XR Gesture State Machine Using Gaze to Reject Unintentional Input

The patent extracts only the necessary components for hand pose detection: standard joint position tracking and gaze information. By removing specialized computer vision algorithms and focusing on spatial relationships between hand and head, the system achieves sufficient accuracy without the complexity of advanced vision systems.

The patent uses universally available tracking data (joint positions and gaze information) that can be obtained from standard sensors. This multi-functional approach allows the same data to serve both pose detection and gesture intention determination, eliminating the need for specialized algorithms.

The patent segments gesture detection into distinct criteria: hand pose determination and gaze criterion satisfaction. Each criterion can be evaluated independently and in parallel, reducing processing time while maintaining comprehensive reliability through multiple checks.

The system preliminarily determines hand pose using simple spatial relationships before applying the gaze criterion. This staged approach allows quick rejection of obvious non-gestures while reserving more complex analysis for cases where the hand pose is ambiguous or potentially intentional.