FBNet: Feedback Network for Point Cloud Completion

TL;DR

FBNet introduces a feedback mechanism with cross attention for point cloud completion, enabling iterative refinement of features and improving shape reconstruction accuracy over existing methods.

Contribution

The paper proposes a novel feedback network architecture with cross transformer for enhanced point cloud completion, addressing feature dimension mismatch issues.

Findings

FBNet outperforms state-of-the-art methods on multiple datasets.

Feedback connections improve feature refinement and shape quality.

Cross attention effectively integrates feedback features for better completion.

Abstract

The rapid development of point cloud learning has driven point cloud completion into a new era. However, the information flows of most existing completion methods are solely feedforward, and high-level information is rarely reused to improve low-level feature learning. To this end, we propose a novel Feedback Network (FBNet) for point cloud completion, in which present features are efficiently refined by rerouting subsequent fine-grained ones. Firstly, partial inputs are fed to a Hierarchical Graph-based Network (HGNet) to generate coarse shapes. Then, we cascade several Feedback-Aware Completion (FBAC) Blocks and unfold them across time recurrently. Feedback connections between two adjacent time steps exploit fine-grained features to improve present shape generations. The main challenge of building feedback connections is the dimension mismatching between present and subsequent features. To address this, the elaborately designed point Cross Transformer exploits efficient information from feedback features via cross attention strategy and then refines present features with the enhanced feedback features. Quantitative and qualitative experiments on several datasets demonstrate the superiority of proposed FBNet compared to state-of-the-art methods on point completion task.