Saved in:
Bibliographic Details
Main Authors: Sun, Jianan, Huang, Yukang, Wang, Dongzhihan, Fan, Mingyu
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.03598
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917121856897024
author Sun, Jianan
Huang, Yukang
Wang, Dongzhihan
Fan, Mingyu
author_facet Sun, Jianan
Huang, Yukang
Wang, Dongzhihan
Fan, Mingyu
contents Partial dental point clouds often suffer from large missing regions caused by occlusion and limited scanning views, which bias encoder-only global features and force decoders to hallucinate structures. We propose a retrieval-augmented framework for tooth completion that integrates a prototype memory into standard encoder--decoder pipelines. After encoding a partial input into a global descriptor, the model retrieves the nearest manifold prototype from a learnable memory and fuses it with the query feature through confidence-gated weighting before decoding. The memory is optimized end-to-end and self-organizes into reusable tooth-shape prototypes without requiring tooth-position labels, thereby providing structural priors that stabilize missing-region inference and free decoder capacity for detail recovery. The module is plug-and-play and compatible with common completion backbones, while keeping the same training losses. Experiments on a self-processed Teeth3DS benchmark demonstrate consistent improvements in Chamfer Distance, with visualizations showing sharper cusps, ridges, and interproximal transitions. Our approach provides a simple yet effective way to exploit cross-sample regularities for more accurate and faithful dental point-cloud completion.
format Preprint
id arxiv_https___arxiv_org_abs_2512_03598
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Memory-Guided Point Cloud Completion for Dental Reconstruction
Sun, Jianan
Huang, Yukang
Wang, Dongzhihan
Fan, Mingyu
Computer Vision and Pattern Recognition
Partial dental point clouds often suffer from large missing regions caused by occlusion and limited scanning views, which bias encoder-only global features and force decoders to hallucinate structures. We propose a retrieval-augmented framework for tooth completion that integrates a prototype memory into standard encoder--decoder pipelines. After encoding a partial input into a global descriptor, the model retrieves the nearest manifold prototype from a learnable memory and fuses it with the query feature through confidence-gated weighting before decoding. The memory is optimized end-to-end and self-organizes into reusable tooth-shape prototypes without requiring tooth-position labels, thereby providing structural priors that stabilize missing-region inference and free decoder capacity for detail recovery. The module is plug-and-play and compatible with common completion backbones, while keeping the same training losses. Experiments on a self-processed Teeth3DS benchmark demonstrate consistent improvements in Chamfer Distance, with visualizations showing sharper cusps, ridges, and interproximal transitions. Our approach provides a simple yet effective way to exploit cross-sample regularities for more accurate and faithful dental point-cloud completion.
title Memory-Guided Point Cloud Completion for Dental Reconstruction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.03598