Learning Category-Level Manipulation Tasks from Point Clouds with Dynamic Graph CNNs

This paper presents a new technique for learning category-level manipulation from raw RGB-D videos of task demonstrations, with no manual labels or annotations. Category-level learning aims to acquire skills that can be generalized to new objects, with geometries and textures that are different from the ones of the objects used in the demonstrations. We address this problem by first viewing both grasping and manipulation as special cases of tool use, where a tool object is moved to a sequence of key-poses defined in a frame of reference of a target object. Tool and target objects, along with their key-poses, are predicted using a dynamic graph convolutional neural network that takes as input an automatically segmented depth and color image of the entire scene. Empirical results on object manipulation tasks with a real robotic arm show that the proposed network can efficiently learn from real visual demonstrations to perform the tasks on novel objects within the same category, and outperforms alternative approaches.

Authors

• Junchi Liang
• Abdeslam Boularias

Keywords

• Point cloud compression
• Visualization
• Image segmentation
• Shape
• Manuals
• Grasping
• Predictive models
• Point Cloud
• Manipulation Tasks
• Dynamic Graph
• Graph CNN
• Neural Network
• Convolutional Neural Network
• Use Of Tools
• Reference Frame
• Target Object
• Depth Images
• Robotic Arm
• Graph Convolutional Network
• Manual Labeling
• Real Robot
• Dynamic Neural Network
• Raw Video
• Prediction Accuracy
• Coordinate System
• K-nearest Neighbor
• Object Pairs
• Inverse Reinforcement Learning
• Number Of Objects
• High-level Policy
• Objects In The Scene
• Imitation Learning
• Camera Pose
• Unknown Objects
• Vector Of Size
• Video Presentation