In the example below, I downloaded a 310x420 pixels image, and increased its size 4 times. The standard model is supposedly the “one-size-fits-all” algorithm, but the other ones might be better for specific situations.Īs you can guess, the default mode will work, but it will definitely not be the best solution for all situations.

There are also 5 different AI models that you can choose from: standard, lines, artwork, low resolution and very compressed. This last one is the one you use if you want a face to be reconstructed or just amplified. The controls to the left allow you to increase or decrease the noise and blur removal, and also increase or decrease the facial recovery strength. After the photo is processed, you can adjust the result using the controls to the left. All you do is open the application, drag a picture into the window, and wait for the software to process it.

You can either download them when you first launch the application, or download them as they are needed (they download automatically when you process a picture, depending on what the picture has in it. The app uses 1.13Gb of space after installation, but the app also needs to install 3.0Gb of extra ML models to process images. For this article I am using the demo version I downloaded from their website.

The results are varied but I hope the app improves in time. The app apparently uses some machine learning techniques to process the images and figure out what to do. A few days ago I ran into Gigapixel AI, an application by Topaz Labs that supposedly is able to increase a picture size while keeping the details, and enhancing any lost details.

Topaz gigapixel ai model files