spatialcomputing.info

Hi all,

I'm new to gestural and AR work.

I'm trying to make an interactive system where the user acts out gestures detected by the SR305 Realsense. It's meant to be a prototype of/simulate the ability to control a hologram in the user's environment. Of course, it won't be an actual hologram, it'll just be a fake "hologram" responding to gestures with the user's environment as the backdrop. I'd like it to act like AR in the sense of collision/plane detection, etc with the surrounding environment.

Instead of the output being the RealSense camera's feed (facing the user), I want the output to come from a (second?) camera that I guess can process depth / AR - like things. This camera would need to facing opposite the user so their gestures are directionally correct.

Is this possible? I normally use TouchDesigner for interactive work, but I'm not sure it's the best choice for AR. Would it involve pre-loading / scanning assets in (which is fine, just takes more prep time for each individual test), or is this only possible in something like Unity or Houdini even? It doesn't need to be a perfect process, hack-y is fine since this is more a concept test than anything.

Thanks!