Skip to content

Conversation

giltis
Copy link

@giltis giltis commented Apr 16, 2015

This is the running github issue for the image segmentation example for micro-ct glass bead packed column. This example demonstrates how to extract bead volume, pore space and calculated porosity of the packed column.

There is an ipython notebook and a VisTrails example. Both of these are included with this PR and will be updated as new functionality is added in PRs against scikit-xray

IPython notebook link: http://nbviewer.ipython.org/gist/anonymous/529b56087698dd0f9569
(Note that this is an old segmentation workflow and will be updated as the tools get implemented in scikit-xray.)
(Also note that this notebook needs to be edited to work again...)

This example will demonstrate the following functionality:

Step 1

Step 2

  • Image alignment
  • Plot aligned images (Using matplotlib)
  • Image alignment (above the absorption edge and below the absorption were collected at different times)

Step 3

  • Image Filtering (make sure you are plotting what is going to be thresholded)
  • Compute histogram
  • Plot histogram (Using matplotlib)

Step 4

  • Threshold by providing a specific value
  • Automatic thresholding
    • Otsu
    • Yen
    • Adaptive
    • Iterative

Step 5

  • Isolate pore space

Step 6

  • Generate mask that represents the walls of the packed column
  • Plot!

Step 7

  • Create a label field that distinguishes the beads from the container and the empty volume
  • Plot to verify that the label field is correct
  • Quantify the volume and porosity

Step 8

  • Render the label field in 3-D

Gabriel Iltis added 2 commits July 17, 2015 00:01
This is an initial commit for the end-to-end segmentation
demonstration and the transformation cropping demonstration. Both demos
are currently incomplete.
Gabriel Iltis and others added 17 commits July 17, 2015 00:14
These examples need to be modified followed by transfer to the vis_3D fldr, but are currently separated until the mods are complete and tested.
The data folder was created to hold the various types of tomo data that is being used in the example notebooks. This way there is a single data src instead of populating each folder with copies of the source data for various examples. Basically, in an attempt to save space...

This will hopefully also solve the issue with .yaml file metadata dictionaries showing as options for inclusion in version control (GIT)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant