Summary:
The author proposes some code to perform morphological closing and opening on binary images.
This is a common operation in digital image processing, and particularly in Medical Imaging.

Hypothesis:
Not applicable.

Evidence:
The paper is clearly organised and illustrates the code with some examples.

Open Science:
This is clearly Open Source code. Data are available with the code.

Reproducibility:
I compiled the code seamlessly with Visual Studio 8 Express. I ran the code and reproduced the exact same results as shown in the paper. Great!
(I would suggest to add a USAGE line to the application code, so the user can run the executable and know about the input and output arguments.)

Use of Open Source Software:
100% Open Source Software: it is developed as ITK classes.

Open Source Contributions:
All the source code is provided. Usable straight from the shelf.

Code Quality:
Looks excellent ITK style, and platform-independent (the dashboard indicates that the code passes on all platforms - Jan 31, 2006).

Applicability to other problems:
Morphological operators have definitely various applications in scientific imaging.

Suggestions for future work:
Could the boundary be handled differently? I am afraid the padding and cropping would penalize the processing of very large images.

Requests for additional information from authors:
Could the author elaborate on the technical justification for the padding/cropping?
I believe the implementation of SafeBorder=true can treat the boundaries more efficiently for the closing.
(Padding and cropping should be turned off altogether when we select the opening.)

Additional Comments:
The paper layout is perfect. (You may just re-run pdflatex to set the Figure index on pages 2 and 3.)

Summary:
This paper provides binary opening and closing filters to ITK.

Hypothesis:
Not applicable

Open Science:
Source code and tests are provided.

Reproducibility:
I have compiled and run an earlier version of this code.

Use of Open Source Software:
The authors use existing ITK filters to implement these filters.

Open Source Contributions:
The authors provide the source code.

Code Quality:
I suggest renaming these filters to itkBinaryOpeningImageFilter nd itkBinaryClosingImageFilter. Having the word Morphological in the name of the class is extraneous.

Applicability to other problems:

Suggestions for future work:

Requests for additional information from authors:
The closing filter provides a SafeBordersOn() method which establishes a pipeline of a pad, dilate, erode, crop. The author illustrates that just using a dilate+erode pipeline produces an incorrect result on the boundary on the image. The pad, dilate, erode, crop method produces the correct image.

Are the pad and crop operations really necessary? If the dilate+erode is not producing the correct result by itself, then I would suspect that the erode operation is not using a proper boundary condition. Can the author provide an additional discussion on this? Perhaps an example simply using erode that illustrates the erode operation cannot produce the correct result regardless of the boundary condition. This needs a mathematical argument instead of something using the BinaryErodeImageFilter.

Additional Comments:
It is great that the author is taking the time to fill in the gaps in the morphology offerings of ITK.