Image clusters based 3D virtual tour schema

Abstract

This paper presents a 3D virtual tour schema based on image clusters. With the development of the Internet, various image-based applications have generated large amounts of image data. Vast amounts of image data can be effectively organized by clustering algorithms according to their geographic location information and content, thus forming image clusters. These data then can be reconstructed for 3D virtual tour using the state-of-the-art computer vision methods. This paper analyzes the popular photo tours application and Photosynth, and proposes an image clusters based 3D virtual tour schema on the basis of these applications. This paper also points out the room for improvement in the future, and several possible applications based on this schema are discussed finally.