sift readme

Contents: 

1. Intro 

2. Requirements 

3. Running 

4. License 

 

 

1. Intro 

 

This is a collection of code I've put together to detect SIFT features in  

images.  It includes a SIFT function library as well as some executables  

to detect, match, and display keypoints.  For more information on SIFT,  

refer to the paper by Lowe: 

 

Lowe, D. Distinctive image features from scale-invariant keypoints.  

International Journal of Computer Vision, 60, 2 (2004), pp.91--110.

 

Or see Lowe's website:
http://www.cs.ubc.ca/~lowe/keypoints/
 

Some of the code also works with affine-invariant features from the code

by the VGG at oxford:
http://www.robots.ox.ac.uk:5000/~vgg/research/affine/index.html 
 

Also included is a library containing functions to compute transforms from  

image features using RANSAC.  See match.c for an example of how to use

the RANSAC functions. 

 

/************************************************************************/ 

 

2. Requirements 

 

All code in this package requires the OpenCV library (known working

version is 1.0.0): 
http://sourceforge.net/projects/opencvlibrary/ 
 

Some functions require the Gnu Scientific Library (GSL) (known working

version is 1.8): 

(Linux) http://www.gnu.org/software/gsl/
(Windows) http://gnuwin32.sourceforge.net/packages/gsl.htm 
 

/************************************************************************/ 

 

3. Running

 

The following is a list of directories that contain VC++.NET projects

and the functions of the projects:

 

  *siftFeat    -- Detect SIFT features

  *match    -- Find and match SIFT features in two images

  *dspFeat    -- Display features from a file

 

You can mess with the global variables in each of the main C files to

adjust parameters, including input and output files.  You might also have

to adjust the project properties depending on how you installed OpenCV 

and GSL. 

 

 

/************************************************************************/ 

 

4. License 

 

The SIFT algorithm is Patented in the U.S. by the University of British  

Columbia.  Thus, the SIFT feature detection code in this package may not  

be used in any commercial products without permission from UBC.  All other  

code in this package is Licensed under the GPLv2.  See the files  

LICENSE.ubc.txt and LICENSE.txt for more info.