WSEAS Transactions on Information Science and Applications
Print ISSN: 1790-0832, E-ISSN: 2224-3402
Volume 23, 2026
Automatic Detection of Caries in Dental X-Ray Images
Authors: , , ,
Abstract: This paper presents a computer-aided pipeline for automatic caries detection in periapical dental radiographs that combines image processing with a bounded-speed level-set evolution. After denoising and contrast cleanup, Otsu thresholding, and morphological opening/closing, remove spurious structures. An integral projection isolates individual teeth, enabling a Morphological Region-Based Initial Contour (MRBIC) to seed a level-set segmentation driven by a normalized Signed Force Function (SFF). The SFF re-estimates inside/outside means at each iteration to self-calibrate to local intensity statistics, while motion filtering regularizes the evolving front; restricting evolution to a per-tooth window improves robustness near restorations and gaps. The scheme is CFL-stable, has linear per-iteration complexity in the active window, and reduces leakage and false edges compared with baseline methods. Using expert dentist annotations as ground truth, the system was evaluated on 120 periapical radiographs, achieving 90% tooth-level segmentation accuracy and 90% caries detection accuracy. These results indicate the scheme supports dentists with fast, reliable second-opinion screening.
Search Articles
Pages: 80-89
DOI: 10.37394/23209.2026.23.7