Comparative Evaluation of Machine Learning Algorithms for Spectrophotometric Dental Shade Classification

Pei-ting Chung

Journal of Sustainability, Policy, and Practice, 2026, vol. 2, issue 1, 204-214

Abstract: Accurate dental shade matching is critical for achieving optimal esthetic outcomes in restorative dentistry. This study presents a comparative evaluation of machine learning algorithms for spectrophotometric dental shade classification, focusing on Support Vector Machine, Random Forest, and Extreme Learning Machine approaches. Spectral reflectance data from 1,280 standardized dental composite specimens (as controlled surrogates for shade-guide categories) across 16 VITA Classical shades were collected using a calibrated spectrophotometer. Feature extraction methods, including CIELAB coordinates, spectral coefficients, and principal component analysis, were systematically compared. Experimental results demonstrate that the Extreme Learning Machine achieved the highest classification accuracy of 97.8%; its mean ΔE00 was 1.42, and 89.3% of predictions fell below the clinical acceptability threshold of ΔE00 = 1.8, with a b coordinate RMSE of 2.14. Random Forest demonstrated superior robustness in edge-shade classification, achieving 94.2% accuracy. The findings provide practical guidance for selecting algorithms in industrial dental shade-matching applications.

Keywords: dental shade matching; machine learning; spectrophotometric classification; CIEDE2000 (search for similar items in EconPapers)
Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://pinnaclepubs.com/index.php/jspp/article/view/583/569 (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:dba:jsppaa:v:2:y:2026:i:1:p:204-214

Access Statistics for this article

More articles in Journal of Sustainability, Policy, and Practice from Pinnacle Academic Press
Bibliographic data for series maintained by Joseph Clark ().