This tool aims at computing an approximation of a given image by applying SVD to it.
A possible application of such tool may be the creation of abstract images, where the level of abstractness depends on the choice of k, the degree of approximation.

Given the following stock image:

it is possible to obtain an highly abstract image like the following:

by setting a low degree (e.g. k=2).

By increasing the value of k, a progressively more highly-detailed approximation is obtained, like the following (k=50):

The upper bound of k is the number of eigenvalues obtained by SVD; to find out such number, the script must be run first, so it is suggested to initially set a low k (by default, k=2).

Mattia Orlandi