GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloud

TL;DR

The paper presents GSPN, a novel 3D object proposal method using analysis-by-synthesis for improved 3D instance segmentation in point clouds, achieving state-of-the-art results.

Contribution

Introduction of GSPN, a shape reconstruction-based proposal method integrated into R-PointNet for enhanced 3D instance segmentation.

Findings

Achieves state-of-the-art performance on 3D segmentation tasks.

Reduces low-objectness proposals through geometric analysis.

Demonstrates the effectiveness of shape reconstruction in proposal generation.

Abstract

We introduce a novel 3D object proposal approach named Generative Shape Proposal Network (GSPN) for instance segmentation in point cloud data. Instead of treating object proposal as a direct bounding box regression problem, we take an analysis-by-synthesis strategy and generate proposals by reconstructing shapes from noisy observations in a scene. We incorporate GSPN into a novel 3D instance segmentation framework named Region-based PointNet (R-PointNet) which allows flexible proposal refinement and instance segmentation generation. We achieve state-of-the-art performance on several 3D instance segmentation tasks. The success of GSPN largely comes from its emphasis on geometric understandings during object proposal, which greatly reducing proposals with low objectness.