Our Hand Tracking and Gesture Recognition capabilities include a range of multi-dimensional and multi-granularity gesture functions such as detection and tracking of one or both hands, left and right hand recognition, 2D/3D hand keypoints recognition, static and dynamic gesture recognition, and hand mesh reconstruction.
Support Qualcomm, MTK, and other mainstream computing platforms, computing portfolio, and TEE
1. Supports all kinds of equipment;
2. Supports 17 types of static/dynamic gestures by default and can expand to more types of gestures
3. High availability;
4. Low latency.
Supports monocular and multiocular camera equipment, as well as gray scale data and RGB data streams
Detects and tracks gestures within the range of 20 centimeters to 3 meters
Recognizes both one-handed and two-handed gestures, such as finger hearts and cupping one hand over a clenched fist in a bow of respect
Support for retrieving lost gestures,and maintains gesture and pose stability when hands are invisible
Real-time effects can be achieved on mobile devices