Shape from video v.s. still images

In this paper we compare two dense matching approaches. The fi rst one has been developed in the context of shape from video using minimal path search. The se cond one is a PDE-based approach, and would be expected to give better results for shape from (a small number of) still images and for wide baseline situations. Both methods use as input the full y ca ibrated camera parameters, that have been obtained after structure form motion recovery from unc alibrated images. Emphasis lies on the usage of only a small amount of (high resolution) images inst ead of a (low resolution) video sequence. We use ground truth synthetic data as well as real data to comp are the two algorithms.

[1]  Cordelia Schmid,et al.  An Affine Invariant Interest Point Detector , 2002, ECCV.

[2]  Reinhard Koch,et al.  Multi Viewpoint Stereo from Uncalibrated Video Sequences , 1998, ECCV.

[3]  Marc Pollefeys,et al.  Multiple view geometry , 2005 .

[4]  Reinhard Koch,et al.  Metric 3D Surface Reconstruction from Uncalibrated Image Sequences , 1998, SMILE.

[5]  Adam Baumberg,et al.  Reliable feature matching across widely separated views , 2000, Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662).

[6]  Paul A. Beardsley,et al.  Euclidean Structure from Uncalibrated Images , 1994, BMVC.

[7]  Anders Heyden,et al.  Euclidean reconstruction from image sequences with varying and unknown focal length and principal point , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[8]  Luc Van Gool,et al.  PDE-based multi-view depth estimation , 2002, Proceedings. First International Symposium on 3D Data Processing Visualization and Transmission.

[9]  M. Vergauwen,et al.  A hierarchical stereo algorithm using dynamic programming , 2001, Proceedings IEEE Workshop on Stereo and Multi-Baseline Vision (SMBV 2001).

[10]  Luc Van Gool,et al.  Dense matching of multiple wide-baseline views , 2003, Proceedings Ninth IEEE International Conference on Computer Vision.

[11]  Luc Van Gool,et al.  Wide Baseline Stereo Matching based on Local, Affinely Invariant Regions , 2000, BMVC.

[12]  Reinhard Koch,et al.  A simple and efficient rectification method for general motion , 1999, Proceedings of the Seventh IEEE International Conference on Computer Vision.