Core Concepts
Ring-a-Pose, a single untethered ring, can continuously track 3D hand poses using active acoustic sensing.
Abstract
The paper presents Ring-a-Pose, a single untethered ring that can continuously track 3D hand poses. Key highlights:
Ring-a-Pose uses active acoustic sensing, where the ring emits inaudible sound waves that are reflected by the hand and captured by the ring's microphone. The reflection patterns are then analyzed to reconstruct the hand pose.
The authors conducted three user studies to evaluate Ring-a-Pose's performance:
Worn Finger Comparison: Ring-a-Pose performs similarly well when worn on different fingers, allowing users to choose their preferred finger.
Continuous Hand Pose Tracking: Ring-a-Pose achieves a mean per-joint position error of 14.1 mm in user-independent evaluation and 10.3 mm in user-dependent evaluation.
Micro-Finger Gesture Recognition: Ring-a-Pose recognizes 7 micro-finger gestures with 90.60% and 99.27% accuracy in user-independent and user-dependent evaluations, respectively.
Ring-a-Pose is low-cost (around $30), low-profile, and low-power (148 mW), making it suitable for everyday use.
Unlike prior multi-ring systems, Ring-a-Pose can track continuous hand poses with just a single untethered ring, providing a practical solution for hand interaction.
Stats
Mean per-joint position error of 14.1 mm in user-independent evaluation
Mean per-joint position error of 10.3 mm in user-dependent evaluation
90.60% accuracy in 7-class micro-finger gesture recognition (user-independent)
99.27% accuracy in 7-class micro-finger gesture recognition (user-dependent)
Quotes
"Ring-a-Pose is the first single-ring system to demonstrate continuous hand pose tracking."
"Ring-a-Pose exhibits exciting performance even when used out-of-the-box, i.e. no calibration data from the new user, further highlighting the practicality of ring-based hand pose tracking systems."