PointVST: Self-Supervised Pre-training for 3D Point Clouds via View-Specific Point-to-Image Translation
This addresses the problem of limited self-supervised learning methods for 3D point clouds, benefiting researchers and practitioners in computer vision and robotics, though it is incremental as it builds on existing pre-training paradigms.
The paper tackles the lack of self-supervised pre-training advancements in 3D point cloud learning by proposing PointVST, a translative framework that uses cross-modal translation from point clouds to 2D rendered images, achieving consistent and prominent performance superiority over state-of-the-art approaches in various downstream tasks.
The past few years have witnessed the great success and prevalence of self-supervised representation learning within the language and 2D vision communities. However, such advancements have not been fully migrated to the field of 3D point cloud learning. Different from existing pre-training paradigms designed for deep point cloud feature extractors that fall into the scope of generative modeling or contrastive learning, this paper proposes a translative pre-training framework, namely PointVST, driven by a novel self-supervised pretext task of cross-modal translation from 3D point clouds to their corresponding diverse forms of 2D rendered images. More specifically, we begin with deducing view-conditioned point-wise embeddings through the insertion of the viewpoint indicator, and then adaptively aggregate a view-specific global codeword, which can be further fed into subsequent 2D convolutional translation heads for image generation. Extensive experimental evaluations on various downstream task scenarios demonstrate that our PointVST shows consistent and prominent performance superiority over current state-of-the-art approaches as well as satisfactory domain transfer capability. Our code will be publicly available at https://github.com/keeganhk/PointVST.