# Calculating variability using vmtksurfacedistance

**URL:** <https://discourse.slicer.org/t/calculating-variability-using-vmtksurfacedistance/24101>\
**Category:** VMTK\
**Created:** [June 29, 2022, 12:52pm UTC](https://discourse.slicer.org/t/calculating-variability-using-vmtksurfacedistance/24101 "2022-06-29T12:52:39Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![fsneden](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/fsneden/32/13546_2.png) [@fsneden](https://discourse.slicer.org/u/fsneden)\
**Post date:** [June 29, 2022, 12:52pm UTC](https://discourse.slicer.org/t/calculating-variability-using-vmtksurfacedistance/24101/1 "2022-06-29T12:52:39Z")

</div>

I am trying to calculate intra-observer variability in a dataset. I am creating surface data from images, and using the vmtksurfacedistance script to calculate difference from an input image (my 2nd surface) and a reference image (my 1st surface). This should give a distancearray value that will describe difference between the two.

Assuming this is the correct procedure, I am then trying to translate this to numpy in order to get the distancearray value. However, this doesn’t seem to work. My code (minus the specific paths and in 2 parts) is the following:

```auto
## format image for analysis
vmtkimageviewer -ifile "raw_image.ima" --pipe vmtkimagewriter -ofile "formatted_image.vti"

## segment formatted image
vmtklevelsetsegmentation -ifile "formatted_image.vti" -ofile "segmented_image.vti"

## create surface of segmented image
vmtkmarchingcubes -ifile "segmented_image.vti" -ofile "new_surface.vtp"

## calculate distance from new surface and old surface
vmtksurfacedistance -ifile "new_surface.vtp" -rfile "old_surface.vtp" -ofile "surface_distance.vtp"

```

Extraction of numerical values is performed in Jupyter notebook (see below).

```auto
## import packages
from vmtk import vmtkscripts 
import numpy as np

## read surface
surfaceReader = vmtkscripts.vmtkSurfaceReader()
surfaceReader.InputFileName = 'surface_distance.vtp'
surfaceReader.Execute()

## translate surface to numpy
surfaceNumpyAdaptor = vmtkscripts.vmtkSurfaceToNumpy()
surfaceNumpyAdaptor.Surface = surfaceReader.Surface
surfaceNumpyAdaptor.Execute()

## get dictionary output
numpySurface = surfaceNumpyAdaptor.ArrayDict
numpySurface

```