- If necessary - stitch/fuse in Zen
- Alternatively select just one side-view
- Open full resolution channels individually and save as TIFs
- can optionally be converted to 8-bit to save computational resources
- Segment each channel individually (see segmentation instructions below)
- Combine all three met channels to create one "All met" mask
- Segment vessel image manually if possible, alternatively attempt via ilastik
- On the "all met" mask image - run:
- MorpholibJ Plugin
- Binary
- Connected Components Labelling
- Analyse
- Analyse 3d
- Select Volume and Bounding Box
- Analyse 3d
- Binary
- Save Results table as csv
- MorpholibJ Plugin
- Make sure pixel values are correct (shouldn't need changing but might)
- Specify a directory to save met images to (line 6)
- Specify location of bounding box csv (line 8)
- Set volume cut-off (line 18)
- Run
- Set and forget, can take a long time depending on number of mets
- Make sure you have "Neurocyto LUTs" plugin site installed
- Specify dir1 (directory with the met images from the previous step)
- Specify dir2 (directory to save binary lego masks to)
- If you want to use manual thresholds instead of allowing the macro to attempt an autothreshold, then:
- change USE_MANUAL_THRESHOLD to true
- set the c1,c2 and c3 thresholds (c1_th, c2_th, and c3_th)
- Has proper menu
- Specify directory with mets from above
- Specify which met ID to start from
- Specify volume / bounding box csv from above
- Specify the segmented, local thickness vessel image
- Specify pixel and volume size of the thickness image (may be different if vessels were segmented on the scaled down image which may be easier)
- Specify "search radius" (line 44)
Finally we arrive at a spreadsheet with mets scored by volume of each lego channel, plus distance to nearest vessel
All the data analysis from the extracted image data is in the folder data_analysis