Week 45#


Hosted on Sketchviz

DICOM set file#

DICOM data set can include and represent different DICOM objects, included or supported by the standard; images, text reports, audios. The input and output volumes are files .NRRD.

Output data#

According to the process followed, the volume straightened is a file .NRRD, a method for storing and representation of n-dimension raster data. If we analyze the header of one volume obtained, its characterized by:

  • Nrrd file type: one file.

  • Dimension: 3.

  • Size: 99x99x70 (ZxWidthxHeight).

  • Type: int (acording with the NRRD documentation, its means: 4-byte integer).

  • Encoding: gzip.

Working with the file .NRRD its possible to: - Open it with python, using the pynrrd package. - Convert it to different formats like, jpeg and variation, tiff, etc. - Save the file like a DICOM data set with other studies. - Anonymize patient info working only with .nrrd files.

Just curiosity#

For the next steps:

  • There is a pre-processing of the input data needed?

  • There is a data set to train the model?

  • There is a model of prediction?