SFG conversion, xBRZ, and similar weirdness

[Katie Boundary]

:eek:

Has anyone seen these? xBRZ seems like the perfect upscaler for cartoons, and SFG... well, see for yourself: http://hplusmagazine.com/2014/12/11/gain-superpowered-vision-scalable-function-graphics/

DUDE. Just... DUDE. I have never seen black magic this powerful. How many souls of the innocent were fed to an eldrich abomination to get results like that?

[jmac698]

Nice, but need a comparison to something better than bicubic.

The original 40x40 samples weren't available in the article.

I got the original program now, but the problem is it's extremely slow and I can't process the standard images.

Also see this new algorithm for pixel art:
http://research.microsoft.com/en-us/um/people/kopf/pixelart/paper/pixel.pdf

[foxyshadis]

SAR's image upsizing options (http://www.general-cathexis.com/interpolation/index.html), and there are newer options not included in that demo. Hey look, SAR is free now!

BTW, comparison shots of the new pixel upscaler (http://research.microsoft.com/en-us/um/people/kopf/pixelart/supplementary/index.html).

[jmac698]

haha funny thing is, I used SAR on extracted thumbnails from the article already, then I realized the article only showed a 20x20 version. I didn't get good results with SAR in that case. DDL seemed the best.

[Katie Boundary]

do those "swirly" "wispy" artifacts have a name?

[jmac698]

Not that I know of, but I've seen them before, such as in mean curvature noise reduction. In that case, they literally come from curvelets, which is a formula to represent curvy things in pictures.

[Katie Boundary]

Then I shall call them...

curlies