m3resp.PixelMask#

class m3resp.PixelMask(name, values, modality='eit', method=None, metadata=<factory>, *, keep_zeros=False, suppress_value_range_error=False, suppress_zero_conversion_warning=False, suppress_all_nan_warning=False)[source]#

Which pixels of an EIT image belong to a region.

Each pixel holds NaN when it is outside the region, 1 when it is included, or a weight from 0 to 1 when it counts partly. Applying the mask multiplies each pixel by this number, so a NaN pixel drops out and a pixel with weight 0 adds nothing to a sum.

The checks follow eitprocessing’s PixelMask:

A 0 becomes NaN, with a warning. Pass keep_zeros=True to keep it as weight 0, or suppress_zero_conversion_warning=True to skip the warning.
A grid of true/false values is accepted: true becomes 1 and false becomes 0, which then becomes NaN without a warning (or stays 0 with keep_zeros=True).
A number below 0 or above 1 raises ValueError, unless suppress_value_range_error=True.
A grid that is NaN everywhere gives a warning, unless suppress_all_nan_warning=True.

Variables:

name (str) – What the mask is, e.g. 'tiv_lungspace_mask'.
values (numpy.ndarray) – The mask as a 2D grid of shape (row, column), stored as floats.
modality (str) – The device the mask belongs to.
method (str | None) – Name of the method that made the mask.
metadata (dict[str, Any]) – Optional extra information.

Raises:

ValueError – If values is not a 2D grid, or holds a number below 0 or above 1 while suppress_value_range_error is False.
TypeError – If values cannot be stored as floats without changing them.

Parameters:

name (str)
values (ndarray)
modality (str)
method (str | None)
metadata (dict[str, Any])
keep_zeros (dataclasses.InitVar[bool])
suppress_value_range_error (dataclasses.InitVar[bool])
suppress_zero_conversion_warning (dataclasses.InitVar[bool])
suppress_all_nan_warning (dataclasses.InitVar[bool])

Methods

to_dict

Return the mask and its descriptive fields as a dictionary.

Attributes

`included_pixel_count`	Return the count of pixels that are not NaN, weight 0 included.
`keep_zeros`
`method`
`modality`
`shape`	Return the number of rows and columns, in that order.
`suppress_all_nan_warning`
`suppress_value_range_error`
`suppress_zero_conversion_warning`
`name`
`values`
`metadata`