The Neurodata's Registration module or ndreg is a set of python functions which can be used to align pairs of images. The module depends on
See README for platform specific installation instructions
Image volumes from a modern microscope can be over a terabyte in size, far too large to be stored on a personal computer. Thus in our infrastructure are stored on server cluster known as ndstore. Accessing image data of this size poses many challenges. Thus ndstore organizes images in spatial database in such a way to minimize read and write times during random access of image cutouts. In ndstore each image dataset is encapsulated in a project which can consist of one or more channels. Each project has a token which is used to access the image data. A more detailed handling of ndstore's data model can be found here. Given a project token and a channel, images can be downloaded directly from the web to a personal computer using ndstore's RESTful API. In python, the API can be called using the NeuroData's Input/Output, ndio module. ndio functions save image data in numpy arrays for easy manipulation.
Although numpy is a powerful numerics library it's image processing capibilites are limited. Thus ndreg depends on SimpleITK for image processing. A detailed tutorial on the installation and usage of SimpleITK can be found here. SimpleITK images can be converted to and from numpy arrays using the SimpleITK.GetArrayFromImage and SimpleITK.GetImageFromArray
In the following tutorials we use the following naming conventions for modules
In [1]:
import numpy as np
import SimpleITK as sitk
import matplotlib.pyplot as plt
In [ ]: