[Insight-users] OtsuThresholdImageFilter

[Luis Ibanez]

Hi Radhika,

Sorry for the ambiguity of my email,

When I said

  "Note that there is a point where this is actually
   a problem for a Gaussian Mixture Model."


I meant that the Gaussian Mixture Modeling approach
was indeed suitable for solving the "problem" of
classifying your image in three intensity classes.


    Regards,


        Luis



-----------------------------------
Radhika Sivaramakrishna wrote:

> Hi Luis,
> An OtsuMultipleThresholdImageFilter would be a great idea where we can
> Pass in the input image and the number of expected thresholds, and it
> will
> do the whole thing and return a labeled image. Yes, so I would love to
> have such a filter. 
> 
> Thanks for your suggestions on the EM model and Markov random fields.
> However, I didn't understand your comment regarding the problem with the
> GMM. Could you explain?
> 
> Thanks
> Radhika
> 
> 
> 
> -----Original Message-----
> From: Luis Ibanez [mailto:luis.ibanez at kitware.com] 
> Sent: Thursday, October 28, 2004 9:28 AM
> To: Radhika Sivaramakrishna
> Cc: insight-users at itk.org
> Subject: Re: [Insight-users] OtsuThresholdImageFilter
> 
> 
> Hi Radhika,
> 
> It seems that what you want is a sort of
> 
>     OstsuMultipleThresholdImageFilter
> 
> that is, an image filter that will package
> the entire operation for you. Including,
> the instatntiation and computation of a
> histogram, the instantiation and computation
> of an OtsuMultipleThreshold Calculator and
> the actual filtering of the image in order
> to produce a labeled image.
> 
> Please confirm is this is what you are looking
> for, and in that case we can put together such
> a filter, since this seems to be a functionality
> that will be useful for many users.
> 
> Note that there is a point where this is actually
> a problem for a Gaussian Mixture Model. In which
> case, there are other classes in ITK that are well
> suited for solving that classification problem.
> 
> 
> In fact, there is a "Expectation Maximization
> Mixture Model Estimation" example in the ITK
> Software Guide
> 
>        http://www.itk.org/ItkSoftwareGuide.pdf
> 
> Section 10.3.3, pdf-page 475,
> 
> 
> Yet another option is to use Markov Random Fields,
> for which ITK have several filters too.
> 
> 
> 
>     Regards,
> 
> 
> 
>        Luis
> 
> 
> 
> -------------------------------
> Radhika Sivaramakrishna wrote:
> 
>>Hi Luis,
>>Thanks for your prompt response. Yes, I did have a couple of
> 
> questions.
> 
>>1) I checked the documentation for OtsuMultipleThresholds. It is
>>templated on input histogram. It says you have to specify the target
>>histogram. What does this mean? Is there a way to just pass the
> 
> original
> 
>>image with the number of thresholds you want to identify, and it
> 
> should
> 
>>return the thresholds that best separate the multiple classes?
>>
>>2) Do you have recommendations for other types of filters/methods I
> 
> can
> 
>>use to separate an image into essentially three portions, a background
>>which is darker, a foreground, another region within foreground which
> 
> in
> 
>>general is brighter than the rest of the foreground.
>>
>>Thanks
>>Radhika
>>
>>
>>-----Original Message-----
>>From: Luis Ibanez [mailto:luis.ibanez at kitware.com] 
>>Sent: Wednesday, October 27, 2004 3:41 PM
>>To: Radhika Sivaramakrishna
>>Cc: insight-users at itk.org
>>Subject: Re: [Insight-users] OtsuThresholdImageFilter
>>
>>
>>Hi Radhika,
>>
>>
>>In the Doxygen documentation, there are entries for the three classes
>>
>>
> 
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1OtsuThresholdImageFi
> 
>>lter.html
>>
> 
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1OtsuMultipleThreshol
> 
>>dsCalculator.html
>>
> 
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1OtsuThresholdImageCa
> 
>>lculator.html
>>
>>
>>The classes
>>
>>        OtsuMultipleThresholdsCalculator
>>        OtsuThresholdImageFilter
>>
>>are relatively new,
>>
>>
>>the first one dates from March, (contributed by Luca Antiga)
>>the second dates from July
>>
>>
>>
>>OtsuMultipleThresholds allows you to compute thresholds from
>>gray scale statistical distributlis that are multimodal, while
>>the original OtsuThreshold calculator was intended for Bimodal
>>distributions.
>>
>>
>>The OtsuThresholdImageFilter it a thresholding filter that
>>internally computes the threshold for you image by using
>>an OtsuThresholdCalculator.
>>
>>
>>
>>   Please let us know if you have further questions,
>>
>>
>>     Thanks
>>
>>
>>       Luis
>>
>>
>>
>>------------------------------
>>Radhika Sivaramakrishna wrote:
>>
>>
>>
>>>Hi all,
>>>
>>>In the documentation only the OtsuThresholdImageCalculator is written.
>>
>>
>>>However, in the Insight library, we also have 
>>>OtsuMultipleThresholdsCalculator and OtsuThresholdImageFilter.
>>>
>>>Does anyone know how to use these, and why are they not available in
>>
>>the 
>>
>>
>>>documentation?
>>>
>>>Thanks
>>>
>>>Radhika
>>>
>>>
>>>
>>>
>>>
>>
>>
> ------------------------------------------------------------------------
> 
>>>_______________________________________________
>>>Insight-users mailing list
>>>Insight-users at itk.org
>>>http://www.itk.org/mailman/listinfo/insight-users
>>
>> 
>>--------------------------------------------------------
>>
>>This email message is for the sole use of the intended recipient(s)
> 
> and may contain confidential and privileged information.Any unauthorized
> review, use, disclosure or distribution is prohibited. If you are not
> the intended recipient, please contact the sender by reply email and
> destroy all copies of the original message. If you are the intended
> recipient, please be advised that the content of this message is subject
> to access, review and disclosure by the sender's Email System
> Administrator.
> 
>>--------------------------------------------------------
>>
>>
> 
>  
> --------------------------------------------------------
> 
> This email message is for the sole use of the intended recipient(s) and may contain confidential and privileged information.Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message. If you are the intended recipient, please be advised that the content of this message is subject to access, review and disclosure by the sender's Email System Administrator.
> --------------------------------------------------------
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
> 
>