TITLE:
Image-Based Methods for Interaction with Head-Worn Worker-Assistance Systems
AUTHORS:
Frerk Saxen, Omer Rashid, Ayoub Al-Hamadi, Simon Adler, Alexa Kernchen, Rüdiger Mecke
KEYWORDS:
Head Mounted Display, Gesture Recognition, Mobile Assistance System
JOURNAL NAME:
Journal of Intelligent Learning Systems and Applications,
Vol.6 No.3,
August
11,
2014
ABSTRACT:
In this paper, a
mobile assistance-system is described which supports users in performing manual
working tasks in the context of assembling complex products. The assistance
system contains a head-worn display for the visualization of information
relevant for the workflow as well as a video camera to acquire the scene. This
paper is focused on the interaction of the user with this system and describes
work in progress and initial results from an industrial application scenario.
We present image-based methods for robust recognition of static and dynamic
hand gestures in realtime. These methods are used for an intuitive
interaction with the assistance-system. The segmentation of the hand based on
color information builds the basis of feature extraction for static and dynamic
gestures. For the static gestures, the activation of particular sensitive
regions in the camera image by the user’s hand is used for interaction. An HMM
classifier is used to extract dynamic gestures depending on motion parameters
determined based on the optical flow in the camera image.