Double-Dot Network for Antipodal Grasp Detection

TL;DR

This paper introduces Double-Dot Network (DD-Net), a deep learning model for antipodal grasp detection that predicts fingertip positions without relying on pre-set anchors, improving robustness and generalization to unseen objects.

Contribution

The paper presents a novel anchor-free CNN architecture for grasp detection, with a new loss function and fingertip-based representation, enhancing accuracy and robustness over existing methods.

Findings

Achieves state-of-the-art accuracy in simulation and robotic experiments.

Effectively detects grasps on unseen objects.

Outperforms existing methods in robustness and generalization.

Abstract

This paper proposes a new deep learning approach to antipodal grasp detection, named Double-Dot Network (DD-Net). It follows the recent anchor-free object detection framework, which does not depend on empirically pre-set anchors and thus allows more generalized and flexible prediction on unseen objects. Specifically, unlike the widely used 5-dimensional rectangle, the gripper configuration is defined as a pair of fingertips. An effective CNN architecture is introduced to localize such fingertips, and with the help of auxiliary centers for refinement, it accurately and robustly infers grasp candidates. Additionally, we design a specialized loss function to measure the quality of grasps, and in contrast to the IoU scores of bounding boxes adopted in object detection, it is more consistent to the grasp detection task. Both the simulation and robotic experiments are executed and state of the art accuracies are achieved, showing that DD-Net is superior to the counterparts in handling unseen objects.