`skimage.morphology.footprints:rectangle`

source: /dev/scikit-image/src/skimage/morphology/footprints.py :256

Signature

   def     rectangle (    nrows    ,    ncols    ,    dtype    =  <class 'numpy.uint8'>   ,  * ,     decomposition    =  None     )

Summary

Generates a flat, rectangular-shaped footprint.

Extended Summary

Every pixel in the rectangle generated for a given width and given height belongs to the neighborhood.

Parameters

nrows : int: The number of rows of the rectangle.
ncols : int: The number of columns of the rectangle.

Returns

footprint : ndarray or tuple: A footprint consisting only of ones, i.e. every pixel belongs to the neighborhood. When decomposition is None, this is just a numpy.ndarray. Otherwise, this will be a tuple whose length is equal to the number of unique structuring elements to apply (see Notes for more detail)

Other Parameters

dtype : dtype-like, optional: The data type of the footprint.
decomposition : {None, 'separable', 'sequence'}, optional: If None, a single array is returned. For 'sequence', a tuple of smaller footprints is returned. Applying this series of smaller footprints will given an identical result to a single, larger footprint, but often with better computational performance. See Notes for more details. With 'separable', this function uses separable 1D footprints for each axis. Whether 'sequence' or 'separable' is computationally faster may be architecture-dependent.

Notes

When decomposition is not None, each element of the footprint tuple is a 2-tuple of the form (ndarray, num_iter) that specifies a footprint array and the number of iterations it is to be applied.

For binary morphology, using decomposition='sequence' was observed to give better performance, with the magnitude of the performance increase rapidly increasing with footprint size. For grayscale morphology with rectangular footprints, it is recommended to use decomposition=None since the internal SciPy functions that are called already have a fast implementation based on separable 1D sliding windows.

The sequence decomposition mode only supports odd valued nrows and ncols. If either nrows or ncols is even, the sequence used will be identical to decomposition='separable'.

The use of width and height has been deprecated in version 0.18.0. Use nrows and ncols instead.

Aliases

skimage.morphology.rectangle