Deep Learning for 3D Point Cloud Enhancement: A Survey

TL;DR

This survey comprehensively reviews deep learning methods for improving 3D point clouds by denoising, completing, and upsampling, highlighting recent advances, taxonomy, and future research directions.

Contribution

It provides the first systematic survey of deep-learning-based point cloud enhancement techniques, including taxonomy, experimental results, and insights.

Findings

Deep learning effectively denoises point clouds.

Methods successfully recover missing data in 3D scans.

Upsampling improves point cloud density significantly.

Abstract

Point cloud data now are popular data representations in a number of three-dimensional (3D) vision research realms. However, due to the limited performance of sensors and sensing noise, the raw data usually suffer from sparsity, noise, and incompleteness. This poses great challenges to down-stream point cloud processing tasks. In recent years, deep-learning-based point cloud enhancement methods, which aim to achieve dense, clean, and complete point clouds from low-quality raw point clouds using deep neural networks, are gaining tremendous research attention. This paper, for the first time to our knowledge, presents a comprehensive survey for deep-learning-based point cloud enhancement methods. It covers three main perspectives for point cloud enhancement, i.e., (1) denoising to achieve clean data; (2) completion to recover unseen data; (3) upsampling to obtain dense data. Our survey presents a new taxonomy for recent state-of-the-art methods and systematic experimental results on standard benchmarks. In addition, we share our insightful observations, thoughts, and inspiring future research directions for point cloud enhancement with deep learning.