starlet_filter_norms

starlet_filter_norms#

scarlet2.wavelets.starlet_filter_norms(scales, levels=None, generation=2, convolve2d=None)[source]#

Compute the L2 norms of the starlet analysis filters.

The starlet transform is linear, so white noise with standard deviation sigma in the image yields coefficients at level j with standard deviation sigma * norms[j]. The norms are determined from the transform of a unit impulse.

Parameters:

  • scales (int) – The number of starlet scales, as in starlet_transform(). The transform has scales+1 levels, the last one being the coarse residual.

  • levels (list of int, optional) – The levels (0 being the smallest) for which the norms are returned. If None, all scales+1 levels are returned.

  • generation (int) – The generation of the transform. This must be 1 or 2. Default is 2.

  • convolve2d (function) – The filter function to use to convolve the image with starlets in 2D.

Returns:

Filter norms for each of the requested levels.

Return type:

array

Examples

>>> norms = starlet_filter_norms(4)
>>> [round(float(n), 3) for n in norms]
[0.943, 0.23, 0.1, 0.049, 0.03]