linumpy.imaging.visualization#

Volume visualization utilities.

Consolidated from linum_screenshot_omezarr.py and linum_screenshot_omezarr_annotated.py.

Functions#

save_orthogonal_views(image, out_path[, z_slice, ...])

Save orthogonal (XY, XZ, YZ) views of a volume as a figure.

estimate_n_slices_from_zarr(zarr_path)

Try to estimate number of input slices from OME-Zarr metadata.

add_z_slice_labels(ax, n_input_slices, img_height[, ...])

Add Z-slice index labels on the side of a coronal/sagittal view.

save_annotated_views(image, out_path[, ...])

Save anatomically-labelled orthogonal views with Z-slice index annotations.

Module Contents#

linumpy.imaging.visualization.save_orthogonal_views(image, out_path, z_slice=None, x_slice=None, y_slice=None, cmap='magma', percentile_max=99.9)[source]#

Save orthogonal (XY, XZ, YZ) views of a volume as a figure.

Parameters:

  • image (array-like) – 3D volume (Z, Y, X) - as returned by read_omezarr.

  • out_path (str) – Output figure path (e.g. ‘view.png’).

  • z_slice (int or None) – Slice indices. Default: center of each axis.

  • x_slice (int or None) – Slice indices. Default: center of each axis.

  • y_slice (int or None) – Slice indices. Default: center of each axis.

  • cmap (str) – Colormap (default ‘magma’).

  • percentile_max (float) – Values above this percentile are clipped for display.

Return type:

None

linumpy.imaging.visualization.estimate_n_slices_from_zarr(zarr_path)[source]#

Try to estimate number of input slices from OME-Zarr metadata.

Checks custom metadata fields, multiscales metadata, sibling slice files in the directory, and falls back to a heuristic estimate.

Parameters:

zarr_path (str or Path) – Path to the OME-Zarr file.

Returns:

Estimated number of input slices, or None if undeterminable.

Return type:

int or None

linumpy.imaging.visualization.add_z_slice_labels(ax, n_input_slices, img_height, font_size=7, label_every=1, show_lines=False, side='left', slice_ids=None)[source]#

Add Z-slice index labels on the side of a coronal/sagittal view.

Parameters:

  • ax (matplotlib axis) – The axis to annotate.

  • n_input_slices (int) – Number of input slices stacked (e.g. 64 physical slices).

  • img_height (int) – Height of the displayed image in pixels (Z dimension).

  • font_size (int) – Font size for labels.

  • label_every (int) – Label every Nth slice.

  • show_lines (bool) – Draw horizontal lines at slice boundaries.

  • side (str) – ‘left’ or ‘right’ for label placement.

  • slice_ids (list of str or None) – Actual slice IDs (e.g. [‘05’, ‘12’]). If None, uses sequential numbers.

Return type:

None

linumpy.imaging.visualization.save_annotated_views(image, out_path, n_input_slices=None, x_slice=None, y_slice=None, font_size=7, label_every=1, show_lines=False, slice_ids=None, zarr_path=None, orientation=None, voxel_size=None, crop_to_tissue=False)[source]#

Save anatomically-labelled orthogonal views with Z-slice index annotations.

Parameters:

  • image (array-like) – 3D volume (Z, Y, X).

  • out_path (str) – Output figure path.

  • n_input_slices (int or None) – Number of input slices. Auto-detected if zarr_path provided.

  • x_slice (int or None) – Slice indices. Default: center.

  • y_slice (int or None) – Slice indices. Default: center.

  • font_size (int) – Font size for slice labels.

  • label_every (int) – Label every Nth slice.

  • show_lines (bool) – Draw horizontal lines at slice boundaries.

  • slice_ids (list of str or None) – Actual slice IDs to display.

  • zarr_path (str or None) – If provided, try to auto-detect n_input_slices from metadata.

  • orientation (str or None) – 3-letter RAS orientation code (e.g. 'RIA'). When provided, panel titles and axis labels use anatomical names (Axial/Coronal/Sagittal) derived from this code instead of the generic 'Coronal (ZY)' / 'Sagittal (ZX)' defaults.

  • voxel_size (list or None) – Voxel size as [z, y, x] in any consistent unit (e.g. millimetres from read_omezarr). Used to set the correct physical aspect ratio so that cross-sections look geometrically correct. If None, aspect=’equal’ (1 pixel = 1 pixel, which distorts anisotropic volumes).

  • crop_to_tissue (bool) – When True, crop the volume to the non-zero bounding box (with a small margin) before rendering. This removes empty space caused by motor drift and canvas inflation, making the tissue fill the panels.

Return type:

None