object-classification-experiments

This project 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. (Note: there are technical differences with the original Serre et al, implementation ). 1. Unsupervised normalized graph cut (Shi-Malik algorithm) based clustering technique (Split-Merge approach). This can be used for grouping images returned by a search engine into categories close to the perceived ones. 1. Downloader utility for downloading images from Google Image Search. The code was used in experiments for the thesis entitled "Explorations on a neurologically plausible model of image object classification". Details: http://www.cse.iitk.ac.in/users/vision/sourabhd/index.html

object-classification-experiments
loadvectors.cpp
sdd_include.h
CarClassifier2.cpp
ParameterExtractor.h
GaborFilterModel2.cpp
ncut.m
dir.h
README
SPCarFilter.cpp
ParameterExtractor.cpp
GaborFilterModel.cpp
dir.cpp
.svn
downloadCarDataset.py
Makefile
display.m
CarClassifier1.cpp
jaguar.conf
GaborFilterModel.h
emd.c
demo.conf
AdjMatrix.h
createAdjMatrix.cpp
emd.h
caltech101.conf
gpl.txt