Summary:
This papers describes an image analysis framework inspired in the concepts of artificial life. The framework has been ported to ITK and its source code is submitted here in a form that is suitable for processing its inclusion in ITK. The capabilities of the framework are demonstrated in several image segmentation scenarios.

Evidence:
The authors provide the source code of their framework along with specific examples that make possible to reproduce the segmentation scenarios presented in the paper.

Open Science:
The paper is a perfect example of Open Science. Not only the author provide the source code, image and parameters required for using their method, they also go throught great care in their paper for describing users how to take advantage of this new framework. The details are not only sufficient for replicating their work, they also provide guidance on how to further extend the framework.


Reproducibility:
I was able to replicate several of the examples described by the authors.
The source code was available and compiled without problems in my Debian Linux with gcc 3.4 and a CVS version of ITK (August 15 2006). The examples ran without problem. The authors configured the CMakeLists.txt file in a very convenient way. Thanks to their organized submission, it was trivial for me to replicate their example of cube segmentation and XXX segmentation.

Use of Open Source Software:
The authors used several Open Source software packages in their work. They mainly used the Insight Toolkit (ITK), and integrated their framework very nicely using existing ITK classes, and creating new classes for the new concepts that their framework brings. The authors give a balanced description of the capabilities of their framework. In their examples, they also used VTK and SOViewers for visualization, and KWWidgets for building a GUI interface.

Open Source Contributions:
The authors provide their source code in a form that is very easy to use. They describe the code details in their paper, an additional PDF manual, and provide a web site with Doxygen documentation for their classes. I was able to use the code in about 30 minutes.

Code Quality:
The code is of excellent quality. It is easy to read and follows most of the ITK coding style. I only tried the code on Linux, but didn't find a reason why the code couldn't run in other platforms. The authors provided binaries for Windows, so it is clear that the framework builds at least in Linux and Windows.

Applicability to other problems:
The Deformable Organisms framework is an overdue contribution to ITK. It brings very powerful capabilities for introducing heuristicst in the process of image analysis. The authors provided examples on the use of this framework in the context of image segmentation and feature extraction, but it is clear that it can also be applied to image registration.

Suggestions for future work:
The framework provided by the authors is an excellent foundation for many applications. Future work can be developed in many directions, mainly by looking at specific applications. Two interesting conceptual additions come to mind: Evolutionary Algorithms, and Learning. In the front of evolutionary algorithms it will be a natural extension to provide the Deformable Organisms with artificial genomes that could be suitable to evolve under the control of an Evolutionary Algorithm. In this way, image analysis problems could be solved by applying "Directed Evolution" methodologies that are commonly used today in Biology and applied fields such as nanosciences. On the learning front, it will be interesting to use the Neural Network framework in order to provide trainable brains for the Deformable Organisms.


Requests for additional information from authors:
All the information needed for using the framework was available in the authors submission. Additional information was provided in their web site.

Additional Comments:
This is one of the best papers I have seen in the Insight Journal. I was aware of the concept of Deformable Organisms from the publications of the authors in 2001. It is of great interest for ITK to bring this framework into the toolkit. It will give a major benefit to the users community.

Summary:
A novel extensive, generic framework for deformable organism segmentation that fits neatly into ITK.

Hypothesis:
The authors present a new framework appropriate for deformable organism based segmentation.

Evidence:
The authors present only a few examples, but no real set of data with parameters etc, in order to reproduce segmentation results. But the example are quite illustrating and it seems quite straightfoward to use the framework and extend it.

Open Science:
The contribution to open science is mainly the novel framework. More cases and examples would have been great and would have allowed also to compare the methodology to totally different segmentation algorithms. There are otherwise no major open source implementations for deformable organisms and thus this fills a void.
I could though not find the source code of the BYU conversion filter that is mentioned in the manuscript, as well as the sources for the KWWidgets Viewer.

Reproducibility:
I did not reproduce the results, but the code has tests that resolved successfully on the submission system

Use of Open Source Software:
Heavy use of open source using ITK and VTK.

Open Source Contributions:
The authors contribute an extensive framework for deformable organism based segmentation.

Code Quality:
I have not looked in detail at the code, but I inspected a few classes and the quality seems to be high. It fits very nicely into ITK and follows all the standard ITK guidelines. The class hierarchy and the integration of the different methodologies is performed in exemplary fashion.

Applicability to other problems:
As this is a generic framework, it seems quite straightforward to extend the current base to new applications of deformable organism based segmentation.

Requests for additional information from authors:
It would have been great to have more detail about the methodology in the paper itself, rather than referring to the journal publications alone.