I am trying to reconstruct what I had in mind when I first put this this down.
Vision-based tracking (e.g. Jinkun's work) may be in image space and not the world frame so we may need additional components for localization.
At the same time, we have localization for the existing person tracker, and the naive association issue (#54) would build on that in a separate thread.
I am trying to reconstruct what I had in mind when I first put this this down.
Vision-based tracking (e.g. Jinkun's work) may be in image space and not the world frame so we may need additional components for localization.
At the same time, we have localization for the existing person tracker, and the naive association issue (#54) would build on that in a separate thread.