MLSQ: A Multimodal-based System for Learning Material Summarization and Question Generation 


Vol. 14,  No. 12, pp. 1004-1015, Dec.  2025
10.3745/TKIPS.2025.14.12.1004


PDF
  Abstract

While the proliferation of digital learning environments has increased the use of diverse multimedia materials, this often leads to passive learning. Existing text-based automatic question generation technologies are insufficient to overcome this limitation. Therefore, this study proposes an AI-based system (MLSQ) that integrates and analyzes video lectures and written materials. This system precisely fuses text data extracted via OCR and STT, along with handwriting information detected from the lecture video as a key emphasis point, based on temporal information and inputs it into a Large Language Model (LLM) to summarize learning content and automatically generate both short-answer and multiple-choice questions. Performance evaluation results showed that the proposed multimodal fusion method demonstrated improved performance over single-modal approaches, with a maximum increase of 3.4%p in BLEU score and 3.1%p in ROUGE-L score. Furthermore, in a 5-point Mean Opinion Score (MOS) user evaluation, the system demonstrated its educational practicality and effectiveness by achieving high scores above 4.0 in all categories, with ‘Speech Conversion Reliability’ and ‘Learning Suitability of Question Generation’ receiving an average of 4.33.

  Statistics
  Cite this article

[IEEE Style]

G. Yu, S. Lee, J. Ahn, M. Woo, S. Kim, J. Lee, H. Choi, Y. Kim, W. Lee, "MLSQ: A Multimodal-based System for Learning Material Summarization and Question Generation," The Transactions of the Korea Information Processing Society, vol. 14, no. 12, pp. 1004-1015, 2025. DOI: 10.3745/TKIPS.2025.14.12.1004.

[ACM Style]

Geonwoo Yu, Sangyoon Lee, Jinyoung Ahn, Minha Woo, Sugyeong Kim, Jungoo Lee, Hyeonwoo Choi, Yaeran Kim, and Woonghee Lee. 2025. MLSQ: A Multimodal-based System for Learning Material Summarization and Question Generation. The Transactions of the Korea Information Processing Society, 14, 12, (2025), 1004-1015. DOI: 10.3745/TKIPS.2025.14.12.1004.

Forms

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)

Old Journals

Indexing

All publications of TKIPS is indexed in DOI, EBSCO, Google Scholar, Crossref, and CrossCheck.

TKIPS is also selected as the Journal for Accreditation by NRF (National Research Foundation of Korea).

Related Journals