object-classification-experiments

Automatically exported from code.google.com/p/object-classification-experiments

SerrePoggioClassifier

Author

Sourabh Daptardar saurabh.daptardar@gmail.com

Version

1.0

LICENSE

SerrePoggioClassifier 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 3 of the License, or (at your option) any later version.

SerrePoggioClassifier 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 SerrePoggioClassifier. If not, see http://www.gnu.org/licenses/.

DESCRIPTION

This code has been developed in partial fulfillment of the M.Tech thesis "Explorations on a neurologically plausible model of image object classification" by Sourabh Daptardar, Y7111009, CSE, IIT Kanpur.

Project website: http://www.cse.iitk.ac.in/users/vision/sourabhd/

This code implements:

1. Biologically inspired model of Serre-Wolf-Poggio http://cbcl.mit.edu/projects/cbcl/publications/ps/serre-PID73457-05.pdf and a number of variations with the model.

2. Unsupervised normalized graph cut (Shi-Malik algorithm) based clustering technique (Split-Merge approach).

3. Downloader uitility for downloading images from Google Image Search

PROGRAMMING LANGUAGES

1. C++

2. Octave

3. Python

DEPENDENCIES

1. Boost

2. OpenCV

3. OpenMP (libgomp)

4. OpenMPI

5. liboctave

6. Yossi Rubner's implementation of Earth Mover's distance: http://vision.stanford.edu/~rubner Two files emd.c emd.h included here are from Rubner's code.

USAGE

1. Serre-Poggio model related experiments / generation of C2 feature vectors:

   make

2. Split-Merge graph clustering algorithm:

   make unsupervised

3. Downloader

   python downloadCarDataset.py