Robot Learning Manipulation Action Plans by "Watching" Unconstrained Videos from the World Wide Web

TLDR

Robots need computational tools that automatically interpret and represent human actions beyond simple learned schemas. This paper presents a system that learns manipulation action plans by processing unconstrained World Wide Web videos to robustly generate the sequence of atomic actions of longer actions for robot knowledge acquisition. The system employs two CNN-based modules—one for hand grasp classification and one for object recognition—followed by a probabilistic manipulation action grammar parsing module that generates visual sentences for robot manipulation. Experiments on a publicly available unconstrained video dataset show the system can learn manipulation actions by watching videos with high accuracy.

Abstract

In order to advance action generation and creation in robots beyond simple learned schemas we need computational tools that allow us to automatically interpret and represent human actions. This paper presents a system that learns manipulation action plans by processing unconstrained videos from the World Wide Web. Its goal is to robustly generate the sequence of atomic actions of seen longer actions in video in order to acquire knowledge for robots. The lower level of the system consists of two convolutional neural network (CNN) based recognition modules, one for classifying the hand grasp type and the other for object recognition. The higher level is a probabilistic manipulation action grammar based parsing module that aims at generating visual sentences for robot manipulation. Experiments conducted on a publicly available unconstrained video dataset show that the system is able to learn manipulation actions by ``watching'' unconstrained videos with high accuracy.