The submission provides an efficient implementation of the Stochastik Rank Correlation Metric. The implementation offers a variety of options in order to "adapt" the metric to specific needs. 

The paper is very well written and gives detailed information on Stochastik Rank Correlation, the ITK implementation of the metric. Moreover it provides very comprehensive examples on how to use the metric for 2D/3D registration. Based on these examples one can immediately start to apply this useful contribution on the test data, but also on own datasets.   

The authors provide comprehensive test data+test example for 2 different testing scenarios in combination with a very detailed description on the used configruation settings. Moreover, detailed evaluation of the metric based on the test scenarios was performed.

The authors proved source code + all test data that was used for the test examples in the paper. 

I donwloaded the code and repoduced some test examples without any problems. Expecially the IntensityBased2D3DRegistration2 example in combination with the configuration file provides a complete registration framework that made it very easy to start some test runs using my own data. 

The whole code package is highly useful. The description on how to use the code as well as the documentation of the sources is exemplary.

The code quality is excellent, conforms with ITK coding style and is perfectly documented. 

The quality of the data is good and very weel suited for testing the code.

I did not only use this metric for 2D/3D registration, but also for statistical model-to-CT registration. The integration of the classes in my exisiting framework worked without any problems and I was able to obtain very good registration results using SRC.

I would strongly recommend to  add this contribution to ITK as soon as possible. SRC is a very useful metric especially (but not only!) for 2D/3D registration scenarios and the code quality is exemplary.