Year: 2008
Type: book-chapter
Abstract: A number of issues need to be taken into consideration for the output of an augmented reality application to appear realistic. Such issues include the level of detail of the virtual scene, the accuracy of alignment between the two worlds and the correspondence of illumination between the real and virtual components. This paper reviews registration techniques used to perform the alignment process and discusses a research project that investigates methods matching light conditions. Although a number of registration techniques exist, feature based registration has not yet achieved popularity within production environments because of the associated difficulties. Instead, fiducial marker based systems are being implemented for their robustness and ease of implementation. Further work to be undertaken is discussed towards the end of the paper.
Authors: Mark Bingham, Deanna L. Taylor, Duke Gledhill, Zhijie Xu
Institution: University of Huddersfield
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2125536262