DMF-Net: Image-Guided Point Cloud Completion with Dual-Channel Modality Fusion and Shape-Aware Upsampling Transformer

TL;DR

This paper introduces DMF-Net, a dual-channel fusion network that effectively combines global shape information from images and local details from partial point clouds to improve 3D point cloud completion.

Contribution

It proposes a novel dual-channel modality fusion approach with a shape-aware upsampling transformer for more accurate point cloud completion.

Findings

DMF-Net outperforms existing methods on ShapeNet-ViPC dataset.

The dual-channel fusion effectively balances global and local information.

Shape-aware upsampling improves point cloud density and detail.

Abstract

In this paper we study the task of a single-view image-guided point cloud completion. Existing methods have got promising results by fusing the information of image into point cloud explicitly or implicitly. However, given that the image has global shape information and the partial point cloud has rich local details, We believe that both modalities need to be given equal attention when performing modality fusion. To this end, we propose a novel dual-channel modality fusion network for image-guided point cloud completion(named DMF-Net), in a coarse-to-fine manner. In the first stage, DMF-Net takes a partial point cloud and corresponding image as input to recover a coarse point cloud. In the second stage, the coarse point cloud will be upsampled twice with shape-aware upsampling transformer to get the dense and complete point cloud. Extensive quantitative and qualitative experimental results show that DMF-Net outperforms the state-of-the-art unimodal and multimodal point cloud completion works on ShapeNet-ViPC dataset.