bioimageflow-restoration-tools¶
bioimageflow-restoration-tools provides restoration inference wrappers, simple image-processing comparisons, and image-quality metrics.
It is designed for workflows that compare noisy or blurred microscopy images with restored outputs.
Install-time libraries are imageio and NumPy.
CAREamicsPredict runs CAREamics inference from a checkpoint.
The simple filters are documented as image-processing methods rather than learned restoration models.
SAIRPICO command-line restoration tools live in bioimageflow-sairpico-tools.
Tools¶
RestoreImage: apply a TV, bilateral, Gaussian, or unsharp-style restoration method.
CAREamicsPredict: CAREamics restoration prediction from a checkpoint.
RestorationMetrics: compare degraded and restored images against a clean reference.
GaussianDenoise: Gaussian smoothing.
MedianDenoise: median denoising.
BackgroundSubtract: subtract a smoothed background estimate.
UnsharpMask: sharpen by adding high-frequency residuals.
RichardsonLucyRestoration: Richardson-Lucy deconvolution.
Workflow Use¶
Use CAREamicsPredict in restoration workflows when a CAREamics checkpoint is available.
Use RestorationMetrics to compare noisy and restored outputs against a clean reference or curated validation crop.