Home > Volume 16 Number 3 (2026) >
IJML 2026 Vol.16(3): 17-24
DOI: 10.18178/ijml.2026.16.3.1187

CLIO: Climate-aware and Legacy-based Intelligent Outfit Recommendation System

Rio Kawasaki and Keiichi Takahashi *
Department of Information and Computer Science, Faculty of Humanity-Oriented Science and Engineering, Kindai University, Iizuka, Japan
Email: 2211140043v@ed.fuk.kindai.ac.jp (R.K.); ktakahas@fuk.kindai.ac.jp (K.T.)
*Corresponding author

Manuscript received July 5, 2026; revised August 12, 2026; accepted September 3, 2026; published September 22, 2026

Abstract—Choosing an appropriate outfit for the day is a common daily challenge. While many fashion applications exist, they typically recommend items from external catalogs without considering the user's own wardrobe or the current weather conditions. Furthermore, outfit logging applications record past outfits but offer no recommendations based on that history. This paper presents CLIO, a Climate-aware and Legacy-based Intelligent Outfit Recommendation System that addresses these gaps. CLIO recommends outfits exclusively from the user's personal wardrobe by scoring each registered outfit according to the similarity between the current weather and the weather on the day the outfit was worn. CLIO also prevents unintentional repetitive wearing by excluding recently worn or visually similar outfits from recommendations for a fixed period. Clothing regions are extracted using an instance segmentation model trained on a custom-annotated clothing dataset, and outfit similarity is computed via cosine similarity of deep convolutional feature vectors. A five-day deployment with a single user confirmed that both algorithmic components run end-to-end under real daily-use conditions. This paper is a systems contribution: we report the design, implementation, and proof-of-concept deployment of the pipeline, and we neither claim nor evaluate the effectiveness of the recommendations it produces. We specify the controlled study that establishing effectiveness would require, together with the quantitative evaluation the perception components require and the directions in which the weather model and preference handling should be extended.

Keywords—outfit recommendation, weather-aware recommendation, instance segmentation, personal wardrobe

[PDF]

Cite: Rio Kawasaki and Keiichi Takahashi , "CLIO: Climate-aware and Legacy-based Intelligent Outfit Recommendation System," International Journal of Machine Learning vol. 16, no. 3, pp. 17-24, 2026.

Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).

PREVIOUS PAPER
First page
NEXT PAPER
Last page

General Information

  • E-ISSN: 2972-368X
  • Abbreviated Title: Int. J. Mach. Learn.
  • DOI: 10.18178/IJML
  • Frequency: Quarterly
  • Average Days to Accept: 68 days
  • Acceptance Rate: 27%
  • Editor-in-Chief: Dr. Lin Huang
  • Executive Editor:  Ms. Cherry L. Chen
  • DOI: 10.18178/IJML
  • Abstracing/Indexing: Google Scholar, Crossref, ProQuest, Electronic Journals LibraryCNKI.
  • E-mail: editor@ijml.org
  • APC: 500USD



Article Metrics in Dimensions