PERANCANGAN PENGENAL KATA DALAM AKSARA SUNDA MENGGUNAKAN METODE DETEKSI TEPI DAN LVQ BERBASIS PENGOLAHAN CITRA PADA ANDROID

DEA DELIA LESTARI

Informasi Umum

Kode

15.04.1443

Klasifikasi

C -

Jenis

Karya Ilmiah - Skripsi (S1) - Reference

Subjek

Engineering - Signal Processing

Informasi Lainnya

Abstraksi

Each area has a distinctive writing or script where the script is different from one region to another. But with a growing era, little by little, a lot of people forget about writing their respective regions. Not a few are also people who do not know any of the characters. Sundanese script has been declared as protected by original letter and required to preserve continuity. In 1996 the first level local government of western Java has enacted a law that regulation no.6 on the preservation, development, and the development of language, literature and script Sunda.

Seeing these conditions, then this final project designed an application that can facilitate the introduction of Sundanese script. In addition to ease of learning, the learning Sundanese script through the application expected Sundanese script can continue to be preserved. This application was made to identify the characters Sunda using Eclipse and the Android OS software.

From the testing results of the 10 words in Sundanese script without cropping background image obtained accuracy of the system's most valued 60.90% with the number of training data is 9 pieces. While the image with cropping the background obtained the highest accuracy rate of 61.53% with a value of learning rate is 0.01 and epoch 100. And for the detection of 30 words in Sundanese script obtained the value of the highest accuracy of 17.78% with the same value of learning rate and epoch.