Implementasi OCR dengan Metode Autoencoder pada Aplikasi Bukutamu berbasis WEB

Abstract

A guestbook is a tool to record the identities of visitors who come to a place or event such as weddings, birthday celebrations, parties etc. Not just recording, the guestbook also functions as evidence and traces to avoid something unwanted. So it is not surprising that in some places it is required to submit an identity card such as a KTP when filling out a guestbook to be allowed to enter the place. KTP itself is an identity card that contains data such as name, place, date of birth etc. The data contained in the KTP can be utilized in the guestbook filling process so that officers only need to take a picture of the KTP and visitor data will be filled in automatically with the help of optical character recognition (OCR). To get a good OCR result, an image with clear words is needed, its position is not tilted and its size is not too small.  Therefore, various preprocessing steps are needed before performing the OCR process, one of which is by applying denoise using the Autoencoder which has succeeded in making the image cleaner and the OCR results become more accurate.