Filter

Filtering images for AFMSlicer processing.

Classes:

Name	Description
SlicingFilter	Filtering and flattening of images for AFMSlicer.

SlicingFilter

Bases: Filters

Filtering and flattening of images for AFMSlicer.

Parameters:

Name	Type	Description	Default
topostats_object	AFMSlicer	An AFMSlicer object of the image to be filtered and flattened.	required
row_alignment_quantile	float	Quantile on which to align values, default is 0.5 (i.e. the median).	None
gaussian_size	float	Amount of Gaussian blurring to perform.	None
gaussian_mode	str	Method of Gaussian blurring to use.	None
remove_scars	dict[str, bool | int | float]	Whether to remove scars or not. This is performed using TopoStats' scar removal method.	None

Examples:

The final image after all stages is stored in the SlicingFilter.image dictionary with the key gaussian_filtered.

>>> import numpy as np
>>> from topostats.classes import TopoStats
>>> from afmslicer.filter import SlicingFilter

>>> rng = np.random.default_rng(seed=32424308)
>>> small_array = rng.random((5, 5), dtype=np.float64)

>>> afmslicer_object = TopoStats(
>>>     image_original=small_array,
>>>     filename="small_array",
>>>     pixel_to_nm_scaling=1.0,
>>>     img_path="./")

>>> filter_config = {
>>>    "row_alignment_quantile": 0.5,
>>>    "gaussian_size": 1.012139,
>>>    "gaussian_mode": nearest,
>>>    "remove_scars": {
>>>        "run": False,
>>>        "removal_iterations": 2 # Number of times to run scar removal.
>>>        "threshold_low": 0.250 # lower values make scar removal more sensitive
>>>        "threshold_high": 0.666 # lower values make scar removal more sensitive
>>>        "max_scar_width": 4 # Maximum thickness of scars in pixels.
>>>        "min_scar_length": 16
>>>    }
>>> }

>>> slicing_filter = SlicingFilter(
>>>     topostats_object=afmslicer_object,
>>>     **filter_config)
>>> slicing_filter.filter_image()

>>> slicing_filter.images["gaussian_filtered"]

Methods:

Name	Description
filter_image	Filter the image.

filter_image

filter_image()

Filter the image.

The following steps are performed to filter the image.

• Median flattening.
• Tilt removal.
• Quadratic removal.
• Nonlinear polynomial removal.
• Scar removal (optional).
• Zero average background.
• Gaussian filtering.

The methods used are inherited from TopoStats.Filters class.