ROI Annotation (regrender roi)

Label cells/ROIs on raw histology, then project them into Allen CCF space and render with brainrender. The workflow order is roi → register → project: ROIs are marked on the full-resolution raw image (in raw pixel coordinates, no registration needed yet), and the saved registration transform is applied later to project them into the atlas.

Labeling cells per channel (right), with projected ROIs rendered in 3D via brainrender (left).

Labeling

regrender roi -D <slices_dir>

In the napari GUI:

• Step through serial sections with Prev/Next.

• Choose the channel to label (merge / R / G / B) for multi-channel images.

• Click cells to add ROIs; they are saved to roi/roi_points_raw.csv in raw pixel coords.

• Verify warps the slice into atlas space (once registered) to show where ROIs land.

Projecting & rendering

After the slices are registered with regrender register, project the raw ROIs into CCF space and render:

• In the GUI, click Project + Render, or

• Run headless:

  regrender roi --project --render -D <slices_dir>
  

Projection writes roi/roi_points_ccf.csv (adds AP/DV/ML_location mm + region acronym), then shells out to neuralib.atlas.brainrender.roi to render one color per channel.

Options

Option	Meaning
-D, --directory	Folder of serial sections (reads transformations/<stem>_transform.json).
-I, --image	Single image (alternative to -D).
--transform-dir	Where the *_transform.json live (default <dir>/transformations).
-O, --output	Output CSV path (default <dir>/roi/roi_points_raw.csv).
--project	Headless: project a saved raw-ROI CSV into CCF space (no GUI).
--render	With --project, also launch brainrender.
--roi-radius	Rendered ROI sphere radius in µm (default 30).
--roi-color	Rendered ROI point color (default orange).