Prior-less Compressible Structure from Motion

Information

This is the release of the Prior-less Compressible Structure from Motion, described in the paper http://ci2cv.net/media/papers/Chen-CVPR-2016.pdf.

Project webpage: http://www.cs.cmu.edu/~chenk/projects/cvpr_2016.html.

Code git: git@github.com:kongchen1992/compressible-sfm.git.

For questions concerning the code please contact Chen Kong at chenk@cs.cmu.edu and Simon Lucey slucey@cs.cmu.edu. The code was developed when Chen Kong and Simon Lucey were at Carnegie Mellon University.

Cite

If you use this code in your research, please cite: @article{KongCVPR16, title = {Prior-Less Compressible Structure from Motion}, author = {Kong, Chen and Lucey, Simon}, booktitle={Computer Vision and Pattern Recognition (CVPR)}, year = {2016}, organization={IEEE} }

Folders

CamRecovery/ contains the necessary code for recovering camera motions and 3D structures.

Exps/ contains the necessary code for various experiments mentioned in the paper.

FacMethod/ contains the necessary code for block sparse dictionary learning including block KSVD, block OMP, and block FOCUSS.

Viz/ contains the necessary code for visualization.