JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Using optical coherence tomography to assess luster of pearls: technique suitability and insights

  • 0School of Information and Electronic Engineering, Zhejiang University of Science and Technology, Hangzhou, 310023, Zhejiang, China. zybuaa@163.com.
Scientific Reports +

|

May 15, 2024

|

May 15, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

Optical coherence tomography (OCT) analyzes pearl texture for grading. This non-destructive method accurately predicts pearl luster using speckle patterns and machine learning.

Area Of Science

  • Materials Science
  • Optical Engineering
  • Gemology

Background

  • Pearl grading relies on luster, a key quality indicator.
  • Current grading methods can be subjective, time-consuming, or destructive.
  • Developing objective, rapid, and non-destructive grading techniques is essential.

Purpose Of The Study

  • To introduce Optical Coherence Tomography (OCT) as a tool for predicting pearl luster.
  • To investigate the correlation between OCT-derived texture features and pearl luster grades.
  • To establish a fast, non-destructive, and low-cost method for pearl luster grading.

Main Methods

  • Image processing techniques including background removal, flattening, and segmentation were applied to OCT images.
  • Seven texture features (CSAC, FD, Gabor, GLCM, HOG, LAWS, LBP) were extracted from the speckle patterns.
  • Support Vector Machines (SVM) and Random Forest Classifier (RFC) were employed to classify luster grades based on texture features.

Main Results

  • Texture features extracted from OCT images effectively characterized pearl speckle patterns.
  • Machine learning models (SVM and RFC) achieved high performance metrics (precision, recall, F1-score, accuracy > 0.9).
  • Accurate luster classification was maintained even after dimension reduction of features.

Conclusions

  • Texture analysis of OCT images provides a viable method for pearl luster grading.
  • OCT offers a fast, non-destructive, and accurate approach to assessing pearl quality.
  • Speckle pattern analysis via OCT can be reliably used to classify pearl luster.