Exploring AI Image Editors: Features, Limitations, and Recommendations

[eleanormiles]

Hi everyone,

I’ve recently started experimenting with AI image editors and am fascinated by the possibilities they offer, from enhancing photos to generating creative art.

However, I’m still learning about the best tools and techniques. I’m curious about which AI image editors you recommend for both beginners and advanced users.

Specifically, I’d like to know about features like background removal, style transfer, and realistic image generation.

Are there any limitations or common pitfalls I should be aware of?

Also, which platforms provide the best balance between usability and creative control?

 

[eleanormiles]

Hi everyone,

I’ve recently started experimenting with AI image editors and am fascinated by the possibilities they offer, from enhancing photos to generating creative art.

However, I’m still learning about the best tools and techniques. I’m curious about which AI image editors you recommend for both beginners and advanced users.

Specifically, I’d like to know about features like background removal, style transfer, and realistic image generation.

Are there any limitations or common pitfalls I should be aware of?

Also, which platforms provide the best balance between usability and creative control visit site?

thanks in advance for any help

 

[Asher Kelman]

Eleanor,

Your question is timely and apt. The new AI potential of our familiar programs are opening a wide and terrifying array of new compositional possibilities. If one is a graphic designer, one can bypass us photographers!

I don’t pretend to be any expert in the available AI options but I can share my perspective as it relates to my creative needs.

So, speaking as a photographer, I want my best work to be the main subject I chose to photograph!

Start with Photoshop: it’s very easy to select a person or object and remove the background. That’s a very mature process as is simply removing objects or people you don’t want in your picture using fill with AI reference to continuing patterns, textures and completing partial object to fill in the empty space.

For sky replacement, ON1 appears to be the most mature.

For shifting skins color distribution of the faces, Phase One, “Capture One” could be the easiest and most advanced to use.

For very large magnification, your photographic integrity and style can be protected as you enlarge, but without losing fine detail. Topaz Megapixel AI brilliantly creates more pixels that appear native to the photograph. All this maintains the integrity of the photographers creative work. Topaz focus can correct focus or movement blur.

However, we can go further than that to completely omitting the camera!

I have little experience with AI creation of photos or modifications such as adding people in a helicopter but these could be fun to experiment with.

I might try asking the background to a full length portrait to be redwood Forrest or dawn by a pond in the woods with a mist clinging to the ground. That’s what I would find worth exploring.

Asher

 

[GarethKitchener]

I'm coming at this from a "camera club competition secretary" viewpoint. For our competitions, the rule is that everything in the photo must be your own work. AI selection tools are fine - they're just better versions of the tools we already had, but any generated content would make an entry ineligible. (For example: You can swap the sky - but the new sky must be a real sky from another one of your photos.) I extend that mindset to all of my own photos - It wouldn't feel right to present something as "my" photo if some of it had been AI-generated.
I've used Google Gemini, ChatGPT and Apple Image Playground to create images for "fun" purposes*, but this is totally separate to photography in my mind. Of these, I think Gemini is probably the best at the moment.

*as an example - I recently asked ChatGPT to analyse a series of detective novels and produce some images of the main characters.

 

[Doug Kerr]

Hi, Gareth,

I'm coming at this from a "camera club competition secretary" viewpoint. For our competitions, the rule is that everything in the photo must be your own work. AI selection tools are fine - they're just better versions of the tools we already had, but any generated content would make an entry ineligible. (For example: You can swap the sky - but the new sky must be a real sky from another one of your photos.) I extend that mindset to all of my own photos - It wouldn't feel right to present something as "my" photo if some of it had been AI-generated.

Well said. But the definition of what is not "AI generated" is very slippery.

Suppose for example in a "landscape" shot I took there was a bird flying by that was completely out of the premise of the shot. And I take it out with some advanced photo editor. It might seem that doing so did not comprise the "integrity" of the image (any more than if I cropped it to eliminate some irrelevant material at the edge of the taken frame, or just cropped the original frame so not all of the sky appeared in the "deliverable")..

But the photo editor might well "generate" a continuation of the sky background so there was no gap where the errant bird had been. Now it might well "generate" that with a skilled interpolation of the surrounding background, so one could perhaps justifiably say it was still my work. Or it might decide to "generate" that from part of some library of sky backgrounds the editor used. So is that result an "illicit" inclusion of "generated" materiel that was not my work?

I don't suggest any answer to this conundrum.

I'm sure all this makes the job of someone reviewing submitted images to confirm that they qualify under the "rules" of the specific competition (a role I think you sometimes have) very tricky.

Best regards,

Doug

 

[GarethKitchener]

Hi Doug.
In the scenario you described, it would be permitted if the pixels that replaced the bird were copied from elsewhere in the image, whether that was by cloning, healing or content aware fill. However, 'generative fill' would not be permitted as that would take pixels from whatever library the AI used.
Policing is is obviously impossible though - I tell the members what the rules are and have to trust them not to "cheat"!

 

[Asher Kelman]

For competitions, one is allowed to

1. crop at will,

2. color correct and one has some license here too

3. adjust an S curve on the range of values in an histogram to have features defined to the extent one wishes.

4. One can likely blur/lighten or darken the background

5. One can delete an offending piece of trash however, wholesale generation of a new composition of forms and idiom that the camera couldn’t have captured at that instant, and one couldn’t, (even at a stretch), have noticed within the original scene deserves disqualification.

So removal of an OOF happenstance bird, that entered the frame is no problem with the smart fill looking to continue surrounding texture.

However, removing a flock of birds, even if they are replaced with cloned pixels or else, introducing “an eagle swooping down on its prey”, would be, of course unacceptable.

So one has to look closely at the rules of a photography competition before using AI!

Asher

 

[Doug Kerr]

Hi, Asher,

For competitions, one is allowed to

1. crop at will,

2. color correct and one has some license here too

3. adjust an S curve on the range of values in an histogram to have features defined to the extent one wishes.

4. One can likely blur/lighten or darken the background

5. One can delete an offending piece of trash however, wholesale generation of a new composition of forms and idiom that the camera couldn’t have captured at that instant, and one couldn’t, (even at a stretch), have noticed within the original scene deserves disqualification.

I had no idea. Are these the guidelines for some particular competition? Or is this what seems to be typical today?

So removal of an OOF happenstance bird, that entered the frame is no problem with the smart fill looking to continue surrounding texture.

However, removing a flock of birds, even if they are replaced with cloned pixels or else, introducing “an eagle swooping down on its prey”, would be, of course unacceptable.

So one has to look closely at the rules of a photography competition before using AI!

How might we "not use AI" if our current editor has that as part of its toolbox? Can we tell the typical modern photo editor "no AI, please"?

Best regards,

Doug

 

[Asher Kelman]

My advice, is to limit AI with generation of original additions to the world of your actual photograph. Further, should be confined to a personal portfolio of a related series of your treasured novel work, where AI was organically necessary to add the parts that do not exist for you, for whatever reason.

To express your imagined scene, you assembled it in the Cathedral of your mind. To express and export it as a “physical reality” using a photograph plus generative AI is fine and kosher, LOL!

However, to be successful with others, it had better be something that stands out as a work of art with your voice, fingerprints, thematics, iconography and/style that says your name and does not appear generic!

Then the solo exhibition could successful or you might choose to be united and share costs with several like-minded colleagues, but each with their own distinctive oeuvres.

If your work would be lost amongst your colleagues, keep it to yourself.

It should be that art moves some boundaries and turns some heads.

So this is a blessing for AI in the context of original photography as a work of art!

Asher