OrthoAI v2: From Single-Agent Segmentation to Dual-Agent Treatment Planning for Clear Aligners

Lansiaux Edouard; Leman Margaux

cs.CVMar 10, 2026

OrthoAI v2: From Single-Agent Segmentation to Dual-Agent Treatment Planning for Clear Aligners

Authors: Lansiaux Edouard, Leman Margaux

Organizations: STaR-AI Research Group, Emergency Department, Lille University Hospital

Abstract

We present OrthoAI v2, the second iteration of our open-source pipeline for AI-assisted orthodontic treatment planning with clear aligners, substantially extending the single-agent framework previously introduced. The first version established a proof-of-concept based on Dynamic Graph Convolutional Neural Networks (\dgcnn{}) for tooth segmentation but was limited to per-tooth centroid extraction, lacked landmark-level precision, and produced a scalar quality score without staging simulation. \vtwo{} addresses all three limitations through three principal contributions: (i)~~a second agent adopting the Conditioned Heatmap Regression Methodology (\charm{})~~\cite{rodriguez2025charm} for direct, segmentation-free dental landmark detection, fused with Agent~1 via a confidence-weighted orchestrator in three modes (parallel, sequential, single-agent); (ii)~a composite six-category biomechanical scoring model (biomechanics $\times$ 0.30 + staging $\times$ 0.20 + attachments $\times$ 0.15 + IPR $\times$ 0.10 + occlusion $\times$ 0.10 + predictability $\times$ 0.15) replacing the binary pass/fail check of v1; (iii)~a multi-frame treatment simulator generating $F = A \times r$ temporally coherent 6-DoF tooth trajectories via SLERP interpolation and evidence-based staging rules, enabling ClinCheck 4D visualisation. On a synthetic benchmark of 200 crowding scenarios, the parallel ensemble of OrthoAI v2 reaches a planning quality score of $92.8 \pm 4.1$ vs.\ $76.4 \pm 8.3$ for OrthoAI v1, a $+21\%$ relative gain, while maintaining full CPU deployability ( $4.2 \pm 0.8$ ~s).

Figures & tables

Movement	Incisor	Canine	Premolar	Molar
$T_{x}$ MD (mm)	4.0	3.5	3.5	2.0
$T_{y}$ BL (mm)	2.5	2.5	3.0	2.5
$T_{z}$ intrusion	2.0 mm ( $\eta=0.69$ )
$T_{z}$ extrusion	1.5 mm ( $\eta=0.42$ , push-only)
$R_{x}$ torque ( ∘ )	15	12	10	8
$R_{y}$ tip ( ∘ )	10	10	10	8

Table 1: Biomechanical movement limits per tooth type. Limits from Glaser [ 13 ] , predictability scores from Kravitz et al. [ 14 ] .

System	Quality $\uparrow$	$\sigma\downarrow$	Feas. (%) $\uparrow$	Time (s)
OrthoAI v1 (DGCNN only)	76.4	8.3	78	$2.1\pm 0.4$
v2: DGCNN only	$88.3\pm 5.2$	5.2	78	$3.4\pm 0.6$
v2: CHaRM only	$89.1\pm 4.8$	4.8	81	$0.9\pm 0.1$
v2: Sequential	$91.2\pm 4.3$	4.3	86	$2.1\pm 0.4$
v2: Parallel (ens.)	$\mathbf{92.8\pm 4.1}$	4.1	89	$4.2\pm 0.8$

Table 2: OrthoAI v1 vs. v2 on 200 synthetic crowding scenarios. Quality: planning score (0–100). Feas.: fraction of feasible plans. CPU time on Intel i7-11800H.

† Estimated from centroid localisation error reported in [ 1 ] .
Method	MEDE (mm) $\downarrow$	MSR (%) $\uparrow$	GPU Time (s)
ALIIOS [ 5 ]	2.17	48.0	24.70
TS-MDL [ 4 ]	1.80	59.0	3.55
OrthoAI v1 (centroids) [ 1 ]	$\sim$ 1.80 †	–	2.1
CHaRNet [ 2 ]	1.12	68.5	0.24
Agent 2 v2 (ours)	$1.38\pm 0.21$	$64.2\pm 3.1$	$0.31$

Table 3: Landmark detection performance. SOTA from [ 2 ] . Agent 2 on our synthetic dataset.

