How to Create ROIs for Autosegmentation

This guide walks through creating regions of interest (ROIs) to use with the autosegmentation tool in PyReconstruct.

Steps

Create a stamp the size of a dendrite of interest.
Place the stamp in the center of the dendrite on every section.
Copy and rename the stamp to RO# to define a region of interest surrounding that dendrite.
Set the ROI radius. Use the Object List to reset the radius of the RO# stamp to 4 µm. This makes all stamps that radius. (8 µm is the largest manageable radius.)
Align the series. Press Ctrl + \ with the field zoomed to the size of the RO# hex.
Refine alignment. Page through the entire series to achieve a precise alignment around that RO#.
Handle folds and artifacts. If a fold or artifact occurs near the dendrite during alignment, set the contrast to 100 so the image becomes white. The autosegmentation tool will then interpolate traces from the surrounding sections and assign them the correct names.
Run autosegmentation. Use the autosegmentation tool to generate traces of all objects within each region of interest.
Use unique JSERs for differing contrasts. If your ROIs have different contrast levels, you must use a separate JSER file for each ROI.
Proofread. Press X to enter proofread mode, then reassign each auto_### label to the correct object name.
Clean up. Delete any off-target segmentations, or bring in additional ROIs using multiple alignments.