Is Azure AI Vision model good for tracking hands?

So for my internship assignment I plan to make an AI model that checks if you sign certain signs in sign language correctly (specifically dutch sign language.)

Last schoolyear I worked on a project to translate signlanguage to written dutch to imitate a tolk. It was a proof of concept and we used ML5 and Mediapipe. My internship company prefers to use Azure.

Does anyone have any experience with Azure AI Vision for hand tracking motion? How well does it work? Can it see a difference between fingers well, even if they might be obstructed?