GRNet: Gridding Residual Network for Dense Point Cloud Completion

TL;DR

This paper introduces GRNet, a novel neural network architecture that uses 3D grids and differentiable layers to improve dense point cloud completion, preserving structural details better than existing methods.

Contribution

GRNet employs innovative differentiable layers for converting between point clouds and 3D grids, enhancing detail preservation in point cloud completion tasks.

Findings

Outperforms state-of-the-art methods on ShapeNet, Completion3D, and KITTI datasets.

Effectively preserves structural details in point cloud completion.

Introduces novel differentiable layers for grid conversion and feature sampling.

Abstract

Estimating the complete 3D point cloud from an incomplete one is a key problem in many vision and robotics applications. Mainstream methods (e.g., PCN and TopNet) use Multi-layer Perceptrons (MLPs) to directly process point clouds, which may cause the loss of details because the structural and context of point clouds are not fully considered. To solve this problem, we introduce 3D grids as intermediate representations to regularize unordered point clouds. We therefore propose a novel Gridding Residual Network (GRNet) for point cloud completion. In particular, we devise two novel differentiable layers, named Gridding and Gridding Reverse, to convert between point clouds and 3D grids without losing structural information. We also present the differentiable Cubic Feature Sampling layer to extract features of neighboring points, which preserves context information. In addition, we design a new loss function, namely Gridding Loss, to calculate the L1 distance between the 3D grids of the predicted and ground truth point clouds, which is helpful to recover details. Experimental results indicate that the proposed GRNet performs favorably against state-of-the-art methods on the ShapeNet, Completion3D, and KITTI benchmarks.