[Insight-users] ITK, Random Forests, and OpenCV

Julien Malik julien.malik at c-s.fr
Fri Oct 18 04:32:40 EDT 2013


Hi Jared,

You may look into what we have in Orfeo Toolbox, where we have wrapped
the OpenCV machine learning framework in an ITK-compatible way [1].

There is :
- generic ListSample/CvMat conversion
- MachineLearningModel abstract class with derivated classes for each
machine learning model algorithm, which can be trained from ListSample,
and results can be saved to file and loaded. Once trained or loaded from
file, it predicts classes from input sample.
- an I/O mechanism inspired from the ImageIO framework, capable of
loading an arbitrary machine learning model file and returning a
MachineLearningModel object.
- a generic classification filter working on an abstract
MachineLearningModel and classifying VectorImage.

Hope this helps,
Julien

[1]
http://www.orfeo-toolbox.org/SoftwareGuide/SoftwareGuidech19.html#x45-29300019.3.3


On 10/18/2013 12:42 AM, Becksfort, Jared wrote:
> Matt,
>
> Wow, you'd think I might have googled "convert itk to opencv" at least once in the several hours I looked into this.  The bridge is on the first page of the results.  I guess I spent all my time looking at comparisons between ITK and OpenCV rather than conversions.
>
> Thanks for the help.  I'll see what I can do with it.
>
> Jared
>
> -----Original Message-----
> From: Matt McCormick [mailto:matt.mccormick at kitware.com] 
> Sent: Thursday, October 17, 2013 5:22 PM
> To: Becksfort, Jared
> Cc: Insight Users
> Subject: Re: [Insight-users] ITK, Random Forests, and OpenCV
>
> Hi Jared,
>
> There is a bridge [1] to/from ITK and OpenCV.
>
> I understand your frustration of articles without code!  If you get something going, please consider submitting it to the Insight Journal [2].
>
> Thanks,
> Matt
>
> [1] http://itk.org/ITKExamples/src/Video/BridgeOpenCV/ConvertAnITKGrayScaleImageToCVMat/Documentation.html?highlight=opencv
> [2] http://insight-journal.org/
>
> On Thu, Oct 17, 2013 at 9:19 PM, Becksfort, Jared <Jared.Becksfort at stjude.org> wrote:
>> Hello,
>>
>> I am trying to use random forests on medical images.  I am converting a Java Weka program I wrote to C++.  I want to use ITK for its medical imaging tools and file format support, but it doesn't seem like there is a random forest package in ITK.  Am I wrong?
>>
>> If there is not a random forest package, I had planned to use the machine learning component of OpenCV, but it does not support the file types I have to support, such as nii, dicom, and mhd/mha and isn't really designed for medical imaging.  It also seems like I am making it much more difficult than I need to.  I have seen papers published that used ITK with random forests, but the code is not available.
>>
>> Does anyone have any suggestions for using a random forest package with or within ITK?  Or using OpenCV for medical images (yes, I know I should ask that on the OpenCV mailing list)?
>>
>> Thanks,
>> Jared
>>
>> Email Disclaimer:  www.stjude.org/emaildisclaimer Consultation 
>> Disclaimer:  www.stjude.org/consultationdisclaimer
>> _____________________________________
>> 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.php
>>
>> 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
>
> _____________________________________
> 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.php
>
> 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
>



More information about the Insight-users mailing list