Stitching and preprocessing example

This example will show how tiled data is stitched and prepared before multivariate analysis.

This page is intentionally workflow-oriented. For the detailed meaning of every stitch setting, see 01b Stitching tile folders. For the detailed meaning of physical-units and rolling-ball settings, see 04 Physical units and rolling-ball correction.

Planned workflow

  1. Select a tile folder.
  2. Configure filename parsing.
  3. Set overlap and scan direction.
  4. Choose grid or correlation-assisted stitching.
  5. Inspect the stitched image.
  6. Apply physical units and optional background correction.
  7. Save the stitched TIFF.
  8. Run analysis on the stitched result.

Points to highlight

  • Tile filename parsing.
  • Variable overlap.
  • Scan direction correction.
  • Correlation-assisted alignment.
  • Physical pixel size.
  • Rolling-ball or background handling.
  • Export of stitched and analyzed data.

Decision guide

Use this workflow when the main question is not only "how do I stitch tiles?" but also "which preprocessing choices should I commit to before analysis?".

Typical decisions are:

  • whether the tile folder is geometrically reliable enough for grid placement alone,
  • whether correlation should refine the tile positions,
  • whether broad background should be removed as preprocessing or kept for later modeling,
  • and whether the stitched result should already carry correct physical scale information for export.

Recommended defaults:

  • start with the correct filename parsing and preview table first,
  • use the microscope overlap values before tuning them manually,
  • try correlation when small seams remain after a grid stitch,
  • prefer reference illumination correction when many similar tiles share one stable shading pattern,
  • and verify physical units before exporting or placing publication scale bars.

Placeholder: insert tiled example dataset.

GIF placeholder: stitch tiles, correct scan direction, and run analysis.