GaPro: Box-Supervised 3D Point Cloud Instance Segmentation Using Gaussian Processes as Pseudo Labelers

TL;DR

GaPro introduces a weakly supervised 3D point cloud instance segmentation method using Gaussian Processes to generate pseudo labels from bounding boxes, outperforming previous weakly supervised approaches and rivaling fully supervised methods.

Contribution

The paper presents a novel Gaussian Process-based pseudo labeling approach for weakly supervised 3D instance segmentation using bounding boxes, enabling effective training without dense masks.

Findings

GaPro outperforms previous weakly supervised methods.

GaPro achieves competitive results with fully supervised methods.

The approach is robust and adaptable to various state-of-the-art models.

Abstract

Instance segmentation on 3D point clouds (3DIS) is a longstanding challenge in computer vision, where state-of-the-art methods are mainly based on full supervision. As annotating ground truth dense instance masks is tedious and expensive, solving 3DIS with weak supervision has become more practical. In this paper, we propose GaPro, a new instance segmentation for 3D point clouds using axis-aligned 3D bounding box supervision. Our two-step approach involves generating pseudo labels from box annotations and training a 3DIS network with the resulting labels. Additionally, we employ the self-training strategy to improve the performance of our method further. We devise an effective Gaussian Process to generate pseudo instance masks from the bounding boxes and resolve ambiguities when they overlap, resulting in pseudo instance masks with their uncertainty values. Our experiments show that GaPro outperforms previous weakly supervised 3D instance segmentation methods and has competitive performance compared to state-of-the-art fully supervised ones. Furthermore, we demonstrate the robustness of our approach, where we can adapt various state-of-the-art fully supervised methods to the weak supervision task by using our pseudo labels for training. The source code and trained models are available at https://github.com/VinAIResearch/GaPro.