NSF Award
9961284
This Small Business Innovation Research Phase I project has as its primary objective to develop a fully automatic and intelligent software tool that is able to mosaic (i.e., align and merge) multiple 3D images of the same object taken from different viewpoints, without a priori knowledge of camera positions. Three-dimensional (3D) modeling of physical objects and environment is an essential part of the challenges for many multimedia tasks. However, most physical objects self occlude, and no single view 3D image suffices to describe the entire surface of a 3D object. Multiple 3D images of the same object or scene from various viewpoints have to be taken and integrated in order to obtain a complete 3D model of the 3D object or scene. We call this process the '3D mosaic'. The main innovations of this proposed effort are threefold:<br/><br/>1. an intelligent alignment method that is able to register multiple uncalibrated 3D images without needing a priori knowledge of camera location and orientation;<br/><br/>2. a seamless merge method to 'stitch' together the aligned 3D images using the fuzzy logic principle; and<br/><br/>3. an intelligent 3D image compression algorithm that preserves 3D image geometric features while achieving high compression ratio.<br/><br/>The 3D Mosaic technique to be developed has enormous commercial applications, including industrial design and prototyping, reverse engineering, manufacturing part inspection, part replacement and repair, animation, entertainment, 3D modeling for WWW documents, archiving, virtual reality environment, education, virtual museum, commercial on-line catalogues, etc. If successful, 1t could become an important part of future 3D TV technology.
