S4G: Amodal Single-view Single-Shot SE(3) Grasp Detection in Cluttered Scenes

TL;DR

This paper introduces a learning-based, single-shot neural network approach for 6-DoF grasp detection in cluttered scenes from a single viewpoint, trained on synthetic data and tested in real-world scenarios, outperforming existing methods.

Contribution

It presents a novel single-shot grasp proposal network trained with synthetic data and an innovative data synthesis pipeline for effective real-world grasp detection.

Findings

Outperforms state-of-the-art methods significantly

Effective in synthetic and real environments

Efficient amodal grasp prediction

Abstract

Grasping is among the most fundamental and long-lasting problems in robotics study. This paper studies the problem of 6-DoF(degree of freedom) grasping by a parallel gripper in a cluttered scene captured using a commodity depth sensor from a single viewpoint. We address the problem in a learning-based framework. At the high level, we rely on a single-shot grasp proposal network, trained with synthetic data and tested in real-world scenarios. Our single-shot neural network architecture can predict amodal grasp proposal efficiently and effectively. Our training data synthesis pipeline can generate scenes of complex object configuration and leverage an innovative gripper contact model to create dense and high-quality grasp annotations. Experiments in synthetic and real environments have demonstrated that the proposed approach can outperform state-of-the-arts by a large margin.