read all the dicom files in a directory?

[qs5d]

Does DCMTK provide some method to read all the dicom files under the same directory and write into one file? Thank you.

[Jörg Riesmeier]

This is not supported by the DCMTK since you cannot write multiple SOP instances into one. Or do you mean a DICOMDIR file? This can be created using the tools dcmgpdir/dcmmkdir or the class DicomDirInterface.

[qs5d]

Maybe I didn't express my idea clearly in the previous question. Now I have several hundred of dicom files under a directory, those files are randomly named. I need to read those randomly named dicom files into one file with only pixel data in order to create a 3d volume of raw image. This 3d volume are saved in an array that reads in the pixel data of each dicom file slice by slice. I'm not sure whether it could be done with DCMTK? Thank you.

[wrgben]

You could use dcmdump with the +W option. This will dump the pixel data into a .raw file. Then you could concatenate all the .raw files together to make your volume dataset.

[Jörg Riesmeier]

Alternatively, you could write a little program that performs this task (load DICOM file, extract pixel data, append to output file). I guess that "qs5d" would also like to sort the individual images before creating the image volume.

Btw, I could also imagine that some of the 3D toolkits for the visualization of medical images provide such functionality.

[Max]

Alternatively, you could write a little program that performs this task (load DICOM file, extract pixel data, append to output file).

Hi,
This is exactly what I am trying to do, I started from the example in the DCMTK documentation and have written the following code:

Code: Select all

FILE *stream;
int main()
{
int list[65536];	//image is 256*256
DicomImage *image =new DicomImage("test.dcm");
if (image != NULL)
{
  if (image->getStatus() == EIS_Normal)
  {
    if (image->isMonochrome())
    {
      image->setMinMaxWindow();
      Uint8 *pixelData = (Uint8 *)(image->getOutputData(8 /* bits */));
      if (pixelData != NULL)
      { /* do something useful with the pixel data */ 
	     fopen_s(&stream, "test.raw", "wb" );
	     for (int i = 0; i < 65536; i++)
	     { list[i]=pixelData[i];
	        fwrite( list, sizeof( int ), 65536, stream );
              }
       }fclose(stream);
    }
  } else
    cerr << "Error: cannot load DICOM image (" << DicomImage::getString(image->getStatus()) << ")" << endl;
}
delete image;
}

I get a test.raw file which is huge compared to the correct one I get using dcmdump.exe with the +W option.
I'm sorry if the answer sounds easy to you, but I'm quite new to this kind of things.
Thank you.

[Jörg Riesmeier]

Why are you using a for-loop? Just write the whole array of pixel data into the file using fwrite().

Btw, DicomImage::getOutputData() does not give you access to the raw pixel data (in contrast to "dcmdump +W") as you can read both in the documentation and in various postings in this forum.

[Max]

Thank you,
I've modified my code, now it uses the findAndGetUint16Array() function and seems to work fine.
Sorry for the dumb question.