Semantic Keyer Processing

I’ve been trying to flex the ML Semantic Keyer to isolate certain areas of an actors face, lets say a cold sore on the corner of a lip. If i use ML Matte Extraction with two selectives output from Action, comp them downstream to create my cold sore matte (eg Lips x Right Side), will Flame be calculating each selective separately, or will it know that they are the same semantic key, just different selections in Analysis?

Is there a better way to allocate the resources in Action/Image when using multiple semantic keys?

Is it possible to comp two selectives with in Action so I can have the one action do all the work, or am i best comping the outputs with ML Matte Extraction?

Hey Zac,

Have you tried the ML face template and inputting a matte at the position of the cold sore? You’re essentially making a custom shape and positioning it in the right place UV wise. That might be simpler. Happy to make you a screen recording of the workflow if you haven’t had a chance to try it.

I have not! I’ll give it a try tomorrow and get back to you on your offer. Thanks Andy!

1 Like

I just remembered….I covered this on Logik Live #25!

Logik Live Session 025 ML with Andy and the New Forum with Randy!

About the 14min mark.


ANDY! Thank you so much. I really need to get through all these.


This is seriously life changing black magic stuff.