Prompt for ai Generative Fill about remixing existing work of mine?

Question

Hi Ho.

As most of us, I am more impressed with AI implemented in Photoshop, either in Betas or official. The best feature ever! And few quick questions regarding this:

1 - Can I open HD canvas (1920x1080) and generate group of Clouds on transparent layer? If yes, what the prompt would be like?

2 - I opened some of my old Websites I've done in Photoshop. Can I flatten image, select all and do Remix with different skin based on my current work?

3 - In past I've done lots of backgrounds / Wallpapers for my system I still use today. Can I remix and content generate 10, 20, 50 and 100% similar conatent but with different colour, layout or create similar graphics based on original image (source / copied layer / copied wallpaper as source)? Also - if yes, what prompt would be?

4 - Can I use Illustrator (or Photoshop) and use "3 examples of different ideas based on..." then I add link to my website (deetronic.rs for example) as "source" and ask AI to generate me variatious based on source (my site)? I am interested what it can draw and what prompt to use?

... And finally:

5 - I load vector drawing, my own characters, load whole family of it (characters based on my family: mom, dad, son, sister, brother and a dog) and ask AI to draw me "Grandpa", 3 variations, based on loaded style and graphics - could AI draw me other characters (at least ideas) based on my characters? If yes - what to type and where (empty layer, filled layer or flattened layer?).

I believe it is not implemented how I imagine AI would work in Adobe, but we live in a really interesting times 🙂

p.s. one idea / request - I load video of my baby in a stroller, walking through the park, I record my son, camera is shaking (video is from 2013): I stabilize image so only background is moving / jittering while the focused image (baby) is not shaking, it is centered on screen. Imagine how cool it would be to AI fill all missing parts from around the stabilized object and get full, non-existent background (based on what is happening - "Walking with baby in a park" and AI will generate full, stabilized image, without any "black holes" during the recording 🙂

Tina_Irvine · Answer

Hi @Nykk_Deetronic,

Thank you for your questions. Those are really great ideas, and very specific. It would be best to play around and experiment with prompts in Photoshop and other apps (Illustrator, Firefly, Express...) to see the results you can achieve. Firefly's gallery has image + prompt examples which can be inspirational.

As far as transparent layers, the AI won't provide that; however, you can try a prompt such as "clouds on solid green background" which would give you an image that you can use to fairly easily isolate the clouds (image attached)
You can select areas in your image and use generative fill to replace with a prompt of your choosing. It might be ideal to do it this way. You can also try Structure Reference in Adobe Firefly. I think this might work best in your website scenario. https://firefly.adobe.com/
Please give it a try by experimenting with various prompts 🙂
The AI won't pull in / reference your website, sorry.
Please try experimenting with prompts and selection areas for this 🙂

This article is worth checking out for AI video features in Premier Pro (in reference to your video idea/request).

Hope this helps. Have fun with AI, and see what magic it can create for you! I agree that we are certainly living in interesting times!

Sign up

To post, reply, or follow discussions, please sign in with your Adobe ID.

Sign in to Adobe Community

To post, reply, or follow discussions, please sign in with your Adobe ID.

Scanning file for viruses.

This file cannot be downloaded