GaussianDenoising

class giant.image_processing.denoising.GaussianDenoising(options=None)

Uses gaussian smoothing to reduce noise in an image.

All we do is convolve a 2d gaussian kernel with the image. This has the effect of reducing noise spikes but also reduces sharpness in the image. It can be fairly effective for reducing noise in images only of point sources without removing too many dim ones. It is also generally fast.

Parameters:

options (GaussianDenoisingOptions | None) – the options to configure the class with

allowed_dtypes: list[DTypeLike] = [<class 'numpy.float32'>]

The allowed datatype. Technically uint8, int16, uint16, and float64 are also supported but we want the output to be float32 so force everything to that.

Methods

__call__

Call self as a function.