Abstract:Due to limitations in acquisition equipment, noise perturbations often corrupt 3-D point clouds, hindering down-stream tasks such as surface reconstruction, rendering, and further processing. Existing 3-D point cloud denoising methods typically fail to reliably fit the underlying continuous surface, resulting in a degradation of reconstruction performance. This paper introduces fine-granularity dynamic graph convolutional networks called GD-GCN, a novel approach to denoising in 3-D point clouds. The GD-GCN employs micro-step temporal graph convolution (MST-GConv) to perform feature learning in a gradual manner. Compared with the conventional GCN, which commonly uses discrete integer-step graph convolution, this modification introduces a more adaptable and nuanced approach to feature learning within graph convolution networks. It more accurately depicts the process of fitting the point cloud with noise to the underlying surface by and the learning process for MST-GConv acts like a changing system and is managed through a type of neural network known as neural Partial Differential Equations (PDEs). This means it can adapt and improve over time. GD-GCN approximates the Riemannian metric, calculating distances between points along a low-dimensional manifold. This capability allows it to understand the local geometric structure and effectively capture diverse relationships between points from different geometric regions through geometric graph construction based on Riemannian distances. Additionally, GD-GCN incorporates robust graph spectral filters based on the Bernstein polynomial approximation, which modulate eigenvalues for complex and arbitrary spectral responses, providing theoretical guarantees for BIBO stability. Symmetric channel mixing matrices further enhance filter flexibility by enabling channel-level scaling and shifting in the spectral domain.

Progressive Point Cloud Denoising with Cross-Stage Cross-Coder Adaptive Edge Graph Convolution Network

MSPoint: Point Cloud Denoising Network Based on Multiscale Distribution Score

Spatial-Adaptive Network for Single Image Denoising

Point Cloud Completion Cascade Optimization Network Based on Feature Fusion

Dual-scale Point Cloud Completion Network Based on High-Frequency Feature Fusion

Adaptive Recurrent Forward Network for Dense Point Cloud Completion

Point Cloud Denoising With Fine-Granularity Dynamic Graph Convolutional Networks

RePCD-Net: Feature-Aware Recurrent Point Cloud Denoising Network

3D Point Cloud Denoising via Deep Neural Network Based Local Surface Estimation

3D point cloud denoising method based on global feature guidance

GeoGCN: Geometric Dual-domain Graph Convolution Network for Point Cloud Denoising

Graph-Based Point Cloud Denoising

GCN-Denoiser: Mesh Denoising with Graph Convolutional Networks

PathNet: Path-Selective Point Cloud Denoising

Edge Enhancement Oriented Graph Convolutional Networks for Point Cloud Segmentation

MODNet: Multi-offset Point Cloud Denoising Network Customized for Multi-scale Patches

Multi-Patch Collaborative Point Cloud Denoising via Low-Rank Recovery with Graph Constraint

3D Dynamic Point Cloud Denoising via Spatial-Temporal Graph Learning

Non-Local Part-Aware Point Cloud Denoising

Progressive Point Cloud Deconvolution Generation Network