ETHZ: 227-0966-00L

Guest Lecture

Exercises

X-Ray Tomography of Porous, Transition Metal Oxide Based Lithium Ion Battery Electrodes

• Matlab for Image Processing and Analysis
• Image Processing Toolbox

1. Read the supplemental materials for a paper
2. Download the publically available data
3. Reproduce the results

• Original Paper
• Supplementary Material
• Data
• First Dataset

In Matlab run the following code Copy / Paste from Here

cd('~')
unzip('http://jenkins.imagej.net/job/Stable-Fiji/lastSuccessfulBuild/artifact/fiji-nojre.zip','./')

addpath 'Fiji.app/scripts'
Miji % start Miji
cd('~')
% to test if it works
% Miji_Test

1. Download the first binary dataset (we won't worry about thresholding now)
2. Decompress the data into the scratch drive

cd /scratch/
tar -xvf ~/Downloads/binarized.tar

• Basic Loading and Processing Script
• 3d Shape Analysis

1. Compare the watershed and component labeling, what is the difference?
2. How would you implement the additional steps in the supplementary material
   • Join by surface fraction
   • Join by proximity
3. How do these steps change the final results?
4. How does changing the region of interest and scale affect the results?
   • For these data which approach is better

Once the data is loaded you can import the segmentation or distance map into Fiji by using MIJI. Since we have already started MIJI earlier we only need to create an image from out Matlab array. We format it as an unsigned 16-bit integer since it is a label (distance map we would then format as a double)

%% Load Slices into Fiji
MIJ.createImage('Watershed',uint16(cWS),true);

You will need to adjust the brightness and contrast to see the properly labeled images