@article{ART003349315},
author={Ho-Sung Choi and Se-hoon Park},
title={An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking},
journal={Journal of The Korea Society of Computer and Information},
issn={1598-849X},
year={2026},
volume={31},
number={6},
pages={125-131}
TY - JOUR
AU - Ho-Sung Choi
AU - Se-hoon Park
TI - An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking
JO - Journal of The Korea Society of Computer and Information
PY - 2026
VL - 31
IS - 6
PB - The Korean Society Of Computer And Information
SP - 125
EP - 131
SN - 1598-849X
AB - In this paper, we propose a 2-stage hybrid ensemble architecture that combines an AI probability map for candidate band localization with traditional curvature tensor extremum tracking to precisely detect the margin line of dental prostheses. To address the black-box nature and hallucination risks of existing end-to-end deep learning methods, our approach ensures the exclusion of deep learning in the final boundary decision.
Instead, the final boundary is determined by tracking curvature tensor extrema optimized via a snake optimization algorithm. We also integrated a reverse hybrid option for atypical preparations, a Human-in-the-Loop interface for snapping-based user correction, and an independent FDI module for tooth classification to prevent error propagation. Quantitative evaluation on 182 clinical jaw scans using Hausdorff and Chamfer distances demonstrated that the proposed hybrid approach achieves statistically significant superiority (P<0.05) and robustness compared to existing pure algorithm-based commercial software.
KW - Dental Margin Line;Curvature Tensor;Hybrid Ensemble;AI Probability Map;Human-in-the-Loop
DO -
UR -
ER -
Ho-Sung Choi and Se-hoon Park. (2026). An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking. Journal of The Korea Society of Computer and Information, 31(6), 125-131.
Ho-Sung Choi and Se-hoon Park. 2026, "An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking", Journal of The Korea Society of Computer and Information, vol.31, no.6 pp.125-131.
Ho-Sung Choi, Se-hoon Park "An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking" Journal of The Korea Society of Computer and Information 31.6 pp.125-131 (2026) : 125.
Ho-Sung Choi, Se-hoon Park. An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking. 2026; 31(6), 125-131.
Ho-Sung Choi and Se-hoon Park. "An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking" Journal of The Korea Society of Computer and Information 31, no.6 (2026) : 125-131.
Ho-Sung Choi; Se-hoon Park. An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking. Journal of The Korea Society of Computer and Information, 31(6), 125-131.
Ho-Sung Choi; Se-hoon Park. An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking. Journal of The Korea Society of Computer and Information. 2026; 31(6) 125-131.
Ho-Sung Choi, Se-hoon Park. An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking. 2026; 31(6), 125-131.
Ho-Sung Choi and Se-hoon Park. "An Automatic Detection Method of Dental Prosthesis Margin Line by Hybrid Ensemble of AI Probability Map and Curvature Tensor Extremum Tracking" Journal of The Korea Society of Computer and Information 31, no.6 (2026) : 125-131.
