Memory-Guided Point Cloud Completion for Dental Reconstruction

TL;DR

This paper introduces a retrieval-augmented point cloud completion framework for dental reconstruction, leveraging a learnable memory of tooth prototypes to improve accuracy and detail in partial scans.

Contribution

It proposes a novel memory-augmented approach that enhances encoder-decoder models with reusable tooth-shape prototypes for better dental point cloud completion.

Findings

Improves Chamfer Distance on Teeth3DS benchmark

Produces sharper dental features in reconstructions

Memory module is plug-and-play and improves existing models

Abstract

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.