Javanese Character Recognition With Real-time Detection Using Convolutional Neural Network

Leksana, Wahyu Purnama Widya and , Fajar Suryawan, S.T., M.Eng.Sc., Ph.D (2023) Javanese Character Recognition With Real-time Detection Using Convolutional Neural Network. Skripsi thesis, Universitas Muhammadiyah Surakarta.

	PDF (Surat Pernyataan Publikasi) Surat Pernyataan Publikasi.pdf Restricted to Repository staff only Download (1MB) | Request a copy
	PDF (Naskah Publikasi) Naskah Publikasi.pdf Download (5MB)

Abstract

Style Javanese letters are a cultural heritage that needs to be preserved so they don't disappear in the future. In this case, many researchers take Javanese letters as their research object. Various methods were developed to create a model for the most optimal recognition of Javanese letters. The convolutional neural network model was chosen by researchers because of its well-known ability to classify images into the desired classification class. 3240 Javanese letter images were collected and classified manually into 20 labels to be used as data sets. This model is then burned using the data set to produce a model that is ready to be used to identify new Javanese letters obtained from real-time image capture using a webcam. This convolutional neural network model achieves 99.17% accuracy within 7 epochs. The designed system successfully recognize the javanese character at a speed of 1 frame for 0.13 seconds. The trial using the image from the webcam into the convolutional neural network model obtained results that were able to identify 20 javanese characters with the highest being ga character with 96.67% accuracy and the lowest being wa character with 56.24%.

Item Type:	Thesis (Skripsi)
Uncontrolled Keywords:	Javanese Character, Convolutional Neural Network, Tensorflow, Real-time
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science > QA751 Artificial Intelligence
Divisions:	Fakultas Ilmu Komunikasi dan Informatika > Teknik Informatika
Depositing User:	WAHYU PURNAMA WIDYA LEKSANA
Date Deposited:	06 Feb 2023 01:53
Last Modified:	06 Feb 2023 01:53
URI:	http://eprints.ums.ac.id/id/eprint/107853

Actions (login required)

View Item