The authors provide additionnal ITK classes for reading/writing Dcm/Tiff from MEVIS

The authors provide additionnal ITK classes for reading/writing Dcm/Tiff from MEVIS

The authors provides a test file wich I've direclty used with some MEVIS data I had.

The authors provide the source code + a test file.

The authors did not provide dataset to test the code

The authors nicely describe how to use its code. However, It took me some time before being able to compile the code.

I am under Windows7 64bit, VS 2008

I've used:

gdcm 2.0.10 (static)
ITK 3.20 (shared)
tiff-3.8.2-1

WARNING some version of gdcm does not compile or does not compile together with ITK.

gdcm in shared library did not compile

1. To compile ITK, I had to manually edit the properties of the ITK project : itkTestDriver

add:

a. Additionnal Depedencies:

gdcmCommon.lib
gdcmDICT.lib
gdcmDSED.lib
gdcmexpat.lib
gdcmgetopt.lib
gdcmIOD.lib
gdcmjpeg12.lib
gdcmjpeg16.lib
gdcmjpeg8.lib
gdcmMSFF.lib
gdcmopenjpeg.lib
gdcmzlib.lib
libtiff.lib

b. Additionnal Library Directories:

add library directory of TIFF library

c. Add an include file into MevisDicomTiffImageIO.h:

#include "gdcmByteSwapFilter.h"

Maybe not the perfect one but I took one where gdcm::DataSet was used if you do not include it leads to the following error:

itkMevisDicomTiffImageIO.h(152) : error C2039: 'DataSet' : is not a member of 'gdcm'

2. To compile the itkMevisTest.cxx file:

I had again to manually edit the VS project the same way as presented above.

As I've had issues with gdcm library, it tooks me two days to make it compile

I did not extensively review the code, but it seems to be clean C++ of high quality.

This tool is usefull when you are dealing with MEVIS data and should be incorporated into ITK library if possible (issue with gdcm compilation).

In almost every example in ITK there are very confusing uses of the command line arguments. For example, in CannyEdgeDetectionImageFilter.cxx there is a section:

  if( argc > 3 )

    {

    variance = atof( argv[3] );

    }

  if( argc > 4 )

    {

    upperThreshold = atof( argv[4] );

    }

If the arguments are changed, every index must be manually updated. This also leads to very easy logic errors. With this proposed parser, the above section can be changed to simply:

float variance = 1.0;

parser->GetCommandLineArgument("-variance", variance);

float upperThreshold = 2.0;

parser->GetCommandLineArgument("-upperThreshold", upperThreshold);

and the input can be changed from 

CannyEdgeDetectionImageFilter brain.png 2.3 4.5

to a much more clear:

CannyEdgeDetectionImageFilter -input brain.png -variance 2.3 -upperThreshold 4.5

This will certainly improve the clarity of the examples enormously!

Reasonable tests are provided, and the documentation is very clear.

The code compiles and runs correctly on my system.

A brief discussion of why to use this instead of a large library for command line parsing is provided and convincing.

The code is very straight forward and easy to use.

I would like to see this code added to the ITK toolkit.