Bi-PointFlowNet: Bidirectional Learning for Point Cloud Based Scene Flow Estimation

TL;DR

This paper introduces Bi-PointFlowNet, a bidirectional learning architecture for scene flow estimation from point clouds, significantly improving accuracy and efficiency over previous unidirectional methods.

Contribution

It proposes a novel bidirectional flow embedding layer and hierarchical feature extraction, advancing scene flow estimation accuracy and computational efficiency.

Findings

Achieved state-of-the-art results on FlyingThings3D and KITTI datasets.

Outperformed existing methods with a large margin.

Reduced computational overhead through hierarchical features.

Abstract

Scene flow estimation, which extracts point-wise motion between scenes, is becoming a crucial task in many computer vision tasks. However, all of the existing estimation methods utilize only the unidirectional features, restricting the accuracy and generality. This paper presents a novel scene flow estimation architecture using bidirectional flow embedding layers. The proposed bidirectional layer learns features along both forward and backward directions, enhancing the estimation performance. In addition, hierarchical feature extraction and warping improve the performance and reduce computational overhead. Experimental results show that the proposed architecture achieved a new state-of-the-art record by outperforming other approaches with large margin in both FlyingThings3D and KITTI benchmarks. Codes are available at https://github.com/cwc1260/BiFlow.