Respect pixel padding value during interpolation

[alund]

• Fixes Gantry tilt interpolation should respect pixel padding value #262

[neurolabusc]

This looks clear. Can I make one request: instead of four variables
isHasPixelPaddingValue
isHasFloatPixelPaddingValue
pixelPaddingValue
floatPixelPaddingValue
Could we have just
floatPixelPaddingValue
that is set to NAN by default? The structure TDICOMdata is required for each and every DICOM, so keeping these smaller is better. For the case of 16-bit data you can just use the specify pixelPaddingValue = round(floatPixelPaddingValue) in nii_saveNII3Dtilt(). And isHasPixelPaddingValue can be set in nii_saveNII3Dtilt() and nii_saveNII3DtiltFloat32() based on whether floatPixelPaddingValue == NAN.

[alund]

I have pushed a new commit that makes this change. It also fixes a mistake I made: the data type for Float Pixel Padding Value is FL and so should have been read with dcmFloat() not dcmStrFloat(). I also realized that, if a pixel padding value is defined, it should be used for the padding done in nii_saveNII3Dtilt() and nii_saveNII3DtiltFloat32().

Having done all of that, I would like to do some more testing before you merge this. Feel free to review, but even if you have no more comments, I would appreciate it if you wait before merging until I have confirmed that my additional testing is complete.

[neurolabusc]

Looks good to me. I will hold off until I get your OK.

[alund]

The tests I have run so far have been fine. I think you can go ahead and merge. Obviously if I find anything later, I'll fix it and submit another pull request.