Take a look at the image above and ask yourself, which of the three is DEFINITELY an AI image. Then consider, which one HAS TO BE a photo? After completing this and showing it to friends, EVERYONE got both answers wrong. But the amazing part of this test isn’t found in the realism of the imagery, but in the congruent similarity between the real and the AI generated images.
In this article, I’ll explore the IMAGE GUIDANCE feature in Leonardo AI and how it can be used to generate nearly precise imagery and prompt generation based on traditional photography. For example, if you have pictures from a traditional photo shoot where the model’s expression or positioning wasn’t quite perfect, you can now take that image into Leonardo AI, add Image Guidance, and “using the same model” correct to your hearts content. This is more than Face Tuning, and it certainly goes farther than the seed option in Midjourney.
In fact, it can be used to create similar compositions but with more unique opportunities, and flair than you may have considered with standard prompts. But, even better, the more realistic the original photograph, the more realistic the final Leonardo AI Generation. In a previous article, I combined both Midjourney and Leonardo AI for some amazing results, but the final generations still looked a bit too AI.
In Search of AI Perfection: Midjourney AND Leonardo.ai
Another “combination” article that just may change the game this time?
But note the difference in the original photo I used for Image Guidance, and these final generations (by the way, did you get it right based on my original questions above?).
In this case, the overall lighting improved, the colors were toned to a more realistic hue and the faces were softened but still realistic. Now, note the…