SISTEM PENDETEKSIAN WAJAH PADA VIDEO MENGGUNAKAN JARINGAN ADAPTIVE LINEAR NEURON (ADALINE)

Fadlisyah, Fadlisyah (2013) SISTEM PENDETEKSIAN WAJAH PADA VIDEO MENGGUNAKAN JARINGAN ADAPTIVE LINEAR NEURON (ADALINE). Masters thesis, Universitas Sumatera Utara.

Text - Thesis.doc Download (959kB)

Abstract

ABSTRACT Facial detection is an important initial stage which is needed by various types of application of computer vision, biometric system, and various types of multimedia application such as facial tracking, facial recognition, and video surveillance. In this research, a rapid system of facial detection was proposed and analyzed by using Adaptive Linear Neuron (ADALINE). The system was established from conditioning image by applying sobel operator, and then applying Widrow-Hoff learning method. The working method of the system was evaluated, based on two parameters of working method: detection rate and false positive rate. The result of the evaluation showed that the performance of the system had detection rate from about 65% to 90%. In order to increase the quality of the work method of the system for the future researches, the established system can be combined with one of the following: the test of video complexity, segmentation of skin color, and statistical test. Keywords: facial detection, Adaptive Linear Neuron, Widrow-Hoff Learning

Actions (login required)

View Item