Pengenalan Huruf Latin Menggunakan Convolution Neural Network Untuk Anak Usia 3-6 Tahun

Hasan, Baharudin (2023) Pengenalan Huruf Latin Menggunakan Convolution Neural Network Untuk Anak Usia 3-6 Tahun. Undergraduate thesis, Politeknik Negeri Jember.

	Text (Abstract) Abstrak.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (6kB)
	Text (BAB 1. Pendahuluan) BAB 1. Pendahuluan.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (14kB)
	Text (Daftar Pustaka) Daftar Pustaka.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (235kB)
	Text (Laporan Lengkap) E41221604_Baharudin Hasan.pdf Restricted to Registered users only Download (3MB) | Request a copy

Abstract

This research uses Convolutional Neural Network (CNN) model training with four different learning rate epoch experiments, namely (0.001, 0.1, 0.3, 0.5). The training results from the best learning rate were 0.1 for 100 epochs with an error of 0.047% and an accuracy of 100%, while the validation testing accuracy results were 82.4%. This application was tested on Alamanda 105 PAUD children with the introduction of vowels in Indonesian and English, and all children succeeded in learning the vowels well. The research results show that the use of CNN in recognizing Latin letters is effective and efficient in learning letters in early childhood. This application also received a positive assessment from PAUD Alamanda 105 teachers in the User Acceptance Testing test with an average score of 92.8%. These results show that this application has the potential to become a learning aid in early childhood education at the Alamanda 105 PAUD school.

Actions (login required)

View Item