Learning Graph-Convolutional Representations for Point Cloud Denoising
This is an incremental improvement for point cloud processing, enhancing denoising robustness in applications like LiDAR scans.
The paper tackles point cloud denoising by proposing a graph-convolutional neural network that addresses permutation-invariance, significantly outperforming state-of-the-art methods on metrics like Chamfer measure and surface normal quality, especially under high or structured noise.
Point clouds are an increasingly relevant data type but they are often corrupted by noise. We propose a deep neural network based on graph-convolutional layers that can elegantly deal with the permutation-invariance problem encountered by learning-based point cloud processing methods. The network is fully-convolutional and can build complex hierarchies of features by dynamically constructing neighborhood graphs from similarity among the high-dimensional feature representations of the points. When coupled with a loss promoting proximity to the ideal surface, the proposed approach significantly outperforms state-of-the-art methods on a variety of metrics. In particular, it is able to improve in terms of Chamfer measure and of quality of the surface normals that can be estimated from the denoised data. We also show that it is especially robust both at high noise levels and in presence of structured noise such as the one encountered in real LiDAR scans.