A Study on Vision-Based Hand Tracking and Gesture Recognition
| Autor: | Ming-Shan Liu, 劉明山 |
|---|---|
| Rok vydání: | 2009 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 97 Nowadays, most of human computer interface (HCI) use the keyboard or mouse as the command input device. However, gesture is considered one of the most natural ways for human to communicate with other people or to issue commands. Generally, the gesture recognition system is designed based on computer vision to automatically detect/track human hands and also recognize the gesture. In object detection and tracking, this thesis develops a multi-cue similarity measurement algorithm for hand detection, and use an object based method to continuously track the hand. In gesture recognition, due to the fact that the gesture image is highly complex and non-separable, at first the Central Moments analysis is employed to extract the physical features from gesture. Then the Hidden Markov Model (HMM) is used to compare the feature sequences and give suitable actions when a match is found. Experimental results show that the gesture recognition system developed in the thesis has the capability of detecting hands and recognizing the gestures for the distance around 1.7~2 meters that is from a normal family sofa to the camera. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|