Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts
This work is significant for researchers and practitioners in 3D computer vision, as it offers a data-efficient learning approach that substantially reduces the need for extensive 3D point cloud annotations, which are costly and time-consuming to acquire.
This paper addresses the challenge of data scarcity in 3D scene understanding by proposing Contrastive Scene Contexts, a 3D pre-training method. The method achieves state-of-the-art results on benchmarks with limited training data or labels, demonstrating that with only 0.1% of point labels on ScanNet, it can reach 89% of baseline instance segmentation and 96% of baseline semantic segmentation performance.
The rapid progress in 3D scene understanding has come with growing demand for data; however, collecting and annotating 3D scenes (e.g. point clouds) are notoriously hard. For example, the number of scenes (e.g. indoor rooms) that can be accessed and scanned might be limited; even given sufficient data, acquiring 3D labels (e.g. instance masks) requires intensive human labor. In this paper, we explore data-efficient learning for 3D point cloud. As a first step towards this direction, we propose Contrastive Scene Contexts, a 3D pre-training method that makes use of both point-level correspondences and spatial contexts in a scene. Our method achieves state-of-the-art results on a suite of benchmarks where training data or labels are scarce. Our study reveals that exhaustive labelling of 3D point clouds might be unnecessary; and remarkably, on ScanNet, even using 0.1% of point labels, we still achieve 89% (instance segmentation) and 96% (semantic segmentation) of the baseline performance that uses full annotations.