Abstract

Registering 2 or more range scans is a fundamentalproblem, with application to 3D modeling. While this prob-lem is well addressed by existing techniques such as ICPwhen the views overlap significantly at a good initializa-tion, no satisfactory solution exists for wide baseline registration. We propose here a novel approach which leverages contour coherence and allows us to align two wide base-line range scans with limited overlap from a poor initialization. Inspired by ICP, we maximize the contour coherenceby building robust corresponding pairs on apparent con-tours and minimizing their distances in an iterative fashion.We use the contour coherence under a multi-view rigid reg-istration framework, and this enables the reconstruction ofaccurate and complete 3D models from as few as 4 frames. We further extend it to handle articulations, and this allowsus to model articulated objects such as human body. Experimental results on both synthetic and real data demonstrate the effectiveness and robustness of our contour coherence based registration approach to wide baseline range scans, and to 3D modeling.