A Real Time Human Motion Recognition Using Stick Figure Based on Motion Library

Autor:	Agusta wicaksono
Druh dokumentu:	學位論文 ; thesis
Popis:	99 Computer systems are increasingly being used for Human motion behavior. But some of them don’t use stick figure in comparison. This research presents a novel method for real time detection of human motion based on motion library. In the first step, we do background subtraction to get silhouette image of human body. In the second step, the silhouette image of human body is processed into stick figure by using thinning algorithm. In the third step, motion analysis is used in comparison between input motion and motion library. With considering some feature points in both vector motions, we present motion human motion recognition using stick figure that gives fast computation speed for real-­‐time application.
Databáze:	Networked Digital Library of Theses & Dissertations
Externí odkaz:	http://ndltd.ncl.edu.tw/handle/40069499698377168533 Zobrazit plný text záznamu