bigstitcher-documentation

name: bigstitcher_documentation description: A Fiji plugin for stitching and fusing multi-tile, multi-angle, multi-TB microscopy datasets.Stores all state in an XML project file; uses BigDataViewer for interactive display. Primary use cases are cleared-tissue lightsheet stitching, tiled confocal reconstruction, multi-view lightsheet registration. Read the files listed at the end of this SKILL for verified commands, GUI walkthroughs, scripting examples, and common pitfalls.

BigStitcher — Skill Quick Reference

Can BigStitcher Be Automated via Groovy?

YES — fully. All BigStitcher processing steps are exposed as macro-recordable commands under Plugins › BigStitcher › Batch Processing. These are callable from Groovy (and any other Fiji scripting language) via IJ.run(). The complete pipeline — dataset definition, phase correlation, global optimization, fusion — can be run from the Fiji Script Editor:

IJ.run("Calculate pairwise shifts ...", "select=/path/to/dataset.xml ...")
IJ.run("Optimize globally and apply shifts ...", "select=/path/to/dataset.xml ...")
// NOTE: In many BigStitcher versions the fusion command is recorded as "Image Fusion" (not "Fuse dataset ...").
// Always confirm the exact command name via Plugins › Macros › Record…
IJ.run("Image Fusion", "select=/path/to/dataset.xml ...")

Use the Macro Recorder (Plugins › Macros › Record…) to capture exact parameter strings while working interactively — then paste them into your Groovy script.

Processing Pipeline (Stitching Mode)

Step 1 — Define dataset         →  IJ.run("Define Multi-View Dataset") *(or similar; command name is version-dependent — use Macro Recorder)*
Step 2 — Calculate pairwise shifts  →  IJ.run("Calculate pairwise shifts ...")
Step 3 — Filter links               →  IJ.run("Filter pairwise shifts ...")
Step 4 — Global optimization        →  IJ.run("Optimize globally and apply shifts ...")
Step 5 — ICP refinement (optional)  →  IJ.run("ICP Refinement ...")
Step 6 — Fuse / Image Fusion         →  IJ.run("Image Fusion")  *(command name depends on BigStitcher version; use Macro Recorder)*

All steps are fully automatable via IJ.run(). Each call is synchronous and operates on the shared XML project file.

Key Parameters (All Steps)

Critical Pitfalls

1. grid_type trailing spaces cause "unrecognized command" — this is the most common failure with Define Dataset. The value [Snake: Right & Down ] has exactly 6 trailing spaces. Missing even one produces a silent mismatch. Use the Macro Recorder once to capture the exact value for your scan direction.
2. Trailing comma in fix_group_0-0, — required; missing it breaks the global optimization reference frame.
3. Re-save to a different path than input — writing HDF5 into the same folder as raw files is allowed but can cause confusion; use a subdirectory.
4. Fusion before optimization — fusing with unregistered positions produces a broken output. Always run optimize → (ICP) → fuse in order.
5. ICP requires sufficient shared signal — if tiles share little overlap content, affine ICP will diverge. Use translation model as fallback.

Automation Pathways Summary

File Inventory