Head-Movement Gesture Recognition Using 3-Axis Sensors and Image Processing
| Autor: | Sheng-Feng Li, 李昇峰 |
|---|---|
| Rok vydání: | 2013 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 101 This study aims to establish a system which can recognize the gestures of the head movement. Two real-time and effective methods are presented herein to identify the gestures of the head movement. The first method is to utilize the 3-axis sensor with an accelerometer, a gyroscope and a magnetometer mounted on user’s head for respectively measuring the gravity acceleration, angular velocity and magnetic flux to estimate the inclination and orientation angles of the head gesture. Finally, Kalman filter is employed to eliminate the Gauss noise. On the basis of inclination and orientation angles prior to filtering, the angle form integration of the angular velocity is calibrated to be the gesture angle as derivation gestures of the head movement that is shown with 95% accuracy in recognition rate. Moreover, the other method is the face detection capturing image information on user’s face via a camera including the relative positions of the face, eyes, lips as well as mandible and then filtering the noise via image processing as the reference points. Hence, the gesture and orientation of user’s head movement can be recognized based on the relationships between these reference points. The recognition result is shown on the screen that it still has 90% accuracy of recognition rate. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|