Design of Batak Toba Script Recognition System Using Convolutional Neural Network Algorithm

Indonesia is one of the countries with diversity and abundant cultural wealth, which Batak Toba script as wealth originating from tribe. However, existence decreasing along rapid development times, due to lack interest younger generation public awareness in preserving script. From these problems, author conducted research create a model introducing script, an effort preserve Indonesia's wealth. The purpose this recognition using digital handwriting dataset, has output form visual text audio pronunciation each method used Convolutional Neural Network algorithm combined RMSprop optimizer. that deep learning methods good performance on image data. results study incised relatively high level accuracy, equal 99,54% was tested dataset handwriting. Through research, able produce for recognizing