XavierJiezou/arch2dc-models

Arch2DC Model Weights

This repository contains the pretrained model weights for Arch2DC, an arch-aware two-stage framework for missing dental crown restoration.

Model Overview

Arch2DC consists of two stages:

• Stage I: AdaPoinTr-based point cloud completion with local-global dual constraints (α=0.8)
• Stage II: SAP-based surface reconstruction with dual-constraint supervision (β=0.9, δ=0.05)

File Structure

arch2dc-models/
├── environment.yml                  # Conda environment (tooth)
├── arch2dc/
│   ├── stage1_best_l1_cd.pth        # Ours Stage I  (AdaPoinTr + global, α=0.8)
│   └── stage2_model_best.pt         # Ours Stage II (SAP dual-constraint, β=0.9)
├── stage1_pcn/
│   └── best_l1_cd.pth               # PCN baseline Stage I
├── stage1_adapointr_bpa/
│   └── best_l1_cd.pth               # AdaPoinTr+BPA baseline Stage I
└── dmc/
    └── best_val.pth                 # DMC baseline

Environment Setup

conda env create -f environment.yml
conda activate tooth

Citation

If you use these models in your research, please cite:

@article{arch2dc2025,
  title={Arch2DC: An Arch-Aware Two-Stage Framework for Missing Dental Crown Restoration},
  author={...},
  journal={...},
  year={2025}
}