Kyouiku Kanji Grade 1 Recognition Using MobileNet V2 Based on Android

Character recognition has become a popular research topic in the fieldof pattern recognition and machine learning, including handwritingrecognition, specifically kanji handwriting. This study performshandwriting recognition of kyouiku kanji grade 1, which is the kanjirequired to be learnt by grade 1 elementary school students in Japan.This research uses ETL-9B dataset from Electrotechnical Laboratory(now AIST), uses CNN MobileNet V2 deep learning method that hasbeen customized for mobile devices, and uses Android application asthe user interface implementation. Based on the study results, thehighest accuracy model was obtained with an accuracy of 96,6875% anda size of 27.4MB for the alpha 1.0 hyperparameter. It can be concludedthat the CNN MobileNet V2 deep learning method has performed quitewell in the process of recognizing handwritten kyouiku kanji grade 1