Vanilla javascript and some css on SVG on canvas and back from bitmap input.

Press "s" to download a 1200x1200 pixels png image (if"opened" in full page)
Press "l" to load and use an image of yours (if"opened" in full page)

The most interesting aspect of this work is the use of some public domain artwork bitmap images as input dataset to generate a randomized grid of blended triangular surfaces.

Triangle is the basic shape from which it's possible to construct all other shapes, at least on a finite discrete grid like our screens. How not to adore triangles blessing our screen?

The goal of this work is not trying to catch the "essential", the "vibe", or the "atmosphere" from an input artwork, neither to parasite it and shine with it's reflected light. The input images are a way to random noise, but you can use your own and you'll see that Triagulatria's own character will be preserved and you'll have even more good time.

For this new version (3.0.x) I want to reduce the gamma of possible outputs of each release by progressive enhancements of the codebase by refactoring and additions and the introduction of culling algorithms in the process exploring different variations of the configuration Glassamatz.

GLASSAMATZ-ZINCBACH

The Zincbach configuration variation is a test of the filterCircle culling set having concentric circles containing progressively 3,2 and 1 vertices.
I also substituted some input images with better scaled ones (the same painting) and I added a diiferent version of the cover of "Ver Sacrum - Januar 1898" by Alfred Roller (red ink print, the other one is brown with a different "trim")