design - копия

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 1

In this example we use SimpleITK's VectorConfidence Connected filter (Ibáñez et al.

, 2005) to segment
an MRI data set from the Neuroimage Analysis Center's Multi-modality MRI-based Atlas of the Brain
(Halle et al., 2013). This filter begins with user-selected seed points then iteratively grows a region based
on the similarity of the local vector image statistics. To create a vector image we combine the MRI T1
and T2 images into a single vector image.

We developed this script interactively in an IPython Notebook (Pérez and Granger, 2007) session. We
implemented custom display functions called showimg and showimg3d, which allow for the easy tiled
display of multiple slices. For display we export the tiled slices to numpy and utilize matplotlib for inline
display in the IPython notebook.

First we read the T1 and T2 brain images and create unsigned char version for display (see Listing 3: lines
5–10). Taking advantage of SimpleITK's Python array slicing, extracting axial slices from the MRI volumes
is simple (see Figure 2 and Listing 3: lines 13–14). Then from the original T1 and T2 volumes we create a
vector volume using SimpleITK's Compose function (see Listing 3: line 18).

You might also like