With the rapid development of 3D imaging sensors, such as depth cameras and laser scanning systems, 3D data has become increasingly accessible. Meanwhile, the increasing popularity of deep learning algorithms, such as convolutional neural networks and deep reinforcement learning, has further increased the usability of 3D vision systems. Driven by these factors, 3D vision has become an emerging and core component for numerous applications, such as autonomous driving, AR/VR, and robotics. Although remarkable progress has been achieved in this area during the last few years, there are still several challenges that need to be addressed, such as the noisy, sparse, and irregular nature of point clouds, and the high cost to label 3D data. 3D data produced by different 3D imaging sensors (e.g., structured light, stereo, LiDAR, time-of-flight) have different characteristics. It is therefore necessary to study general algorithms that can mitigate the domain gap between different types of 3D data. Besides, how to effectively integrate geometry-based and learning based techniques to develop 3D vision systems is still an open problem. The aim of this special issue of IET Computer Vision is to collect and present the latest research development in learning-based 3D vision theories and their applications, and to inspire future research in this area. Papers working on 3D data acquisition, 3D modelling, 3D data (including point cloud, voxels, meshes) analysis, and their applications with deep learning are within the scope of this special issue. Note that papers purely working on 2D vision tasks (for example regular video processing) are out of the scope of this special issue.

Guest Editors:  Yulan Guo, Hanyun Wang, Stefano Berretti, Ronald Clark and Mohammed Bennamoun.

Submissions must be made through ScholarOne by 31 December 2021.