3DSGrasp: 3D Shape-Completion for Robotic Grasp

TL;DR

3DSGrasp introduces a Transformer-based network for completing partial 3D point clouds, enabling more reliable robotic grasping by predicting missing geometry and improving success rates in real-world applications.

Contribution

The paper presents a novel pose-invariant Transformer-based PCD completion network that enhances robotic grasping accuracy from partial point clouds.

Findings

Outperforms state-of-the-art PCD completion methods.

Significantly improves grasping success rate in real-world tests.

Generates geometrically consistent and complete point clouds.

Abstract

Real-world robotic grasping can be done robustly if a complete 3D Point Cloud Data (PCD) of an object is available. However, in practice, PCDs are often incomplete when objects are viewed from few and sparse viewpoints before the grasping action, leading to the generation of wrong or inaccurate grasp poses. We propose a novel grasping strategy, named 3DSGrasp, that predicts the missing geometry from the partial PCD to produce reliable grasp poses. Our proposed PCD completion network is a Transformer-based encoder-decoder network with an Offset-Attention layer. Our network is inherently invariant to the object pose and point's permutation, which generates PCDs that are geometrically consistent and completed properly. Experiments on a wide range of partial PCD show that 3DSGrasp outperforms the best state-of-the-art method on PCD completion tasks and largely improves the grasping success rate in real-world scenarios. The code and dataset will be made available upon acceptance.