Show HN: Use machine learning and enlarge your image without losing quality

Very interesting concept, but it only kinda works with simple shape images, and doesn't work with more complex images (photograph) Using the resize filter, I ended up with images where the image elements edges are sharpened. So yes, the final images were not blurry but it definitely lost quality.

Update: To be fair letsenhance.io didn't do a good job with images I tested either. Maybe worse. And you are correct, their website tricks you into signing up just to test it.

I provided a fairly low resolution image which was capture from a 3d map tile. It enlarged and didn't lose quality, but what I was most impressed by was removing the jpg artifacts.

we know that letsenhance.io was relases some time ago, but: a) our _core_ code for this was relased 2 years ago, b) LE requires signing up c) there is no information how they approach this subject, as we give all information with detailed alogritms

Does anyone know how this compares to superresolution using fractal image coding?

Only result I get is:

"Something went wrong. Please upload image once again."

Why three are only a few filters? Can one add a new one via PR?