DBDNet:Partial-to-Partial Point Cloud Registration with Dual Branches Decoupling

TL;DR

This paper proposes an effective registration method with dual branches decoupling for partial-to-partial registration, dubbed as DBDNet, and introduces a dual branches structure to eliminate mutual interference error between rotation and translation by separately creating two individual correspondence matrices.

Abstract

Point cloud registration plays a crucial role in various computer vision tasks, and usually demands the resolution of partial overlap registration in practice. Most existing methods perform a serial calculation of rotation and translation, while jointly predicting overlap during registration, this coupling tends to degenerate the registration performance. In this paper, we propose an effective registration method with dual branches decoupling for partial-to-partial registration, dubbed as DBDNet. Specifically, we introduce a dual branches structure to eliminate mutual interference error between rotation and translation by separately creating two individual correspondence matrices. For partial-to-partial registration, we consider overlap prediction as a preordering task before the registration procedure. Accordingly, we present an overlap predictor that benefits from explicit feature interaction, which is achieved by the powerful attention mechanism to accurately predict pointwise masks. Furthermore, we design a multi-resolution feature extraction network to capture both local and global patterns thus enhancing both overlap prediction and registration module. Experimental results on both synthetic and real datasets validate the effectiveness of our proposed method.

Figures (6)

• Figure 1: Overall architecture of the proposed DBDNet. Overlapping points are first filtered by the overlap prediction module. Then, a multi-resolution network and attention blocks are used to extract and fuse pointwise features. Finally, the rotation and translation are separately calculated from dual branches.
• Figure 2: Structure of multi-resolution feature extraction network.
• Figure 3: Structure of attention block.
• Figure 4: Structure of overlap prediction module.
• Figure 5: The qualitative results on Modelnet40 (a) unseen shapes with Gaussian noise, (b) unseen categories with Gaussian noise.