scitran/hdft-subsampled-recon repository overview

⁠hdft-subsampled-recon

Build context for a Flywheel compatible Gear⁠ for the HDFT Subsampled Recon algorithm from Schneider Lab, University of Pittsburgh.

This sample code computes a transformation of multi-shell diffusion weighted data to a set of Spherical Harmonic coefficients. Outputs 4D Spherical Harmonic coefficient data. This is a first step in the Schneider Lab HDFT diffusion reconstruction process.

⁠Inputs

• dwi_filename: input filename of 4D DWI data
• subsampling_vec: input vector to select volumes from dwi_filename (can be a text file or CSV)
• bvals_filename: input file of Nx1 b values
• bvecs_filename: input file of Nx3 b vectors

⁠Parameters

• spherical_harmonics_order: The maximum order of spherical harmonics. Defaulted to 8.
• mean_diffusion_length: The mean diffusion length for reconstruction of GQI matrix. Defaulted to 1.2.

⁠Output

• sh_filename: 4D Spherical Harmonic coefficient data.

⁠Reference

Pathak, S. K., Fissell, C., Krishnaswamy, D., Aggarwal, S., Hachey, R., Schneider, W. (2015). Diffusion reconstruction by combining spherical harmonics and generalized q-sampling imaging. ISMRM, Toronto, Canada.

⁠License

All code is copyright University of Pittsburgh unless alternate authorship noted. This code is not for public distribution, please contact Schneider Lab re distribution. http://www.lrdc.pitt.edu/schneiderlab/⁠