Point Cloud Denoising
Point cloud denoising aims to remove noise and outliers from 3D point cloud data, improving the accuracy and reliability of subsequent analyses and applications. Current research focuses on deep learning approaches, employing architectures like graph convolutional networks, transformers, and diffusion models, often incorporating geometric information and multi-scale features to achieve optimal denoising while preserving fine details. These advancements are crucial for various fields, including autonomous driving, 3D modeling, and robotics, where high-quality point cloud data is essential for accurate perception and decision-making. The development of efficient and robust denoising methods is driving progress in these areas by enabling more reliable and accurate processing of 3D data.
Papers
LBF:Learnable Bilateral Filter For Point Cloud Denoising
Huajian Si, Zeyong Wei, Zhe Zhu, Honghua Chen, Dong Liang, Weiming Wang, Mingqiang Wei
GeoGCN: Geometric Dual-domain Graph Convolution Network for Point Cloud Denoising
Zhaowei Chen, Peng Li, Zeyong Wei, Honghua Chen, Haoran Xie, Mingqiang Wei, Fu Lee Wang