Abstract
The itk::GaussianDerivativeImageFunction computes the derivative at the specified physical or pixel location. Unfortunately it has a number of deficiencies, for which I suggest possible solutions.
Keywords
Source Code and Data
Reviews
Julien Jomier
Friday 28 April 2006
Summary: This paper proposes an improvement to the GaussianDerivativeImageFunction while respecting backward compatibility.
Reproducibility: I was able to compile the code but only after modifing the main CMakeLists.txt. Note that since the source code is templated there is no need to create a library for it. CMake was confused about this. The following lines:
#Add Source #SUBDIRS(Source)
should be removed from the main CMakeLists.txt This is why the code is not showing up on the Insight Journal website.
Code Quality: The code looks good and respects ITK style for the most part.
Requests for additional information from authors: a) I could not find the RecomputeContinuousGaussianKernel() function in the current ITK cvs class. b) The plan is to test that the current ITK cvs is broken. The best way is to add the current test Insight/Testing/Code/Common/itkGaussianDerivativeImageFunctionTest.cxx in your publication and add a test for spaced images at the end of this test. It would be great if you can submit a new revision with this modification. c) Why is this code not working in 3D?
Other comments This is a very good work. Thanks for the contribution!