Copy link to clipboard
Copied
It should be doable and you probably have already everything you need to make it happen.
/Imagine: I just added a new collection to LR. After it generates thumbnails and do face recognition, it recognises scenes. Since I always do a lot of pictures on every event, I have like... 2137 images. Almost half of the pictures were taken in one space. Now when I edit selected ones I can remove entire people or objects because LR actually knows what's behind. Or I can add a person from another pictures. Or generate clean, perfect wide shots of entire venue with people. And with people. I can change angle of a scene or change the leneses. I generate scenes that I miss when reviewing all this. I can make a cinematic story with rich dataset. I can do this all night...
I wanted to say that I know it needs a tremendeus amount of raw pictures to feed the model but there's no "but". I wake up in different world everyday these days. Every idea ever posted here is doable now. For years in my photographers community there was funny story easy to hear: "so there was this wedding couple once, they didn't like that someone important was standing back to the camera on one image and... no way, you would never guess lol - they asked the photographer to rotate that man so they could see his face!" The same happened with this tale - the joke turned the other way now.
Have something to add?