A Gesture Recognition Algorithm Based on Threedimensional Projection and Direction Chain Code

Abstract

In this paper, the accelerometer of intelligent terminal is used for gesture recognition. Most of the current motion recognition algorithms in this field have great restrictions on the standardization of motion, and the recognition rate for slightly substandard motion is greatly reduced. Aiming at the existing "$3" algorithm, this paper proposes to project the 3D motion data to the 2D plane for trajectory pattern recognition, and uses the angle and radius to identify the curved area. This method can make up for the people greatly. Physiological habits lead to difficulty in the recognition of gestures, which can greatly improve the recognition accuracy when the movement is not standardized.