Joint estimate of disparity and optical flow in a stereo camera setup.
Temporally coherent unambigous results.
Occlusion boundaries are well preserved, no smoothing artifacts.
Allows large displacement between frames (>30 pixels).
Low computational complexity due to efficient seed growing [2, 3].
Download
Code (a software package as a Matlab toolbox, source code of binaries available),
please cite us by [1].
Example
The video shows resulting disparity and optical flow maps of our GCSFs
algorithm [1] and a comparison with spatiotemporal stereo by
Sizintsev [4] and variational optical flow by Brox [5].
The maps are color coded. For disparity, warmer colors are closer to
the camera. In optical flow, green color is zero motion, warmer colors
is left and up motion, colder colors is right and down motion
respectively. Black color denotes unmatched pixels.
The video contains 3 sequences: INRIA and INRIA 2, which are from CAVA
dataset of INRIA , and ETH, which is from a pedestrian dataset of ETH Zurich .
Please, find the comments on our results and comparison with other
methods in the paper [1] in Section 3.2.
References
[1]
Jan Cech, Jordi Sanchez-Riera, Radu Horaud.
Scene Flow Estimation by Growing Correspondence Seeds. In Proc. CVPR , 2011. [pdf]
[bib]
[2]
Jan Cech, Jiri
Matas, Michal Perdoch. Efficient Sequential Correspondence Selection
by Cosegmentation. IEEE Trans. on PAMI, 32(9), 2010. [pdf]
[bib] [software]
[3]
Jan Cech, Radim
Sara. Efficient Sampling of Disparity Space for Fast and Accurate
Matching. In Proc. BenCOS Workshop CVPR, 2007. [pdf]
[bib] [software]
[4]
M. Sizintsev and R. P. Wildes.
Spatiotemporal stereo via spatiotemporal quadratic element (stequel) matching. In Proc. CVPR , 2009.
[5]
T. Brox and J. Malik.
Large displacement optical flow: de-
scriptor matching in variational motion estimation. IEEE Trans. on PAMI , 2010.
