Dynamic Threshold Gesture Recognition Apparatus

The patent applies dynamics by transitioning from a fixed threshold to a dynamic threshold that automatically adapts based on observed hand movement characteristics. The threshold updating unit continuously adjusts the threshold value according to the actual moving distances detected in the video stream, enabling the system to adapt to different users and gestures while maintaining low device complexity.

The patent changes the parameter of the threshold from a static value to a dynamically updated value based on statistical analysis of hand movement distances. By calculating the average moving distance and using it to update the threshold, the system achieves adaptive detection precision without increasing operational complexity.

The patent applies preliminary action by using the threshold updating unit to learn and adapt to the user's hand movement characteristics before making recognition decisions. The system accumulates data about normal hand movements during operation and adjusts the threshold accordingly, allowing it to distinguish between intentional gestures and preliminary actions based on learned patterns rather than fixed criteria.

The patent implements feedback by continuously updating the threshold based on observed hand movement distances. The threshold updating unit uses the average moving distance as feedback to adjust the threshold dynamically, creating a closed-loop system that improves recognition reliability while preserving information about different types of hand movements through adaptive learning.