PointVoteNet: Accurate Object Detection and 6 DoF Pose Estimation in Point Clouds

TL;DR

PointVoteNet introduces a learning-based approach that directly estimates 6 DoF object poses from unordered point clouds, achieving high accuracy and surpassing some existing methods.

Contribution

It is the first to directly estimate 6 DoF poses from raw point clouds with or without RGB, improving accuracy over prior RGB-dependent methods.

Findings

Achieves state-of-the-art accuracy on pose estimation benchmarks.

Works effectively with point clouds with or without RGB data.

Outperforms some existing methods trained on the same data.

Abstract

We present a learning-based method for 6 DoF pose estimation of rigid objects in point cloud data. Many recent learning-based approaches use primarily RGB information for detecting objects, in some cases with an added refinement step using depth data. Our method consumes unordered point sets with/without RGB information, from initial detection to the final transformation estimation stage. This allows us to achieve accurate pose estimates, in some cases surpassing state of the art methods trained on the same data.