Robot Learning Manipulation Action Plans by "Watching" Unconstrained Videos from the World Wide Web
From Yezhou Yang, Yi Li, Cornelia Fermuller and Yiannis Aloimonos:
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.
The list of the grasping types.
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.... ( article at Kurzweilai.net ) ( original paper )
Records 5641 to 5641 of 5641
Supply Chain - Featured Product
NORD Delivers Powerful, Reliable Drive Solutions for Demanding Packaging Applications
When you need quality, flexibility and high efficiency, NORD has you covered with complete drive solutions for primary and secondary packaging, end-of-line packaging, and higher-level applications. NORD supports you through the entire life cycle of your drive systems with pre-sales engineering support, online tools, and a global service network. Drive systems for both vertically and horizontally networked technology with all the options for sensor incorporation and integration into the higher-level control system.
