본문 바로가기
  • Home

Culture mining with deep learning algorithms

  • The Japanese Language Association of Korea
  • Abbr : JLAK
  • 2022, (73), pp.93-106
  • DOI : 10.14817/jlak.2022.73.93
  • Publisher : The Japanese Language Association Of Korea
  • Research Area : Humanities > Japanese Language and Literature
  • Received : June 30, 2022
  • Accepted : August 29, 2022
  • Published : September 20, 2022

LEE JUNSEO 1 Sangsoon Lim 1

1성결대학교

Accredited

ABSTRACT

This paper is a paper conceived to improve the Cultural Image Frame Network (CIFN) by utilizing deep learning-based image learning datasets that have recently received much attention in major research fields such as computer vision and Natural Language Processing (NLP). In particular, CNN, which uses convolutional filters for images to calculate quickly and considers the entire image, including specific objects as well as backgrounds, is a very suitable algorithm for extracting cultural elements that constitute the cultural image frame of this culture mining study. In addition, by utilizing images in the form of refined images verified with deep learning experimental and test datasets, the limitations of existing research, such as (1) reliability of tagging information, (2) inaccuracy of the segmentation method, and (3) redundancy of images, can contribute to more sophisticated research.

Citation status

* References for papers published after 2023 are currently being built.

This paper was written with support from the National Research Foundation of Korea.