Photoshop’s new AI feature quadruples the amount of pixels in your photos — WOW!
Adobe Photoshop has got a new AI feature that can quadruple the number of pixels in your photos.
The tool, called Super Resolution, is now shipping in Camera Raw 13.2 and will be coming soon to Lightroom and Lightroom Classic.
The feature uses a machine learning model trained on millions of photos to enlarge images while preserving their clean edges and fine details.
Programmer Eric Chan said it’s very simple to use:
Press a button and watch your 10-megapixel photo transform into a 40-megapixel photo. It’s a bit like how Mario eats a mushroom and suddenly balloons into Super Mario, but without the nifty sound effects.
The feature could breathe new life into photos shot with older cameras.
While it won’t generally be as handy for modern high-res cameras, it could still prove useful for certain tasks, like working with tightly-cropped pictures.
Adobe developed the feature by training the model on millions of pairs of low-resolution and high-resolution image patches.
These patches are crops from detailed regions of real photos, such as flowers and fabrics. When the model is given enough examples of different subjects, it learns to upscale low-res images while maintaining the details.
To use Super Resolution in Photoshop, just right-click on a photo and select “Enhance…” from the context menu.
The system will then spit out an enhanced Digital Negative (DNG) file that can be edited like any other image.
While Super Resolution works best on raw files taken straight from a camera, it can also enhance other formats, such as JPEGs, PNGs, and GIFs.
The feature is currently limited to images smaller than 500 megapixels, but that should be more than enough for everything except enormous panoramas.
Published March 11, 2021 — 14:28 UTC