Skip to content

Latest commit

 

History

History
56 lines (47 loc) · 3.25 KB

process.md

File metadata and controls

56 lines (47 loc) · 3.25 KB

process

Calibrate and deskew an acquisition, optionally deconvolving it. Saves uint16 output beside the acquisition by default.

With --deconvolve, 3D scan refinement is inferred from recorded spacing toward 0.4 micrometers. For example, a 1.2 micrometer acquisition uses factor 3. Already finer acquisitions retain their spacing; coarser steps must be integer multiples of 0.4 micrometers. Measured PSFs must be sampled at the inferred reconstruction spacing.

Option Description
--deconvolve Apply RLGC deconvolution before deskewing.
--flatfield-correction Include illumination in the deconvolution measurement model, or divide calibrated data by it when deconvolution is disabled.
--save-float32 Preserve fractional intensities in float32 output.
--skip-empty-below VALUE Zero channel tiles with insufficient signal above this intensity.
--skip-empty-min-signal-fraction VALUE Required fraction above the cutoff; default 0.01.
--resume Continue compatible completed tiles. If omitted, partial outputs offer Resume / Overwrite / Cancel (default: Cancel); completed outputs are overwritten.
--output PATH Set the output directory.
--live PATH Process during acquisition using a supplied CYX illumination TIFF; pass the acquisition directory.
--live-wait Wait for successful live completion and all raw tiles, then process offline. Combine with --flatfield-correction to estimate missing illumination. Mutually exclusive with --live; pass the acquisition directory.
--time-range START STOP Select timepoints.
--pos-range START STOP Select positions.
--z-downsample-level VALUE Deskewed Z reduction factor; default 2.
--crop-after-deskew Crop Y to the region supported across all Z planes.
--no-max-projection Disable per-tile and stage-placed maximum-Z images.
--no-create-fused-max-projection Disable the stage-placed maximum-Z mosaic.
--write-fused-max-projection-tiff Also save that mosaic as OME-TIFF.
--decon-crop-scan VALUE Set scan planes per deconvolution chunk; default automatic.
--decon-gpu-id VALUE Select the CUDA device; default 0.
--decon-verbose VALUE Opt-in solver diagnostics: default 0 (off); 1 reports chunk geometry and iterations.
--decon-fallback-step-scan VALUE Set chunk reduction after GPU allocation failure.
--decon-psf-paths PATH Supply channel-ordered PSFs; repeat for each channel.
--eager-mode Legacy planar KL-stopping option; has no effect with MATLAB-default GC stopping.
uv run process "/path/to/acquisition" --deconvolve --flatfield-correction

Start before acquisition finishes, without a precomputed illumination file:

uv run process "/path/to/acquisition-directory" --live-wait --flatfield-correction --deconvolve

See the live acquisition guide for required controller sidecars and completion behavior.

Append --help to list all options and defaults.

The photon-GC solver revision is recorded with the processing configuration. Resume rejects incompatible solver configurations; start a fresh output or intentionally overwrite instead. Processing without deconvolution retains its existing calibration and flatfield behavior.