprising: a user emotion information processing unit that collects user facial expression data and analyzes user emotions to extract user final emotion data; an artist and artwork emotion information processing unit that collects artwork information and analyzes artwork emotions to extract artwork emotion information; and an artwork matching and recommendation processing unit that measures the similarity between user final emotion data and artwork emotion information to perform user-artwork matching and make recommendations, and collects user evaluation data based on recommendations.

Inventors

  • 신원섭

Assignees

  • 신원섭

Dates

Publication Date
20260507
Application Date
20241029

Claims (13)

  1. An electronic device comprising at least one processor for real-time customized artwork recommendation and author matching based on Korean-specific emotions, A user emotion information processing unit that collects user facial expression data, analyzes user emotions, and extracts final user emotion data; An author and a work emotion information processing unit that collects work information, analyzes work emotions, and extracts work emotion information; A device for real-time customized artwork recommendation and author matching based on Korean-specific emotions, characterized by including: an artwork matching and recommendation processing unit that measures the similarity between user final emotion data and artwork emotion information to perform user-artwork matching and make recommendations, and collects user evaluation data based on recommendations.
  2. In claim 1, reinforcement learning is performed using user evaluation data collected in the work matching and recommendation processing unit, A device for real-time customized artwork recommendation and author matching based on Korean-specific emotions, characterized by enhancing user-artwork matching and recommendation algorithms and building a dataset.
  3. In claim 1, the user emotion information processing unit, A user expression data collection unit that acquires images of a user's face captured in high-speed burst mode, detects a face region in the collected data, removes noise, and collects user expression data; A user emotion analysis unit that analyzes micro-expression data using a deep learning model, classifies emotions by recognizing basic and complex emotions based on the emotions most frequently felt by Koreans, and assigns weights based on the ranking of emotions for which one wishes to receive empathy from others, and A device for real-time customized work recommendation and author matching based on Korean-specific emotions, characterized by including a user final emotion data extraction unit that extracts the user's final emotion data by summing the weights assigned to the emotions.
  4. In Clause 3, the user facial expression data collection unit, A micro-expression data collection unit that acquires images of the user's face captured in high-speed burst mode through the user's mobile phone camera, and A device for real-time customized artwork recommendation and author matching based on Korean-specific emotions, characterized by including a data preprocessing unit that detects face regions in collected data and performs a preprocessing process to remove noise.
  5. In Paragraph 3, the user emotion analysis unit, An emotion recognition unit that recognizes basic and complex emotions tailored to the emotions most frequently felt by Koreans by analyzing micro-facial expression data using a deep learning model, and An emotion classification unit that classifies emotions recognized by the emotion recognition unit into positive, negative, and neutral clusters, and A device for real-time customized work recommendation and author matching based on Korean-specific emotions, characterized by including a weighting unit that assigns weights to the emotion ranking of the emotion one wants to receive empathy from others (1st rank: 0.3), complex emotion (2nd rank: 0.2), and positive emotion cluster (3rd rank: 0.1).
  6. In Article 1, the author and work appraisal information processing unit, A work information collection unit that processes work uploads and supports the author in directly labeling subjective feelings about the work, and A work emotion analysis unit that analyzes the correlation between color information of the work image and labeled emotions to cluster emotions and assigns weights to each emotion cluster based on the ranking of emotions desired to receive empathy from others, and A device for real-time customized artwork recommendation and author matching based on Korean-specific emotions, characterized by including an artwork emotion information extraction unit that extracts the main emotions of an artwork by reflecting weights and feedback assigned to each cluster.
  7. In Article 6, the Work Information Collection Department, A work upload section that supports uploading work images, and A device for real-time customized work recommendation and author matching based on emotions specialized for Koreans, characterized by including an author labeling section that supports the author in directly labeling subjective emotions regarding the work.
  8. In Clause 6, the Artwork Appraisal Analysis Department, An image analysis unit that analyzes color information of images based on color psychology and analyzes the correlation with labeled emotions, and An emotion clustering unit that generates emotion clusters by synthesizing author labeling data and image analysis results to group works with similar emotions, and A device for real-time customized work recommendation and author matching based on Korean-specific emotions, characterized by including a weighting unit that assigns weights to each emotion cluster according to the ranking of emotions desired to receive empathy from others.
  9. In Clause 6, the artwork appraisal information extraction unit, A final emotion derivation unit that extracts the main emotions of the work in a set number based on weights and feedback data assigned to each cluster, and A device for real-time customized artwork recommendation and artist matching based on Korean-specific emotions, characterized by including a learning and dataset construction unit that performs reinforcement learning using extracted key emotions and artwork information, stores the results in a database, and enables the use of the data for artwork matching.
  10. In claim 1, the work matching and recommendation processing unit, A user-work matching unit that measures the similarity between the extracted user's emotion information and the work emotion information of the work dataset to match and recommend the work with the highest similarity, and A device for real-time customized work recommendation and author matching based on Korean-specific emotions, characterized by including a recommendation result reflection unit that collects user satisfaction regarding recommendation results for each work, stores user feedback in a dataset, and supports reinforcement learning to improve user satisfaction.
  11. In Clause 10, the user-work matching unit, A similarity measuring unit that measures the similarity between the user's emotions and the work dataset, and A matching unit that matches the work with the highest similarity based on the similarity measurement results from the similarity measurement unit, and A device for real-time customized artwork recommendation and artist matching based on Korean-specific emotions, characterized by including a recommendation section that displays artwork, atelier location information, and artist information on a curation map corresponding to the number of settings matched in order of shortest distance based on the user's location.
  12. In Clause 10, the recommendation result reflection section is, A feedback collection unit that collects user satisfaction regarding recommendation results for each work, and A device for real-time customized work recommendation and author matching based on Korean-specific emotions, characterized by including a feedback analysis unit that analyzes user feedback data, stores the analysis results in a dataset, and supports the improvement of user satisfaction through reinforcement learning.
  13. An operation for real-time customized work recommendation and author matching based on Korean-specific emotions is performed in an electronic device comprising at least one processor, and A step of collecting user facial expression data, analyzing user emotions, and extracting final user emotion data; A step of collecting artwork information and analyzing artwork emotions to extract artwork emotion information; A step of measuring the similarity between user final emotion data and artwork emotion information to perform user-artwork matching and make recommendations, and collecting user evaluation data based on recommendations; A method for real-time customized work recommendation and author matching based on Korean-specific emotions, characterized by including the step of reinforcing a user-work matching and recommendation algorithm and constructing a dataset by performing reinforcement learning using collected user evaluation data.

