3DSGrasp: 3D Shape-Completion for Robotic Grasp
This addresses the challenge of robust robotic grasping in real-world scenarios where objects are viewed from sparse viewpoints, though it appears incremental as it builds on existing point cloud completion techniques.
The paper tackles the problem of robotic grasping with incomplete 3D point cloud data by proposing 3DSGrasp, a method that predicts missing geometry to generate reliable grasp poses, resulting in outperforming state-of-the-art methods on completion tasks and significantly improving real-world grasping success rates.
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.