WienerFilter

class of dascore.proc.wiener inherits from: dascore.proc.filter._WindowFilter, PatchProcessor, DascoreBaseModel, pydantic.main.BaseModel

source

WienerFilter(
    *args ,
    noise: Any = None,
    samples: Any = False,
    **kwargs ,
)-> None

Apply a Wiener filter to reduce noise in the patch data.

The Wiener filter is an adaptive filter that reduces noise while preserving signal features. It estimates the local mean and variance within a sliding window and uses these statistics to determine the optimal filtering.

Parameters

Parameter	Description
patch	Input patch.
noise	The noise-power to use. If None, noise is estimated as the average of the local variance of the input.
samples	If True, values specified by kwargs are in samples not coordinate units.
**kwargs	Used to specify the window sizes for each dimension. Each selected dimension must be evenly sampled. It works best when the window samples are odd.

Returns

Patch with noise-reduced data.

Examples

import numpy as np
import dascore as dc
# Get an example patch and add noise
patch = dc.get_example_patch()
noisy_data = patch.data + np.random.normal(0, 0.1, patch.data.shape)
noisy_patch = patch.update(data=noisy_data)

# Apply Wiener filter along time dimension with 5-sample window
filtered = noisy_patch.wiener_filter(time=5, samples=True)
assert filtered.data.shape == patch.data.shape

# Apply filter with custom noise parameter
filtered_custom = noisy_patch.wiener_filter(time=5, samples=True, noise=0.01)
assert isinstance(filtered_custom, dc.Patch)

# Apply filter along multiple dimensions
filtered_2d = noisy_patch.wiener_filter(time=5, distance=3, samples=True)
assert filtered_2d.data.shape == patch.data.shape

Note

This implementation uses scipy.signal.wiener which performs adaptive noise reduction based on local statistics within the specified window.

Methods

Name	Description
check	Refuse a patch which does not carry what the operation needs.
get_metadata	Return the window, once one is known to be given.
model_copy	Copy the model, dropping cached values the update invalidates.
new	Create new instance with some attributed updated.
numpy_kernel	Return the Wiener-filtered data.
wiener_filter	Apply a Wiener filter to reduce noise in the patch data.
reconcile	Return metadata or a patch holding the final data; default as is.
run	Run the operation: check, get_metadata, kernel, reconcile, record.