readme.txt

快速高斯滤波工具箱（fast guassian toolbox）

The zip file includes:
readme.txt ---> this file
gaussian.c ---> a c file which has to be mex-compiled
prova.m    ---> an M-file to test the routine


Gaussian.m is a 2D and 1D recursive implementation of the Gaussian
filter. This implementation yields an infinite impulse response 
filter that has 6 MADDs per dimension independent of the value
of sigma in the Gaussian kernel.

The 2D Gaussian filter has the following form:

          1
-------------------  exp [ - ( x^2 + y^2 ) / (2 * sigma ^2 ) ]
    2*pi*sigma^2


The 1D Gaussian filter has the following form:

          1
-------------------  exp [ - ( x^2 ) / (2 * sigma ^2 ) ]
  sqrt(2*pi)*sigma


Better results are obtained if input image is padded with a 
proper border of zeros.


Examples of use.

y1=gaussian(image,sigma);

This performs convolution between "image" 
(a mono or bi-dimensional array) with a Gaussian
Kernel given the desired Gaussian sigma.

y2=gaussian(image,sigma,padding);

The input image is padded with a border of zeros.


The output is very close to

out=conv2(input_image,gaussian_kernel,'same');

where input_image is the input vector (1- or 2-dimensional)
and gaussian_kernel is the Gaussian filter (1- or 2-dimensional).


References:

For a complete publication list of Lucas J. van Vliet please visit the following URL:
http://www.ph.tn.tudelft.nl/~lucas/publications/papersLJvV.html



-------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------
     Luigi Rosa
     Via Centrale 35
     67042 Civita di Bagno
     L'Aquila --- ITALY
     mobile [phone]
     email [e-mail]
     website http://utenti.lycos.it/matlab
-------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------