Abstract
Soft-tissue-driven orthognathic surgical planning requires re- constructing anatomically coherent facial soft-tissue volumes from a given target outer facial surface. However, existing implicit neural representa- tion approaches often lack volumetric constraints and fail to preserve anatomical consistency. In this paper, we present a Projection-Aware Implicit Neural Transformer (PAINT) for patient-specific soft-tissue volumetric reconstruction. Given a voxelized outer-surface representa- tion, PAINT predicts a soft-tissue implicit grid that reconstructs the enclosed volume conditioned on the input surface. PAINT combines a hybrid 3D CNN encoder with transformer layers for global anatomi- cal context, a 3D decoder with triplet skip attention to capture input- conditioned volumetric relationships, and volume-constraint and projection- aware reconstruction losses to enforce local fidelity and global volumetric consistency. Experiments demonstrate the effectiveness of key compo- nents of PAINT and superior performance over existing 3D networks.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/ODIN_007.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=DnGZ1NLLJS
BibTex
@InProceedings{YanSu_PAINT_MICCAISAT2026,
author = { Yang, Su AND Kim, Daeseung AND Gu, Kevin AND Dharia, Rohan AND Jung, Lois AND Gateno, Jaime},
title = { { PAINT: Projection-Aware Implicit Neural Transformer for Soft-Tissue Surface Generation in Orthognathic Surgical Planning } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17270},
month = {pending},
page = {pending}
}