Automatic 3D Object Reconstruction from a Single Image

Bryson R. Payne, James F. Lay, Markus A. Hitz
University of North Georgia
Department of Computer Science and Information Systems

The goal of our research is to develop a robust, fully automatic, unsupervised system for recognizing and reconstructing various classes of 3D objects in single-object, 2D images.

Our system currently (as of March 2014) takes a single 2D image as input and provides a correctly textured 3D VRML, X3D, or WebGL file for two classes of objects: boxes and spheres.

Click on any of the links below to find sample output in VRML (boxes), X3D (spheres), or our one example with WebGL above.

• Daylight (VHS tape box) - JPG file - WRL file - Full Folder
• Frosted Flakes (cereal box) - JPG file - WRL file - Full Folder
• Rubik's Cube - JPG file - WRL file - Full Folder
• Volleyball (sphere example) - JPG file - X3D file - Full Folder

Note: To view these samples, you'll need a free VRML/X3D browser. Try FreeWRL http://sourceforge.net/projects/freewrl/.

©2014 Payne, Lay & Hitz, University of North Georgia.