Aplikasi Android untuk Pengenalan Citra Karakter Jepang dengan Library Tesseract

Abstract

Japanese is one of difficult language to learn because the writing is relatively complicated and it has three kind of writing hiragana, katakana, and kanji. In this study, the Japanese character recognition system based on Android is designed to recognize Japanese character image and translate it into Indonesian using the library Tesseract OCR (Optical Character Recognition). OCR is a technique to convert non digital text into digital text or literally be interpreted as optical character recognition. In this study there were five functional requirements taking a picture with a camera, taking pictures form gallery, the conversion from image to text, text editing of Japanese character form OCR results, and display the results of Japanese text translations. Based on 10 samples of Japanese characters that have been tested, the accuracy values obtained from the test image from the camera around 80% and 94% image from the gallery. Besides test of applications using the questionnaire reached 91%. Thus this application is expected to be able to recognize the image of Japanese characters and translate them into Indonesian according to user needs.