If you’re nervous about Generative AI imaging, or opposed to it, this isn’t going to help. I’m going to demonstrate a new ability of generative AI just released by Adobe.
The image to the left and featured above is generated by Adobe Firefly using Google’s Gemini with Nano Banana. Adobe Firefly can now take a finished image and use it in a generated image, either by editing it according to instructions or by including the given image in a larger one generated by the AI. Experimenting with this, I’ve discovered this incredibly powerful method.
This is based on my photograph of the New Mexico countryside that I have already published. My prompt directed the AI to keep this image but edit it in the style of Georgia O’Keeffe. It blows my mind how good is the result.
For comparison, this is the original photograph.
Here is another one that is, to me, even more startling.
I published this tulip a while back. Preparing the image for this exercise, I removed the foliage background. I expanded the image horizontally to create a larger image. I sharpened the image and increased exposure and contrast ever so slightly. Then I added a neutral background texture. The result at this stage is below.
Then, as with New Mexico image above, I asked Adobe Firefly using Google’s Gemini with Nano Banana to edit this image in the style of Robert Mapplethorpe. I have long admired the photographs of Mapplethorpe and especially his exquisite flower images. Here is the result. Note that I gave no further instructions. The choice of black-and-white, and the background, for instance, were the choice of the AI, echoing of course Mapplethorpe’s style.
This makes a marvelous black-and-white image that rewards prolonged study. As with the O’Keeffe above, it is remarkable how much this looks like a photograph by Mapplethorpe. But it’s not a photograph by Mapplethorpe, and it would be quite wrong to pass it off as one.
As many of my mages in this New Images series make clear, I am not uncomfortable using AI to develop and/or edit interesting and appealing images. This new technique or feature in Adobe Firefly is interesting but may have limited legitimate uses.
However, I have found one practical and I think legitimate use of this technique.
I created a new feature image for my page on buying art (still in development). Frequently, while offering our work for sale, we want to show the works as they would appear in a room. There are apps and programs to do this, but I found the new AI technique useful.
The picture is one of the Los Banos, Callifornia, images. I asked the new AI to produce a typical contemporary living room with the Los Banos photograph, framed, on the wall behind a sofa.
As always, your comments are welcome. Visit again next week for a new installment of New Images.
Leave a Reply