[Main]  [Publications]  [Abstracts and Posters]  [Software]   [Awards and Scholarships]  [Computer Vision Related Notes]  [Links]

 

IMPORTANT: All downloadable material listed on this page is available for public use. Please note that while great care has been taken, the software, code and data are provided "as is" and that the author does not accept any responsibility or liability.

 

Comments? Suggestions? Bugs?  Feel free to send me email at kosta at cs dot yorku dot ca

Table of contents:

 

Generalized Integral Image C++ code

The following C++ are implementation of the generalized integral image detailed in "Fast Scale-Space Feature Representations by Generalized Integral Images" by Konstantinos G. Derpanis, Erich, T.H. Leung, and Mikhail Sizintsev, International Conference on Image Processing, 2007.

 

Generalize Integral Image: C++ code

.

3D G2/H2 Separable Steerable Filters Matlab code

The following Matlab scripts are implementations of the 3D G2 and H2  separable steerable filters detailed in "Three-Dimensional nth Derivative of Gaussian Separable Steerable Filters" by Konstantinos G. Derpanis and Jacob M. Gryn, Technical Report CS-2004-05 York University, 2004.

 

3D G2 / H2 Steerable library: Matlab code

 

2D G2/H2 Separable Steerable Filters Matlab code

The following Matlab scripts are implementations of the 2D G2 and H2 separable steerable filters detailed in "The Design and Use of Steerable Filters" by William T. Freeman and  Edward H. Adelson, IEEE Transactions on Pattern Analysis and Machine Intelligence, 1991.

 

 
The three leftmost images represent the modulated G2 basis filters and the rightmost represents the resulting steered G2 filter.

 

Initialize G2 basis filters (Matlab, Last updated April 20, 2005.): imgG2Init.m

G2 steer image (Matlab, Last updated April 20, 2005.): imgG2theta.m

Intialize H2 Filters (Matlab, Last updated April 20, 2005.): imgH2Init.m

H2 steer image (Matlab, Last updated April 20, 2005.): imgH2theta.m

 

Example application (November 20, 2004): Movement Without Motion

 

Pyramid Matlab code

The following Matlab scripts are efficient implementations of the "expand" and "reduce" routines detailed in "The Laplacian Pyramid as a Compact Image Code" by Peter Burt and Edward Adelson, IEEE Transactions on Communications, 1983.

 

Expand (Matlab, Last updated August 11, 2004.): imgExpand.m

Reduce (Matlab, Last updated August 11, 2004.): imgReduce.m

 

Level set segmentation Matlab code

Chan-Vese intensity segmentation (Matlab, Last updated February 21, 2009.): zip

Cremers-Soatto two-frame motion segmentation (Matlab, Last updated February 21, 2009.): zip

Cremers-Soatto (multi-frame) spacetime motion segmentation (Matlab, Last updated February 21, 2009.): zip

 

PVN video related code

PVN is a video file format loosely based on the PGM format for images.  Further information about the format can be found here.

 

Read/write (Matlab): coming soon

Virtualdub plugin: zip

 

Zone plate Matlab code

The following Matlab scripts create 2D and 3D zone plates (i.e., 2D and 3D Matlab matrices, respectively).  The zone plates are based on generalizations of a 1D chirp signal.

 

Create 2D zone plate (Matlab, Last updated July 23, 2004.): zoneplate2D.m

Create 3D zone plate (Matlab, Last updated July 23, 2004.): zoneplate3D.m

 

1394 Digital Camera  Driver

CMU IEEE-1394 Digital Camera Windows Driver

 

Matching Pursuit demo

Demo

 

Affine video stabilization demo

Demo

 

Panoramic mosaic demo

Demo

 

 

Last Updated: March 18, 2009