While there are many challenges with the current AR top manufacturers, a startup formed by ex-Microsoft engineers already has gestures and precise tracking of the whole body.
Under the Perceptive IO system, the team produced hand geometry and an algorithm that tracks gestures at 1000 frames per second for fast-tracking of bargain hunters. The system also detects high-speed motion tracking of two hands.
The team at SIGGRAPH Asia 2017 showcased the “close-to-quick tracking of two-hand interaction” near-field research algorithms. Through the latest video posted above, you can see how fast it’s recognized.
Compared with other tracking systems (such as Leap Motion), Perceptive IO can efficiently improve tracking accuracy, as well as the interaction between the two hands. That will not only develop the existing gesture interface but will even become the way for the next generation of AR devices to interact.
Also, the soft pair has also developed a Motion2fusion volumetric motion tracking system that “rebuilds any non-rigid scene in real time.” With machine learning, Motion2fusion avoids over-smoothing and visual artifacts. It is reported that the technology can be applied to mainframes, PC games or holographic conferencing systems.
It is not clear whether the company is still open because its domain name has expired and its Twitter account is no longer available. However, to see this technology is also pleased one thing, I believe the next detail will be announced.
Source