[Insight-users] AdaptiveHistgramEqualization too slow

[Luis Ibanez]

Hi Zhenchao,


1) The itkAdaptiveHistogramEqualizationImageFilter took

                        9 minutes 25 seconds

     to run in an input image of size:  3072x2304

     in a laptop LATITUDE E6500, with processor:
     Intel(R) Core(TM)2 Duo CPU T9600  @ 2.80GHz
     Running Ubuntu 9.10,
     with code compiled with gcc 4.4 in Release mode.


     with the following parameters:

                 Radius =   10
                 Alpha   =   1.0
                 Beta     =   0.25


      The radius parameter is obviously critical here.

      You should anticipate that the computation time
      is proportional to the square of the radius
      (for 2D images)

      A radius of 10 is using a region of 21x21 pixels
      for performing the local computations.


2)   As a reference: Using a radius value of 2,
      leads the filter to run in 25 seconds, on the
      same input image.


What value of radius are you using with the ITK filter ?

What value of radius are you using with Matlab ?

and

Very important:
have you compiled your code for "Release" mode ?



     Regards,


              Luis


------------------------------------------------
On Fri, Mar 5, 2010 at 4:32 AM, Zhenchao Li <cockneykevin at gmail.com> wrote:
> Hi,
>    I'm using itk's builtin adaptive histogram equalization to equalize an
> image(3000*3000 in size), with the default arguments the process seems to
> run forever. However if I use matlab's adaptivehisteq function the result
> returns immediately. After googling I find that matlab uses an algorithm
> called contrast limited AHE, while itk uses the original AHE. The difference
> is that the former process the image block by block while the latter
> processes the image globally.(not sure how really it works). So, am I stuck?
> Or is there a way to use CLAHE in itk? The current AHE in itk is really
> slow, any suggestions?
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>
>