← Back to Benchmarks
simmediumgraspingmetric · varies

Evaluating Gaussian Grasp Maps for Generative Grasping Models

Description

Generalising robotic grasping to previously unseen objects is a key task in general robotic manipulation. The current method for training many antipodal generative grasping models rely on a binary ground truth grasp map generated from the centre thirds of correctly labelled grasp rectangles. However, these binary maps do not accurately reflect the positions in which a robotic arm can correctly grasp a given object. We propose a continuous Gaussian representation of annotated grasps to generate g

Source

http://arxiv.org/abs/2206.00432v1