Road map for Parsing Image data to PyTorch

05 Dec 2018

Problem

Since the dataset consists of Analyze 7.5 format files. I want to find if there is already the software package to deal with the dataset.

Trial

1. Nibabel

First, I find the Nibabel

• Advantage: This package provides read +/- write access to some common medical and neuroimaging file formats, including: ANALYZE (plain, SPM99, SPM2 and later), GIFTI, NIfTI1, NIfTI2, MINC1, MINC2, MGH and ECAT as well as Philips PAR/REC.
• Disadvantage: it’s too complicated without a clear representation for the plotting.

2. Nilearn

• Advantage: Diverse plotting method.
• Suitable for dealing with Analyze 7.5 format.

3. Pydicom

• A software package for dealing with DICOM files

4. Nipype

Nipype consists of many parts, but the main ones are Interfaces, the Workflow Engine and the Execution Plugins. example

Find BIDS: a simple and intuitive way to organize and describe neuroimaging and behavioral data.

5. MedicalTorch

MedicalTorch is an open-source framework for PyTorch, implementing an extensive set of loaders, pre-processors and datasets for medical imaging. link

Conclusion

In the first step, Nibabel and Nilearn should be used in our project. And the MedicalTorch will be useful in PytTorch.