PENGENALAN OBJEK SEDERHANA BERDASARKAN CIRI WARNA DENGAN METODE BACK PROPAGATION

Abstract

ABSTRACT

Human being has extraordinary ability for recognizing objects based on these feature knowledge that has been observed.  Computer is a good, fast, accurate and efficient data processing device.  Computer able to solve the unstructured problem, uncomplete information and qualitative data Artificial Intelligence (AI) problem.

            Pattern recognition is one of AI applications for classifying and identifying pattern or complex object by characteristics measurement or object features (color, shape, textute). A system to identify two object that have almost the same shape but defferent colors is made in this research.  Pattern recognition method is made by one of the methods of Artificial Neural Network that is supervised (Supervised Learning) that is Back Propagation. Example object in this research using apple image and orange image. Calculation result in Backpropagation shows with 14 images colour, 10x10 matriks pixel and random selected weight, so sometimes to get large error.  The resercer define if error more than 0.5 the system is not able to recognize tested images.  To get high accurately simulation data is needed. The result of this research shows that learning procces by Backpropagation, systems recognize different objects based on the color feature.