Unsupervised clustering analyzes of features extraction for a caries computer-assisted diagnosis using dental fluorescence images

Michel Bessani; Mardoqueu M. da Costa; Emery C. C. C. Lins; Carlos D. Maciel

doi:10.1117/12.2038633

17 February 2014 Unsupervised clustering analyses of features extraction for a caries computer-assisted diagnosis using dental fluorescence images

Michel Bessani, Mardoqueu M. da Costa, Emery C. C. C. Lins, Carlos D. Maciel

Author Affiliations +

Proceedings Volume 8937, Multimodal Biomedical Imaging IX; 89370U (2014) https://doi.org/10.1117/12.2038633
Event: SPIE BiOS, 2014, San Francisco, California, United States

Abstract

Computer-assisted diagnoses (CAD) are performed by systems with embedded knowledge. These systems work as a second opinion to the physician and use patient data to infer diagnoses for health problems. Caries is the most common oral disease and directly affects both individuals and the society. Here we propose the use of dental fluorescence images as input of a caries computer-assisted diagnosis. We use texture descriptors together with statistical pattern recognition techniques to measure the descriptors performance for the caries classification task. The data set consists of 64 fluorescence images of in vitro healthy and carious teeth including different surfaces and lesions already diagnosed by an expert. The texture feature extraction was performed on fluorescence images using RGB and YCbCr color spaces, which generated 35 different descriptors for each sample. Principal components analysis was performed for the data interpretation and dimensionality reduction. Finally, unsupervised clustering was employed for the analysis of the relation between the output labeling and the diagnosis of the expert. The PCA result showed a high correlation between the extracted features; seven components were sufficient to represent 91.9% of the original feature vectors information. The unsupervised clustering output was compared with the expert classification resulting in an accuracy of 96.88%. The results show the high accuracy of the proposed approach in identifying carious and non-carious teeth. Therefore, the development of a CAD system for caries using such an approach appears to be promising.

Citation Download Citation

Michel Bessani, Mardoqueu M. da Costa, Emery C. C. C. Lins, and Carlos D. Maciel "Unsupervised clustering analyses of features extraction for a caries computer-assisted diagnosis using dental fluorescence images", Proc. SPIE 8937, Multimodal Biomedical Imaging IX, 89370U (17 February 2014); https://doi.org/10.1117/12.2038633

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
9 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Luminescence

Dental caries

Feature extraction

Principal component analysis

Optical filters

RGB color model

Teeth

Show All Keywords

Keywords/Phrases

Search In:

Publication Years