Examples

We include two worked examples in our github repository under the examples folder. In both cases we include an interactive approach that inclues several visualizations, in the form of a jupyter notebook. At the end of each jupyter notebook we show how the same analysis can be run from the command line, using our transformation graph interface, producing our standard output format.

The first describes 3D registration between two human MRI datasets. The second describes 3D to 2D registration between the Allen Nissl atlas, and a sequence of Nissl stained images.

Note that our command line interface has only been validated on Linux systems.

Contents:

• Human MRI example
  • Import libraries
  • Outputs
  • Load images
  • apply the transform to the atlas image
  • apply the transform to the target image
  • Run the same analysis using the command line interface
  • write out a config file for registration
  • write out a config file for the transformation graph
  • Run this example from the command line
  • View all the outputs
• Mouse Serial Section Example
  • Import libraries
  • Load images
  • Perform an initial “slice to neighbor” alignment
  • Run atlas to slice alignment
  • Apply the transform from atlas to registered space
  • Apply the transform from target to registered space
  • Apply the transform from registered target to atlas space
  • Run this example from the command line
  • Write out the registration config file
  • Write out the transformation graph config file
  • Run the command
  • View all the outputs
• Example rat brain image registration
• Import libraries
• Load the data
• Set up configuration for low resolution registration, and run it
• Run it again using a warm start and local contrast
• Make some visualizations of the result to verify accuracy
• save the registration data