Pengenalan Pola Aksara Jawa Tulisan Tangan dengan Jaringan Syaraf Tiruan Perambatan-Balik

Abstract

Automatic reading techniques of handwriting being developed. Computerization of Java script has started to be developed. However, there is no automated reader system for documents written in handwriting Java script. This research applies backpropagation neural network for recignition of handwritten java script. Preprosesing patterns done with Java script normalize the image size to be 40x40 pixels then feature extraction. Performance network training is 99.8% and performance examination is 95.81%.