Skip to content

Comment on Image Scaling Using Deep Convolutional Neural Networks

Comments

I wonder how this approach compare to upscaling via IFS (i.e. "fractal compression"-like techniques). Although the underlying modelling is very different, the fundamental idea seems very similar - learn vectors from the image, apply (possibly a mixture of) them at a different scale. The results certainly look superficially similar.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.