Silhouette-to-Contour Registration: Aligning Intraoral Scan Models with Cephalometric Radiographs

TL;DR

DentalSCR tackles the challenging problem of registering 3D dental IOS models to lateral cephalometric radiographs under real-world imaging conditions. The method builds a reproducible cross-arch reference frame (UMDA), renders radiograph-like coronal-axis projections with a surface-based DRR, and performs silhouette-to-contour registration using a symmetric bidirectional Chamfer distance in a coarse-to-fine optimization. It stabilizes initialization, preserves clinically relevant magnification and distortion, and achieves high-fidelity 3D–2D alignment that is interpretable and robust across cases, as demonstrated on 34 expert-annotated clinical datasets with marked improvements in landmark accuracy and contour metrics. The work offers a practical pipeline for orthodontic diagnostics that integrates geometric priors with contour guidance, paving the way for clinically deployable 3D–2D registration tools and potential extensions to multi-view and uncertainty-aware analyses.

Abstract

Reliable 3D-2D alignment between intraoral scan (IOS) models and lateral cephalometric radiographs is critical for orthodontic diagnosis, yet conventional intensity-driven registration methods struggle under real clinical conditions, where cephalograms exhibit projective magnification, geometric distortion, low-contrast dental crowns, and acquisition-dependent variation. These factors hinder the stability of appearance-based similarity metrics and often lead to convergence failures or anatomically implausible alignments. To address these limitations, we propose DentalSCR, a pose-stable, contour-guided framework for accurate and interpretable silhouette-to-contour registration. Our method first constructs a U-Midline Dental Axis (UMDA) to establish a unified cross-arch anatomical coordinate system, thereby stabilizing initialization and standardizing projection geometry across cases. Using this reference frame, we generate radiograph-like projections via a surface-based DRR formulation with coronal-axis perspective and Gaussian splatting, which preserves clinical source-object-detector magnification and emphasizes external silhouettes. Registration is then formulated as a 2D similarity transform optimized with a symmetric bidirectional Chamfer distance under a hierarchical coarse-to-fine schedule, enabling both large capture range and subpixel-level contour agreement. We evaluate DentalSCR on 34 expert-annotated clinical cases. Experimental results demonstrate substantial reductions in landmark error-particularly at posterior teeth-tighter dispersion on the lower jaw, and low Chamfer and controlled Hausdorff distances at the curve level. These findings indicate that DentalSCR robustly handles real-world cephalograms and delivers high-fidelity, clinically inspectable 3D--2D alignment, outperforming conventional baselines.

Figures (5)

• Figure 1: Overall of the proposed framework. The system first constructs a unified U-Midline Dental Axis (UMDA) to establish a reproducible anatomical coordinate frame shared by both arches. Using this reference, intraoral scan (IOS) meshes are projected into radiograph-like views through a coronal-axis surface-based DRR with Gaussian splatting, preserving clinical magnification and geometric distortion. The resulting silhouette is then aligned to expert-annotated cephalometric contours via a 2D similarity transform optimized with a symmetric bidirectional Chamfer distance under a hierarchical coarse-to-fine schedule. This pipeline yields stable initialization, radiographically consistent projections, and accurate silhouette-to-contour registration.
• Figure 2: Contours on the cephalometric radiograph. Upper arch shown in red and lower arch in green; each arch is formed by concatenating tooth-level labial/buccal polylines in anatomical order.
• Figure 3: Two-dimensional view displaying the corresponding anatomical landmarks on a lateral cephalometric radiograph. Landmark definitions and their corresponding codes are detailed in Table \ref{['tab:landmark_code_map']}.
• Figure 4: Three-dimensional visualization of the upper jaw (UR, blue) and lower jaw (LR, red) dental models with overlaid anatomical landmarks (red points). The anatomical significance of the corresponding landmarks is detailed in Table \ref{['tab:landmark_code_map']}.
• Figure 5: Qualitative comparison on representative cases. Columns: DentalSCR (ours), Bounding Box, Single-Stage. DentalSCR produces tighter IOS model$\rightarrow$CR overlays with fewer gaps, especially at incisors/canines and posterior cusps.