Familar Object Grasping Object Viiew Recog

  • Shafii, Nima, S. Hamidreza Kasaei, and Luís Seabra Lopes. “Learning to grasp familiar objects using object view recognition and template matching.” IROS 2016.
  • Grasp template :
  • Local shape feature for graspable regions
    • spin-image feature
  • Global feature, the radius (distance from the CoM to the selected keypoint)
    • important to represent a stable grasp
  • Finger configuration
  • New objects that are similar to known ones (i.e. they are familiar) can be grasped in a similar way
    • As an example, if the robot knows how to grasp a pen, it may use the same grasp temple to take a marker