|Please use this identifier to cite or link to this publication: http://hdl.handle.net/10380/3227|
This paper describes two new filters for raising an input image by a given power/exponent.Evidence:
The complete source code, tests, and baseline output images are provided.Open Science:
The complete source code, tests, input images, and baseline output images are provided.Reproducibility:
I downloaded the code, compiled it, and ran it. I had some trouble getting the automated tests to run (I'm not sure I correctly configured the ITK test driver program). In the end I ran the tests manually and received the expected output.Use of Open Source Software:
The paper proposes new filters for the ITK.Code Quality :
The files meet the ITK coding guidelines.Free comment :
I would much prefer that only one filter is added: itkPowImageFilter. This single filter should combine the functionality of the proposed itkPowImageFilter and itkConstantPowImageFilter. That is, the single itkPowImageFilter should support a power image, and if not provided fall back to using a constant. This is similar to say the itkFastMarchingImageFilter which expects a speed image, but if not provided uses a speed constant (default = 1.0).
Also, it would be nice to provide a SetPowImage() function to avoid the use of SetInput(1, powImage).
Comment by Gaetan Lehmann: implementation choices
Thanks a lot for your review.
The filters are implemented as subclasses of UnaryFunctorImageFilter and BinaryFunctorImageFilter. I'd like to keep it that way, because it makes the implementation very easy and gives access to in place processing at no cost. It would be quite difficult to implement the two filters in a single one because of that.
It would be a very interesting feature to implement in BinaryFunctorImageFilter though, and would avoid all the recent filters with Constant in their name - Add, Divide, Multiply and Subract filters.
I agree about the SetPowImage() - I'll add it soon.
|Download Paper , View Paper|
|Download Source code|
Linked Publications more
Diffeomorphic Demons Using ITK's Finite Difference Solver Hierarchy
by Vercauteren T., Pennec X., Perchant A., Ayache N.
Surface Meshes Incremental Decimation Framework
by Gelas A., Gouaillard A., Megason S.
Send a message to the author