Limitation	v1	v2
Centroid-only geometry	Yes	Resolved
Scalar quality score	Yes	Resolved
No staging simulation	Yes	Resolved
No SaaS interface	Yes	Resolved
Synthetic data only	Yes	Ongoing
Automatic movement gen.	No	Planned

Table 4: Limitation comparison: v1 vs. v2.

OrthoAI: A Neurosymbolic Framework for Evidence-Grounded Biomechanical Reasoning in Clear Aligner Orthodontics
Feb 23, 2026Edouard Lansiaux, Margaux Leman, Mehdi Ammi3D Point Cloud Segmentation 3D Medical Image Segmentation
STaR-AI, Emergency Department, Lille University Hospital, 59000, Lille, France · Artificial Intelligence and Data Semantics Laboratory, Paris 8 University, 93200, Saint-Denis, France
MATHENA: Mamba-based Architectural Tooth Hierarchical Estimator and Holistic Evaluation Network for Anatomy
Apr 1, 2026Kyeonghun Kim, Jaehyung Park, Youngung Han +17Dental Imaging Medical Imaging
OUTTA, Seoul, South Korea · Seoul National University, Seoul, South Korea · Sangmyung University, Seoul, South Korea +4
OralAgent: Integrating Reasoning, Tools, and Knowledge for Interactive Dental Image Analysis
Apr 9, 2026Jing Hao, Siyuan Dai, Yongxin Zhang +13Medical Image Analysis Dental Imaging
Faculty of Dentistry, the University of Hongkong, Hong Kong SAR, China · Department of Electrical and Computer Engineering, University of Pittsburgh, Pittsburgh, PA, USA · Shenzhen University, China +6

Feb 23, 2026J/K move · Enter open · S save

OrthoAI: A Neurosymbolic Framework for Evidence-Grounded Biomechanical Reasoning in Clear Aligner Orthodontics

Edouard Lansiaux, Margaux Leman, Mehdi Ammi

STaR-AI, Emergency Department, Lille University Hospital, 59000, Lille, France · Artificial Intelligence and Data Semantics Laboratory, Paris 8 University, 93200, Saint-Denis, France

3D Point Cloud Segmentation 3D Medical Image Segmentation Weakly Supervised Learning

Open paper AskPDF arXiv

Automated clinical decision support for clear aligner orthodontics faces a key challenge: bridging geometric perception (3D tooth segmentation) with clinical reasoning (biomechanical feasibility). We address this with OrthOAI, introducing three methodological contributions. First, sparse-supervision segmentation: a landmark-to-point-cloud synthesis protocol enables training from sparse anatomical annotations (6-8 points per tooth) instead of dense labels, combined with a clinically stratified loss mixing label-smoothed cross-entropy and a batch-adaptive Dice term for class imbalance. Second, knowledge-grounded constraint inference: biomechanical feasibility is modeled as a Constraint Satisfaction Problem over a domain ontology of tooth movements, encoding evidence-based per-stage limits as soft and hard constraints. Third, multi-criteria treatment evaluation: treatment quality is scored through a formal Multi-Criteria Decision Analysis framework using a weighted Additive Value Function grounded in clinical priority theory. On landmark-reconstructed point clouds from 3DTeethLand (MICCAI 2024), segmentation reaches 81.4% Tooth Identification Rate with 60,705 parameters. Ablations quantify the impact of each design choice. End-to-end inference runs in under 4 seconds on CPU. We also outline the gap between the current prototype-trained on synthetic ellipsoidal approximations-and clinical deployment, with a roadmap for validation. Code and weights are released.

Abstract

Explore similar work

OrthoAI: A Neurosymbolic Framework for Evidence-Grounded Biomechanical Reasoning in Clear Aligner Orthodontics

MATHENA: Mamba-based Architectural Tooth Hierarchical Estimator and Holistic Evaluation Network for Anatomy

OralAgent: Integrating Reasoning, Tools, and Knowledge for Interactive Dental Image Analysis