ArrowPose: Segmentation, Detection, and 5 DoF Pose Estimation Network for Colorless Point Clouds

TL;DR

ArrowPose is a rapid neural network that accurately detects objects and estimates their 5 DoF pose from colorless point clouds, trained on synthetic data and achieving state-of-the-art results in under 250 milliseconds.

Contribution

The paper introduces a novel fast detection and pose estimation network specifically designed for colorless point clouds, outperforming existing methods in speed and accuracy.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Runs inference in only 250 milliseconds.

Outperforms all existing colorless point cloud methods.

Abstract

This paper presents a fast detection and 5 DoF (Degrees of Freedom) pose estimation network for colorless point clouds. The pose estimation is calculated from center and top points of the object, predicted by the neural network. The network is trained on synthetic data, and tested on a benchmark dataset, where it demonstrates state-of-the-art performance and outperforms all colorless methods. The network is able to run inference in only 250 milliseconds making it usable in many scenarios. Project page with code at arrowpose.github.io