MedianDenoise

MedianDenoise applies a median filter to a 2D intensity image. It uses SciPy when available and a small NumPy fallback otherwise.

Inputs

• input_image: 2D intensity image.

• radius: median window radius in pixels.

Outputs

• output_image: median-filtered float image.

Dependencies and Core Libraries

imageio, NumPy, SciPy median filtering when available, and a small NumPy fallback.

Assumptions

The image is 2D and contains impulse-like noise where local median filtering is a reasonable baseline.

Minimal Example

from bioimageflow_core import Arguments
from bioimageflow_restoration_tools import MedianDenoise

MedianDenoise().process_row(
    Arguments(input_image="noisy.tif", radius=1, output_image="median.tif")
)

Expected Results

Synthetic salt-and-pepper fixtures lose isolated outliers while preserving image shape.

Failure Modes

Unreadable images, unsupported dimensions, invalid radius values, and write failures raise errors.