Related Experiment Video
Updated: Aug 9, 2025
05:42
Detection and Removal of Tooth-Colored Composite Resin Using the Fluorescence-Aided Identification Technique
Published on: July 27, 2022
3.7K
Fuzzy K-Nearest Neighbor Based Dental Fluorosis Classification Using Multi-Prototype Unsupervised Possibilistic Fuzzy
Ritipong Wongkhuenkaew1,2, Sansanee Auephanwiriyakul3, Nipon Theera-Umpon4
1Department of Computer Engineering, Faculty of Engineering, Chiang Mai University, Chiang Mai 50200, Thailand.
International Journal of Environmental Research and Public Health
|February 25, 2023
Summary
This study introduces an automated system to detect dental fluorosis severity using image analysis. The method accurately segments teeth and classifies fluorosis stages, improving upon previous techniques.
Area of Science:
- Biomedical Engineering
- Dental Imaging
- Computer Vision
Background:
- Dental fluorosis is a common condition caused by excessive fluoride intake during tooth development.
- It manifests as enamel discoloration, ranging from white to dark brown stains.
- Accurate screening of fluorosis severity is crucial for dental professionals.
Purpose of the Study:
- To develop an automated, image-based system for segmenting and classifying dental fluorosis.
- To assist dentists in objectively screening the severity of dental fluorosis.
- To improve the accuracy and efficiency of fluorosis diagnosis.
Main Methods:
- Utilized unsupervised possibilistic fuzzy clustering (UPFC) on RGB and HIS color spaces to categorize tooth enamel.
- Employed the fuzzy k-nearest neighbor method for feature classification, optimizing cluster numbers with the cuckoo search algorithm.
- Developed a classification rule based on segmented pixel proportions to determine fluorosis stages (Normal, Stage 1, Stage 2, Stage 3).
Main Results:
- Achieved an average pixel accuracy of 92.24% for the segmented binary tooth mask.
- Obtained an average pixel accuracy of 79.46% for segmenting teeth into white-yellow, opaque, and brown pixels.
- Correctly classified four fluorosis stages in 86 out of 128 blind test images, showing a 13.33% improvement over prior methods.
Conclusions:
- The proposed automated system effectively segments dental fluorosis and classifies its severity.
- The image-based approach offers a promising tool for objective and efficient dental fluorosis screening.
- This method demonstrates significant potential for clinical application in dental diagnostics.
