[Insight-users] Local texture Features

[Nicholas Tustison]

Hi Julien,

I just saw your email regarding the usage of the run length feature
filters I wrote for the Insight Journal.  If it helps, I just made that
contribution itkv4 compliant and uploaded it as a new revision.

Also, FYI, people have used the stochastic fractal dimension image
filter to calculate texture measurements described 

http://www.insight-journal.org/browse/publication/318

but currently available in the ITK directory.

Nick



On Apr 29, 2011, at 5:22 AM, Julien Malik wrote:

> Hi,
> 
> We have filters to compute local textures descriptors in Orfeo Toolbox, based on ITK classes.
> 
> Code is here :
> http://hg.orfeo-toolbox.org/OTB/file/tip/Code/FeatureExtraction
> 
> The filters computing local textures around a neighborhood are :
> otb::ScalarImageToTexturesImageFilter (based on itk::GreyLevelCooccurrenceMatrixTextureCoefficientsCalculator)
> otb::ScalarImageToAdvancedTexturesImageFilter
> otb::ScalarImageToHigherOrderTexturesImageFilter (based on http://www.insight-journal.org/browse/publication/231)
> 
> The filters are streamed and multithreaded, so they can be applied on images of arbitrary size.
> 
> Cheers,
> Julien
> 
> 
> 
> Le 29/04/2011 10:42, M.Staring at lumc.nl a écrit :
>> Hi,
>> 
>> The following at least provides code to get plain GLCM operators working:
>> 
>> https://github.com/ITKTools/ITKTools/tree/master/src/texture
>> 
>> HTH,
>> 
>> Marius
>> 
>> -----Original Message-----
>> From: insight-users-bounces at itk.org [mailto:insight-users-bounces at itk.org] On Behalf Of Neal R. Harvey
>> Sent: donderdag 28 april 2011 18:15
>> To: insight-users at itk.org
>> Subject: Re: [Insight-users] Local texture Features
>> 
>> Neil Panjwani Wrote:
>> 
>>> Hi,
>>> 
>>> I'm interested in computing local texture maps. I found a few filters
>>> in ITK, including the
>>> GreyLevelCooccurrenceMatrixTextureCoefficientsCalculator,
>>> which does compute many of the features I'm interested in. However, it
>>> does so over an entire image, a masked image, or given a cooccurrence
>>> matrix (which is usually generated from the entire image) and only
>>> returns one texture value for each feature given that input. I'd like
>>> to compute these features locally, in say a 9x9 neighborhood much like
>>> the MATLAB function
>>> entropyfilt() (
>>> http://www.mathworks.com/help/toolbox/images/ref/entropyfilt.html)
>>> 
>>> How can I utilize these filters to compute local features?
>>     I am also very interested in doing exactly the same thing.
>>     For example, if I have a segmentation and a number of objects
>>     labeled in an image, I would like to get texture measures for each
>>     individual object.
>> 
>>     When I looked into this I wasn't even able to get the code for the
>>     ITK plain vanilla GLCM operators working properly, let alone
>>     something more complicated. In the end, I went to OpenCV and
>>     modified some of the code there to do what I needed.
>> 
>>     If anyone can provide examples of code that successfully implements
>>     the ITK GLCM operators, that can provide results that correlate well
>>     with applying some other GLCM tools such as those available in
>>     matlab, it would be very much appreciated. If anyone is able to
>>     provide a solution to the more complicated problem as elucidated by
>>     Mr Panjwani, that would be wonderful.
>> 
>>     Cheers
>> 
>> 
> _____________________________________
> 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