I would very much like a “Bokeh AI”
I use Micro Four Third cameras, and I have 2 stops disadvantage of shallow depth of field. My favorite subjects are birds, and you not always can influence the background accordingly.
What your algorithms really do well is masking - at least a lot better than any auto masking algorithm I know so far. So it should be pretty easy for you to create. And in the beginning, I would not need AI to do the blur - just a nice control pin
But I see a lot of potential to improve some of my pictures with such a tool
@martin-0230 Thanks for your suggestion! I’m glad you like our masking functionality. A large portion of our users are wildlife and bird photographers so I can see this being useful for many people.
With increasing use of Apple and Android phones for serious work (my S22U is the all-but-birds camera) the 5 to 10 times depth of field is even more of an issue. Looking ahead to depth maps in meta data you could simulate lens blur for general use, like landscapes, instead of just a simple background blur.
While I the AI masking is good for me at least it is still a long way behind what Photoshop offers and that is still not perfect so until that level of detection is possible I would wait on offering this feature. I know it would be useful to have I think as Billray said unless there is a depth map it is not going to look great. To offer an accurate depth map I think PhotoAI would also need a good database of how about every lens works which may be not something in their roadmap to make.
A good depth mapper would be a nice to have to help guide the sharpening AI as you might be able to “refocus” the image.
That makes sense, we don’t want to build a worse version of what Photoshop already offers. If we do release something like this, there will definitely be an AI-enabled differentiator that makes the results better than what you can do manually.