Changes between Initial Version and Version 1 of libpipi/research/filters


Ignore:
Timestamp:
08/18/2008 12:47:55 PM (16 years ago)
Author:
Sam Hocevar
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • libpipi/research/filters

    v1 v1  
     1== Gaussian kernels ==
     2
     3The usual way to create a Gaussian kernel is to evaluate a Gaussian function at the center of each cell:
     4
     5 k[i][j] = exp(-(i²+j²)/2σ)
     6
     7This usually works well, except when the kernel is thin (σ < 1). It gets worse when using our generalised kernel:
     8
     9 k[i][j] = exp(-((i×cosθ-j×sinθ-dx)²+(j×cosθ+i×sinθ-dy)²)/2σ)