First, I LOVE this product! Second, something I’d REALLY love to see. Photo AI recovers faces by analyzing the existing image and estimating what it’s supposed to look like. Most of the time it does an excellent job but sometimes the results are not so good or even bizarre. But what if the AI could analyze a SECOND image as a guide? Say you had an excellent photo of a person and another, out of focus image of the same person you’re trying to restore. If the AI could analyze the good image as a guide to what the person is SUPPOSED to look like and apply that information to the second image it might get more realistic and consistent results. Or is this TOO much to hope for? Regardless, thanks again for an amazing product!
1 Like
I’m so happy to hear you love the product!! We’re toying with the idea of being able to feed it sample images, eg. other photos of the person, and essentially training the model specifically for the image.
You can image how difficult this is technologically but we’re discussing it.
