Fully Convolutional Grasp Detection Network with Oriented Anchor Box
This work addresses the problem of efficient and accurate grasp pose prediction for robotic manipulation, representing an incremental improvement over existing methods.
The paper tackles robotic grasp detection from RGB images by proposing a real-time method using an end-to-end fully convolutional network with oriented anchor boxes, achieving accuracies of 97.74% and 96.61% on the Cornell Grasp Dataset and outperforming the latest state-of-the-art by up to 1.74%.
In this paper, we present a real-time approach to predict multiple grasping poses for a parallel-plate robotic gripper using RGB images. A model with oriented anchor box mechanism is proposed and a new matching strategy is used during the training process. An end-to-end fully convolutional neural network is employed in our work. The network consists of two parts: the feature extractor and multi-grasp predictor. The feature extractor is a deep convolutional neural network. The multi-grasp predictor regresses grasp rectangles from predefined oriented rectangles, called oriented anchor boxes, and classifies the rectangles into graspable and ungraspable. On the standard Cornell Grasp Dataset, our model achieves an accuracy of 97.74% and 96.61% on image-wise split and object-wise split respectively, and outperforms the latest state-of-the-art approach by 1.74% on image-wise split and 0.51% on object-wise split.