Segmenting Objects through an Autonomous Agnostic Exploration Conducted by a Robot

Abstract

Human's everyday environment is an open environment in which objects with new shapes, colors or textures frequently appear. Enabling robots to deal with such environments and to manipulate those objects raises a difficult challenge: how to recognize an object ? How to distinguish it from the background ? An approach is proposed here to allow the robot to find this segmentation on its own. It relies on an active exploration of the environment aimed at identifying features of things that move after a contact with robot's end-effector. The only assumption made is that objects of interest are solid objects that the robot can move. The proposed approach can thus be applied without modifications to a large range of environments, as shown by the experiments performed by the robot.

Self-Supervised Interactive Object Segmentation Through a Singulation-and-Grasping Approach

“…Interactive Segmentation Previous works dealing with the interactive perception problem focus on generating interactions with the environment based on objectness hypotheses and obtaining feedback after applying actions to update the segmentation results in recognition, data collection, and pose estimation tasks [3,14,22,25]. Many methods use a robot manipulator to distinguish one object from the others by applying pre-planned non-prehensile actions to specific object hypothesis [6,26,36]. However, the non-prehensile action, such as pushing, for a specific object is challenging in cluttered environments due to inevitable collisions with other surrounding objects.…”