Proposal of quantitative index of skin texture by the image processing and its practical application

Hiroshi Kobayashi, Takuya Hashimoto, Kazuhiro Yamazaki, Yoshikazu Hirai

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

There are many cosmetic companies and visual appreciation is undertaken by human expert though, skin texture has not been investigated quantitatively and automatically so far. In this study, image processing method for analyzing skin texture is proposed. Since sulci is very important for expressing skin texture, we apply cross binarization method for extracting sulci and by matching short lines onto sulci extracted, sulci is expressed by short lines. By analyzing short lines in terms of number, direction, and so on, quantitative indices are suggested for displaying feature of sulci. We then investigate whether it is available or not to predict visual appreciation score by quantitative indices. By using multiple linear regression analysis, we find that visual appreciation score is predictable in high rate by using quantitative indices we have proposed. This method comes into practical use since January 2009.

Original languageEnglish
Pages (from-to)922-929
Number of pages8
JournalNihon Kikai Gakkai Ronbunshu, C Hen/Transactions of the Japan Society of Mechanical Engineers, Part C
Volume76
Issue number764
DOIs
Publication statusPublished - Apr 2010

    Fingerprint

Keywords

  • Image processing
  • Multiple linear regression analysis
  • Quantitative index
  • Short line matching
  • Skin texture
  • Visual appreciation

Cite this