TUD Logo

TUD Home » ... » Results of student work » Minor Theses » Mario Höpfner

Computer Graphics

Imagebased Handtracking For Intuitive Navigation In 3D Scenes

Chair for Computer Graphics and Visualization


Student: Mario Höpfner
Supervisor: Sören König
Responsible Professor: Prof. Dr. Stefan Gumhold

Motivation

With the permanently growing amount at computer-aided applications, also its complexity in input and output grows rapidly. Hand gesture recognition gives the possibility to humans of using its most flexible tools - the hand - effective and in natural way.

Description

The existing work presents a complete implementation of a gesture control system. Hand recognition and tracking are realized using image processing algorithms.

Results

After manual initialisation the implemented algorithm is able to track a marked human hand and also to recognize three gestures. For this reason it is possible to control four different 3D demo scenes.

input image
input image
labeled target area
labeled target area
reliability
reliability
qquantized HSX-Image
quantized HSX-Image
probability image
probability image
binary image
binary image
contour
contour
parameters
parameters
Kombination beider Ansichten
combination of both views
Geste Faust
gesture fist
Geste Spreizen
gesture finger spread
Geste Handwippe
gesture seesaw
Puzzle
Puzzle
Flying
Flying
Image Viewer
Image Viewer
Earth Viewer
Earth Viewer

Download

Last modified: 1st Feb 2010, 2.39 PM
Author: Corina Weissbach