P: Inpainting for text to image

[Lavanya232490872d40]

I would really love a feature where we can specify the part of the AI-generated art that we want to change. Instead of having the AI generate new pictures each time, with completely random changes, I want to be able to specify which parts I have a problem with. For example, if the hands are not realistic, I want to be able to select them and generate images with just the hands changed (everything else the same). Or, if I really like the picture, but want to the see it in a purple hue, i want to be able to do that. I understand it might be really difficult to implement though. 

[matthew274739933ocp]

It would be really cool to have a masking tool that allows to brush areas to include or disclude the generated areas and prompts

[Kaan29356951el2n]

I totally agree 

[Tracy.H]

@matthew274739933ocp Merging your thread here so we can track the demand for certain features.

[juniorp11296369]

hola , tengo mucho tiempo usando photoshop y con los avances de la IA creo que adobe puede hacer un boom en cuestion de cambios de fondos , uso mucho el impaiting de stable difusion y he podido hacer muchos cambios de fondos pero la herramienta no es tan genial como podria ser si lo hace adobe , enviare una imagenes para que vean , ese feature aca donde vivo muchos fotografos les encantaria usarlas y mas si viene de adobe 

Merging your thread here so we can track the demand for certain features.

[JacoSteenkamp]

Adding a Mask feature to renders. 

This could give the user flexability to make adjustments to masked areas or objects within the render, by having the freedom to "keep" or "adjust" masked areas of the render and thus maintaining sections of the original render therefore only create new variations of the masked areas, means the process is less distructive. 

I've created a mockup of what this could look like. ( see attachment )

Thank you for another amazing product Adobe !

Jaco Steenkamp 

Merging your thread here so we can track the demand for certain features.

[JacoSteenkamp]

Thank you @Tracy.H the inpanting option is not available in my version of Firefly.
How would I get it activated ?

[Tracy.H]

@JacoSteenkamp Inpainting is not available yet. We will announce when it becomes available in the beta.

[Hannah Nutron]

would be nice to be able to circle an area for it to regenerate that spot, sometimes there are things that don't belong and being able to tell the AI to remove it would be excellent!

Merging your thread here so we can track the demand for certain features.

[CyprianEdits]

Add the ability to highlight and select specific aspects of a generated image, and use those to generate new images while discaring the rest of the image. This would make narrowing down designs easier and allow us to create more specific stylized works

Merging your thread here so we can track the demand for certain features.

[loais85013400]

and when it will released?

[Milena296246255yjk]

a house made of chocolate

[Coolsammy]

Sometimes the image is almost perfect except for a small area or sometimes you want AI to regenerate the majority of the image except for a small area that users want to retain in the next iteration.  A masking brush with a either a "Keep" or "Replace/Fix" toggle would be useful and save processing cycles if only a small area needs to be corrected.

Merging your thread here so we can track the demand for certain features.

[georginab69694769]

Is there an estimated release date for the "in exploration" features?

[Michael_jms]

I want to see this feature live. 🙂 Maybe we could have a little sneakpeek into it in "meeting the team" 😛

[Matthew241430003xuj]

For Firefly Text to Image:

1. It would be nice to have a lasso tool where I could select a part of a result and refresh based off that selection.

2. It would be great to lasso portions of multiple results and refresh based on a combination of those selections.

Merging your thread here so we can track the demand for certain features.

[bfyybfyy9]

hope there will be a feature that let us mask a specific part of the image and re-generate what's in there, e.g. computer screen content, what's outside the window

Merging your thread here so we can track the demand for certain features.

[Serjiio]

having a tool to select a portion or certain subject in the image so you dont have to regenerate the whole result.

Merging your thread here so we can track the demand for certain features.

[Esdras29965127k2f5]

possibilidade de exportar a imagem escolhida para uma área de edição onde seja possível a alteração de detalhes específicos da imagem através de uma ferramenta de borracha e novo prompt ou algo semelhante.

Parabéns pelo trabalho!

Merging your thread here so we can track the demand for certain features. You can upvote the original post to let us know you want this feature.

[Tracy.H]

We have released the Generative Fill module in Firefly and in the Photoshop beta. Please give it a try!

[bogdanm99912979]

I was thinking a cool idea to improve on the image generation is to have some kind of selection tool and then a prompt to iterate on that part of the image that was selected. You could have a really good character image generated, and want to use that, but the hand might be off or the color of the jacket would be better if something else.

So I imagine a selection tool with a prompt, so I can say "give me more options for this hand", or "change the hair style to be more like klingon", or "add a spalsh of color here". 

Does that make sense?  

[Anusha302375353a3o]

Any subject can be easily removed from the image but also keeping in mind to fill the spot with content aware method like in photoshop.

Merging your thread here so we can track the demand for certain features. You can upvote the original post to let us know you want this feature.