Written by Shai Bagon (shaibagon@gmail.com)
This package implements the algorithm for Sketching the Common as described in:
Shai Bagon, Or Brostovsky, Meirav Galun and Michal Irani
"Detecting and Sketching the Common", CVPR 2010
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
If you wish to use this software (or the algorithms described in the aforementioned paper) for commercial purposes, you should be aware that there is a US patent:
Eli Shechtman and Michal Irani
"METHOD AND APPARATUS FOR MATCHING LOCAL SELF-SIMILARITIES"
If you use this package you HAVE to reference the aforementioned CVPR 2010 paper.
Here is a bibtex for that paper:
@INPROCEEDINGS{Bagon2010,
author={Shai Bagon and Or Brostovski and Meirav Galun and Michal Irani},
booktitle={Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on},
title={Detecting and Sketching the Common},
year={2010},
month=june,
url={http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=5540233&tag=1},
pages={33--40},
doi={10.1109/CVPR.2010.5540233},
ISSN={1063-6919}
}
This package contains Linux implementation only. It is unlikely that a Windows/Mac/Unix versions will be added in the future. (if you wish to create such a version by yourself, you may contact me at shaibagon@gmail.com for some assistance).
The software was tested with Matlab version 2008b and 2010b with 64bit Linux machines.
Pre-requisite:
You need to have opencv shared libraries installed on your machine, and
the enviromental variable LD_LIBRARY_PATH pointing to the installation directory.
1. Extract the files from the gz tarball to a folder of your choice
(for example /home/bagon/sketch_cvpr2010).
->% cd /home/bagon/sketch_cvpr2010
->% tar -zxf SketchCommonCVPR10_1.0.tar.gz
2. Open Matlab and compile cpp code for mex function:
>> cd /home/bagon/sketch_cvpr2010 % or where you extracted the package to
>> mexall
Make sure your Matlab is configured with gcc compiler.
If not, you have to configure your mex compiler BEFORE running mexall.
To configure mex type:
>> mex -setup
Matlab should be able to guide you through the rest of the configuration process.
1. The package comes with a simple usage example.
In Matlab type:
>> cd /home/bagon/sketch_cvpr2010 % or where you extracted the package to
>> Sketch_example
When this functions terminates, you should see a figure with the input images
and a figure with the resulting output sketch.
2. Main function.
The main sketching function is sketch_the_common_CVPR10.m
Use
>> doc sketch_the_common_CVPR10
To see how to use this function.
The following Matlab functions are part of this package:
0. mexall.m
Mexing the associated cpp code into mex files and installs the package.
***********************************************************************
If you have warnings/errors running this code, you might not be able to
Use this package!
***********************************************************************
1. Sketch_example.m
Usage example using the thumbnail images in the example folder
2. sketch_the_common_CVPR10.m
Main sketching function
3. self_sim_correlation_matrix.m
Taking images cropped around the common part (and optionally an exact center
of each detected object) and computes the "self-similarity correlation matrix"
for the common ensemble of descriptors.
4. self_sim_win.m
A matlab wrapper for the executable that computes local self-similarity descriptors
5. self_sim_connectivity.m
Taking an ensemble of descriptors represented by an array of descriptors and their
spatial locations, and computes the "self-similarity correlation matrix"
6. LaplacianLinf.m
Defines the quadratic programming (QP) optimization problem that arises from the
sketching functional and optimize it using Matlab's quadratic programming function.
7. my_tformarray.m
Auxilary function for translating all thumbs to be aligned according to the center of
the common part.
8. loc_offset_mex.cpp
Auxilary function that computes the relative offsets between the descriptors in the ensemble