PENGENALAN BENTUK TANGAN DENGAN METODE PCA DAN NFL SEBAGAI PENGGANTI MOUSE KOMPUTER

Abstract

Humans who always do the daily activities with computers, need some ways and means to be able to interact (give inputs and receive outputs) with the computers. This program is made so the people can interact with the computers, using their hand gestures instead of a computer mouse. The program uses the hand image as an input. And then the program uses Principal Component Analysis and Nearest Feature Line methods to recognize user’s hand gestures. The testing of the program is carried out with four scenarios. The first and second scenario with Database of the hand images of one man, show hand gesture recognition success rate of 92% and 76%, respectively. Then the follow-up testing is carried out with two more additional scenarios. The third and fourth scenario is hand gesture recognition testing with Database of the hand images of two people combined (balanced composition).