Automatic color pattern recognition of multispectral printed fabric images

Jie Zhang, Pengpeng Yao, Hochung Wu, John H. Xin (Corresponding Author)

Research output: Journal article publicationJournal articleAcademic researchpeer-review

2 Citations (Scopus)


Printed fabrics often have rich colors and variable patterns in different sizes and shapes, which make it difficult to achieve accurate pattern recognition and color measurement using traditional spectrophotometers and digital cameras. This paper develops a grid-based density peaks clustering (GDPC) algorithm to automatically recognize patterns and extract colors of multispectral images of printed fabrics. The multispectral images captured by a self-developed multispectral imaging system is firstly converted into color images in CIELAB color space and three principal components are calculated by applying principal component analysis to reduce the dimensions of the multispectral images. During the multispectral image processing, the noise pixels are removed by calculating the local stability of each pixel, and then the remaining stable pixels are clustered using proposed GDPC algorithm based on three CIELAB color channels and three principal components. Compared with widely-used color clustering algorithms, the proposed GDPC algorithm can recognize the color patterns from more intricate multispectral printed fabric images with higher accuracy and less computational time.

Original languageEnglish
Pages (from-to)2747-2763
Number of pages17
JournalJournal of Intelligent Manufacturing
Issue number6
Publication statusPublished - Aug 2023


  • Grid-based density peaks clustering algorithm
  • Multispectral imaging method
  • Pattern recognition
  • Printed fabrics

ASJC Scopus subject areas

  • Software
  • Industrial and Manufacturing Engineering
  • Artificial Intelligence


Dive into the research topics of 'Automatic color pattern recognition of multispectral printed fabric images'. Together they form a unique fingerprint.

Cite this