Did aximmtery support Image-Based Keying?
I mean something like
https://learn.foundry.com/nuke/9.0/content/getting_started/tutorial3/image_based_keying.html
or
https://docs.zerodensity.tv/reality25/appendices/appendix-d-reality-keyer/what-is-image-based-keying
and do it have advantages above chroma keying like this video tell?
https://www.youtube.com/watch?v=G82dsYZiePg
Yes, Aximmetry supports it for fixed-camera (VirtualCam*) usages.
On the KEYER panel select CLEAN PLATE GENERATOR as the keyer type.
Pick the Background Color the same way you do with any keyers.
If there isn't any object in front of the green (the camera shows the pure green only) then you're done and you can switch back to any keyer type you want to use.
If there are some objects that cannot be removed from the green Clean Plate Generator will try to eliminate them from the picture. You can play with the Threshold and Erode slider to improve the removal.
When you're done switch back to the keyer type you want to work with. Turn on Use Clean Plate.
Background Color is ignored in this case and the keyer will use the generated plate.
Note that in the current version the generated clean plate cannot be saved with the scene, you have to generate it each time after loading.