Looks like Snap (where many of the authors of this paper work at) are heavily investing (or already invested) in R&D for advancing AR and pose estimation with papers like this.
Not the first time I've seen them publish these papers but I think we'll see more from them which will tell us about their next features or future products.
Not the first time I've seen them publish these papers but I think we'll see more from them which will tell us about their next features or future products.