Improving feature extraction in fingerprint of medicinal herbs via wavelet transform and fractal technique

Jianwei Du*, Yan Tang Yaun, Jingrong Wang, Zhi Hong JIANG

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)

Abstract

In this paper, hybrid features combing both high-frequency and low-frequency components of wavelet transform are applied to fingerprint of medicinal herbs. Through the fingerprints of medicinal herbs by wavelet transform and the fractal dimensions, 13 features are obtained, which are called fractal-wavelet features. In this new approach, the information of each sample can be acquired to the maximum degree. These novel hybrid features have been applied to recognition of the different types of ginseng. Experiments have been conducted, and the result of recognition can match the real situation. Experiments indicate this method is better than the traditional ones.

Original languageEnglish
Pages (from-to)476-483
Number of pages8
JournalJournal of Chemometrics
Volume20
Issue number11-12
DOIs
Publication statusPublished - Nov 2006

Scopus Subject Areas

  • Analytical Chemistry
  • Applied Mathematics

User-Defined Keywords

  • Fingerprint of medicinal herbs
  • Fractal dimensions
  • Fractal-wavelet feature
  • Type recognition of ginseng
  • Wavelet transform

Fingerprint

Dive into the research topics of 'Improving feature extraction in fingerprint of medicinal herbs via wavelet transform and fractal technique'. Together they form a unique fingerprint.

Cite this