Hello Jaka
Thanks for your answer and also to your team for preparing a Unity example. Yes indeed the usecase is to use a 416*416 pipeline to keep load low and then, when a certain object is detected, grab a full resolution image (like "setCaptureStill(True)" in Python), but this is going to work for me.
In the future, it would be nice to have the possibility to retrieve image captures on demand in Unity C# as well if you ever get to it 🙂
Best regards and again thanks for the support
-piro