Summary:
The article gives a short presentation of the watershed method and goes on
to present a new, related method called marked based watershed. The
authors present the relevant algorithms and then compare the performance and
accuracy of the existing ITK watershed method with the new algorithm using a variety of examples.

Evidence:
The authors are very thorough in presenting their evidence. They do not make
bold claims, but rather cite the differences based on reproducable evidence.
The relevant algorithms are presented in the article, as well as references
for the sources. All the claims made in the article can be verified as
everything necessary to duplicate the results are provided.

Open Science:
All the information used in creating the article are made available. The parameters
of the examples, along with the data files and source code are provided, so
the user can easily do an independent verification of the results presented.

Reproducibility:
I did not verify their results, but have made extensive use of the code in my own project.
The work is available as an external project for WrapITK, and is therefore
very easy to build and utilize. It is very easy to reproduce the work when needed.

Use of Open Source Software:
The authors used pieces of technology developed by them and others that are
part of (or candidates to be part of) the ITK. Otherwise, the outside sources
seem to be literary and not in the form of open source software.

Open Source Contributions:
The code related to the article is an model example of how such code should
be presented. The authors make the source code, as well as the example code and data
available. The code is easy to make use of, as it comes as an external
project for WrapITK, only requiring the familiar steps of CMake, Make, and
Make install to get running.

Code Quality:
I have not reviewed the code fully, but the code that I looked at is very
clear, with consistent coding style and relevant comments where they are
needed. There seems to be nothing that would limit the code to a specific
platform.

Applicability to other problems:
As the authors suggest, watershed method is a very flexible method that is
useful in a wide variety of situations. The authors show many examples
that show the differences in the new marker based watershed and the ITK
watershed method, where often using the new method is required to get the correct
segmentation results.

This better applicability to a set of problems as well as the better
performance make the method appealing for many projects that need to utilize watershed
method. As an example, the code is used extensively in the project I'm working
on, where it is used to segmentate confocal microscopy images.

Requests for additional information from authors:
The article is very well presented and as all the relevant material is
available, leaving no open questions at this time.

Additional Comments
The paper follows the requirements of open science very well, provides an open source implementation of a method that is
directly useful for many users (including myself) and the backgrounds of the article are presented well. I certainly recommend anyone
using the watershed method to try the new method, and would like to see it as a part of ITK.