GRConvNet

• Generative Residual Convolutional Neural Network
• Learning an object agnostic function to grasp objects
• Uses multi modal input data (RGB + depth images).
• Generates pixel-wise antipodal grasp configuration.
• State-of-the-art performance (97% on Cornell dataset).
• Use eye-to-hand camera configuration.
• Sulabh Kumra, et al. “Antipodal robotic grasping using generative residual convolutional neural network.” IROS 2020.
•