Cascaded Refinement Network for Point Cloud Completion with Self-supervision

TL;DR

This paper introduces a two-branch cascaded network for point cloud completion that improves detail preservation and can be trained with limited or no ground truth data, achieving superior results.

Contribution

A novel two-branch network with self-supervised strategies for point cloud completion that reduces dependence on ground truth data and enhances reconstruction quality.

Findings

Outperforms state-of-the-art methods in realistic shape completion.

Effective in self-supervised, semi-supervised, and fully supervised settings.

Improves detail preservation in reconstructed point clouds.

Abstract

Point clouds are often sparse and incomplete, which imposes difficulties for real-world applications. Existing shape completion methods tend to generate rough shapes without fine-grained details. Considering this, we introduce a two-branch network for shape completion. The first branch is a cascaded shape completion sub-network to synthesize complete objects, where we propose to use the partial input together with the coarse output to preserve the object details during the dense point reconstruction. The second branch is an auto-encoder to reconstruct the original partial input. The two branches share a same feature extractor to learn an accurate global feature for shape completion. Furthermore, we propose two strategies to enable the training of our network when ground truth data are not available. This is to mitigate the dependence of existing approaches on large amounts of ground truth training data that are often difficult to obtain in real-world applications. Additionally, our proposed strategies are also able to improve the reconstruction quality for fully supervised learning. We verify our approach in self-supervised, semi-supervised and fully supervised settings with superior performances. Quantitative and qualitative results on different datasets demonstrate that our method achieves more realistic outputs than state-of-the-art approaches on the point cloud completion task.