Completing point cloud from few points by Wasserstein GAN and Transformers

TL;DR

This paper introduces a novel point cloud completion method using Wasserstein GANs and Transformers, specifically designed to perform well even with very few input points, addressing a key challenge in vision and robotics applications.

Contribution

The paper proposes an end-to-end encoder-decoder network with Transformers and Wasserstein GANs, pre-trained and fine-tuned for improved point cloud completion from limited data.

Findings

Improves completion performance for many input points.

Maintains stable performance with few input points.

Validated on ShapeNet dataset.

Abstract

In many vision and robotics applications, it is common that the captured objects are represented by very few points. Most of the existing completion methods are designed for partial point clouds with many points, and they perform poorly or even fail completely in the case of few points. However, due to the lack of detail information, completing objects from few points faces a huge challenge. Inspired by the successful applications of GAN and Transformers in the image-based vision task, we introduce GAN and Transformer techniques to address the above problem. Firstly, the end-to-end encoder-decoder network with Transformers and the Wasserstein GAN with Transformer are pre-trained, and then the overall network is fine-tuned. Experimental results on the ShapeNet dataset show that our method can not only improve the completion performance for many input points, but also keep stable for few input points. Our source code is available at https://github.com/WxfQjh/Stability-point-recovery.git.