Description

Device and method for real-time customized work recommendation and author matching based on Korean-specific emotions The present invention relates to the recommendation of artworks, and specifically to an apparatus and method for real-time customized artwork recommendation and artist matching based on Korean-specific emotions, which enables user-customized painting artwork recommendations by utilizing artwork information obtained through user emotion analysis results and artist emotion labeling. In the IT field, active research is being conducted on recommendation algorithms that enable the effective selection of content and recommendation systems utilizing them. Recommendation systems recommend items users are likely to be interested in, reducing the time users spend selecting content and simultaneously promoting content consumption. Representative recommendation algorithms used in recommendation systems include collaborative filtering, cluster models, and search-based methods. Meanwhile, the art market consists of artists as producers and private collectors, corporations, and institutions as consumers. The distribution market that connects these producers and consumers is broadly divided into galleries, auction houses, and art fairs. In major offline art markets, transactions center around high-priced works by famous artists. This method of consuming art is gallery-centric, and due to a lack of information accessibility, there are limitations in general users encountering works suitable for them. In such an art market, technologies are being proposed to recommend artworks or artists tailored to the individual tastes and purposes of users. For example, in the prior art, as a method for recommending works and artists, there is a technology that collects user behavior information on a platform where an illustrator's work is posted to analyze the user's preferences, and recommends illustrators and works to the user using the user's preferences and event information. (Republic of Korea Registered Patent No. 10-2553204) Another technology involves analyzing a user's emotional state to recommend works corresponding to that state, and displaying the recommended works using the metaverse. (Republic of Korea Registered Patent No. 10-2617130) Another technology reduces the opportunity cost for art searchers and identifies the types of artworks with high preference by evaluating the most preferred artwork types through an objective process and identifying the user's preferences and tastes regarding artworks. (Republic of Korea Published Patent No. 10-2020-0026433) However, such conventional technologies recommend works based solely on the user's taste, preference, and emotional state, and thus have limitations in recommending works suitable for the user. In particular, despite the increasing demand for personalized curation services, the lack of services that understand the subtle emotions of Koreans and recommend suitable artworks limits the revitalization of the culture and arts market and the creation of diverse opportunities for consuming art. Therefore, there is a need to develop new technology that considers both the user's emotional state and artwork emotional information to enable efficient artwork recommendations and artist matching based on emotional analysis results. FIG. 1 is a configuration diagram of a device for real-time customized artwork recommendation and author matching based on Korean-specific emotions according to the present invention. Figure 2 is a detailed configuration diagram of the user emotion information processing unit. Figure 3 is a detailed configuration diagram of the author and artwork appraisal information processing unit. Figure 4 is a detailed configuration diagram of the artwork matching and recommendation processing unit. FIG. 5 is an overall detailed configuration diagram of a device for real-time customized artwork recommendation and author matching based on Korean-specific emotions according to the present invention. FIG. 6 is a flowchart illustrating a method for real-time customized artwork recommendation and author matching based on Korean-specific emotions according to the present invention. Hereinafter, preferred embodiments of the apparatus and method for real-time customized artwork recommendation and author matching based on Korean-specific emotions according to the present invention will be described in detail as follows. The features and advantages of the apparatus and method for real-time customized artwork recommendation and author matching based on Korean-specific emotions according to the present invention will become apparent through the detailed description of each embodiment below. FIG. 1 is a configuration diagram of a device for real-time customized artwork recommendation and author matching based on Korean-specific emotions according to the present invention. The terms used in this disclosure have been selected to be as widely used and general as possible