<![CDATA[Abstract - Hand gesture interaction is the most relevant approach when interacting in augmented reality. This approach is natural, simple, inexpensive and accurate. The research has three goals that are: to apply the concept of epipolar geometry for geometric of stereo vision, to calculate accuracy of the projective reconstruction method and to make a model for hand gesture interaction in augmented reality. Hand gesture interaction in augmented reality is affected by the accuracy of feature point detection. In this experiment, the percentage of sum of difference from feature point detection is about 3% while the percentage of sum of difference from feature point reconstruction is about 93%. Based on this experiment, the most accurate interaction is obtained from the combination of the sub-pixel feature point detection and the optimal triangulation algorithm. The interaction can be run in real-time by 2.6 fps.  Kata Kunci: hand gesture interaction, projective reconstruction, augmented reality  ]]>
Copyrights © 2011
