Issue No. 02 - February (2012 vol. 34)
J. Denzler , Lehrstuhl fur Digitale Bildverarbeitung, Friedrich-Schiller-Univ. Jena, Jena, Germany
O. Kahler , Dept. of Eng. Sci., Univ. of Oxford, Oxford, UK
We present a novel approach to the structure-from-motion problem which combines the search for correspondences and geometric reconstruction, rather than treating these as separate steps. Through the combination of the two steps, we achieve an implicit feedback of 3D information to aid the correspondence search, and at the same time we avoid an explicit model for tracking errors. The reconstruction results are therefore optimal in case of, for example, Gaussian noise on image intensities. We also present an efficient online framework for structure-from-motion with our combined approach, thoroughly evaluate the method in experiments and compare the results to state-of-the-art methods.
solid modelling, image reconstruction, optimisation, combined optimization approach, structure from motion problem, geometric reconstruction, 3D information, tracking errors, Gaussian noise, image intensities, online framework, Three dimensional displays, Image reconstruction, Tracking, Cameras, Mathematical model, Estimation, local minima., Structure-from-motion, direct reconstruction, constrained tracking
J. Denzler, O. Kahler, "Tracking and Reconstruction in a Combined Optimization Approach", IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 34, no. , pp. 387-401, February 2012, doi:10.1109/TPAMI.2011.141