I have digital photos from the late 90s and early 00s that just aren't very high resolution. In fact I have my first-ever selfie that was taken on an Apple QuickTake 100 back in 1994 at 640✕480 pixels; ⅓ of a megapixel.
But neural networks and machine learning are making it possible to add detail and resolution to photos like this. It's basically making an "educated guess" about what detail was there and "restoring" it. But it does a pretty good job, and I think it gives a better overall effect to older, smaller, or noisier photos. Science fiction becomes reality. 😊
There are some open-sourced tools out there that you can use for free. But they certainly aren't the easiest to use. I'm sure more user-friendly tools will be coming out in the next year or two. In the meantime, the one I've been using is a site called Let’s Enhance. It’s a paid service, but they do give you 5 free conversions when you sign up. And if you are really strapped for cash, I’m sure you can create multiple accounts to extend your number of free conversions.
Comments