How to use Photoshop Generative Fill to add, remove or extend

Here are some useful tips for adding, removing, or extending using Adobe’s new Generative Fill feature in Photoshop (Beta). Warning: It involves flying saucers.

What is Photoshop Generative Fill?

Adobe has incorporated Firefly (AI-generated art app) into Photoshop (Beta) 12.4.6. This allows you to add, remove and extend your image by typing in text prompts.

Or even not typing in text prompts, as we shall see.

Generative Fill attempts to match the original scene’s perspective, lighting, and style. In other words, it attempts to create an image that also has appropriate shadows, reflections, lighting and perspective.

Ethics

AI-generated art has come under fire. Some characterize it as a plagiarism tool because many systems mine millions of images in an effort to generate its own images. Adobe works around this by only having its model learn from Adobe Stock images and public domain content without copyright restrictions. Additionally, as part of Adobe’s Content Credentials feature, AI images made in Photoshop will be encoded with an invisible digital signature indicating whether it’s human-made or the product of AI.

We’ve described what it is. We’ve discussed ethics.

Now on to the fun stuff.

Generative Fill: Adding, removing and extending

I attempted three general sorts of generative fill: adding, removing, and extending. Yes, that means adding new elements (Bigfoot!), removing unwanted elements (ex-boyfriends or girlfriends?), or extending your image to make it wider or taller. I’ll also offer my opinion for instances you might find helpful while not creating “fake” images.

But I’ll create some fake ones too. You know. Because that can be fun.

Adding to your image using Generative Fill

Of course I wanted to add things first. We’re all curious about this, right? First, I attempted to add “a very realistic-looking metal flying saucer flying in the night sky.”

First, I selected the area of the sky where I wanted my flying saucer to go.

Then, I pressed “Generative Fill.” The field is blank. You can choose to leave it blank and let it attempt to fill in the area, much like it might with Content Aware. Or you can type in a text prompt. You can see my “flying saucer” text prompt above.

This was the first flying saucer that it generated. And also my favorite. It didn’t look as realistic as I would have hoped for, but it is a flying saucer. I found in later experiments that it didn’t seem to matter whether I put “realistic” or “detailed” in the text prompts.

But then again, who am I to judge? I have never seen a flying saucer from outer space. Maybe they do look just like this.

You might also notice that in addition to the flying saucer, it has also re-created the night sky around it.

Just for fun, I thought I’d try a “very realistic jet plane.”

I got a jet plane. I don’t know that it was “very realistic.” The back wings look rather askew.

But nonetheless, this technology is a lot of fun and rather amazing. If one is doing graphic design, this could be extremely handy and convenient, easier than inserting clip art.

For night photography, particularly since I have little interest in putting anything “fake” in my photos. But that doesn’t stop it from being fun.

I thought I would try one more generative fill. This time, my text prompt was “A woman in Western wear.” It produced the above. On a smaller image or on social media, this might be fine.

However, zooming in closely revealed a woman with a rather melted face.

Removing part of your image using Generative Fill

I thought I might experiment with a night photo of an old abandoned Nye County Search and Rescue vehicle in the Nevada desert. I noticed that I had “blown out” the left part of the windshield and part of the headlight. They were overexposed due to my light painting. It happens.

I could clone these two spots out or even use Content Aware. However, I was curious about how Generative Fill would work.

To remove these overexposed hot spots, it was better not to leave a text prompt. This would make Generative Fill act more like Content Aware, only with the added benefit of being able to leverage Firefly’s AI-generated art as well.

Above are the results of my first attempt. The fill looks extremely realistic. Best of all, it did this with no effort on my part, and looked better than Content Aware might have done.

I might consider using Generative Fill in the future to remove unwanted hot spots or minor issues such as graffiti with profanity or other unwanted elements. I suppose that means that my photo might be flagged with an invisible digital signature to indicate that it is created by AI. This of course blurs the lines between AI-generated art and photography. Then again, on a purely technical level, many feel that Lightroom Denoise AI or Topaz Labs Denoise AI already do this.

What’s your opinion?

Extending your image using Generative Fill

This time, I thought I would attempt to extend my image using Generative Fill. I made a selection. I found that since this was a full-sized photo, extending my selection by 30 pixels into the original photo yielded the best results.

This is how the image looked after applying the above. You can see a vertical stripe in the upper left part of the sky. It wasn’t perfect. But this can be fixed very easily.

And this is how the image looked when I added approximately 30% more to extend the image. About half the image is no longer my photo!

This is rather remarkable. I found this to work much quicker and more easily than using Content Aware Fill. And with considerably better results.

Why might I want to extend my photo?

Artistically, I don’t.

However, if I print my photos as a wrapped canvas with many printers, I lose up to 1.5 inches on the sides, especially if I choose to frame it. That’s a lot of the photo to lose.

If I print it on paper or metal, I don’t lose 1.5 inches of the periphery. But for many printers, a canvas wrap is very different.

Therefore, if I really want to print on canvas without having my image cropped, extending the edges is best. One way of doing this is to use Generative Fill. It doesn’t need to be perfect looking, but it can do wonders for these sorts of prints, particularly if your aspect ratio doesn’t exactly fit the frame perfectly.

Bonus tip: How to save a Generative Fill image

Above, the initial “Save As” is rather limited. Photoshop (Beta) only allows you to save this as four file types.

However, by selecting “Save a Copy,” this offers the usual array of file types.

Opinions?

I’ve offered some of my opinions above. I’ve found I have some “good use scenarios” for using Generative Fill for adding, removing, and extending images. And for me, these scenarios don’t feel like I am creating a “fake” photo.

But not everyone has the same view of this, or of AI-generated art or AI in general. What’s your opinion? What are some of the best uses of Generative Fill? Will you use it?