Pengenalan Tulisan Tangan Bahasa Arab Menggunakan Metode Probabilistic Neural Network

Abstract

Technological developments are very rapid at this time does not make documentation of data or information with handwriting removed. Even in various studies, writing by hand has proven to have a very good effect on brain intelligence. In this research, handwriting recognition applications will be built using the Probabilistic Neural Network method with training data. Data with each letter consists of 10 samples used for system training. In pattern recognition, training data has a very important role in determining the accuracy of the system, besides the size of the pixel dimensions of rows and columns of image is also very influential. This handwriting image recognition process goes through several stages, namely training data and testing data. In the training data phase there are some data that must be trained before the data is tested. After training the data will be tested and then converted to RGB in the gray direction. The data tested is initially connected so that it can be separated, then the data is cropped. Then the data extraction feature will be set to produce training data and testing data. In this study, the image pixel matrix will be changed in vector form.
Keywords: Handwriting Recognition, Feature Extraction, Probabilistic Neural Network (PNN).