CD Skripsi

Pengenalan Alfabet Sibi Menggunakan Convolutional Neural Network Sebagai Media Pembelajaran Bagi Masyarakat Umum

---

ABSTRACT
SIBI is a Sign Language developed by hearing people and adapted from American Sign
Language (ASL). SIBI has been established by the government through the Decree of
the Minister of Education and Culture of the Republic of Indonesia Number
0161/U/1994 and is used in SLB in Indonesia. The difficulty of communication between
deaf and hard of hearing people and the general public causes limited social
interaction between people with disabilities and the general public. The initial step that
will be taken for the introduction of Sign Language to the community by creating a
learning tool as an introduction to learning Sign Language by introducing the SIBI
alphabet. One method that can be developed to bridge these problems is Deep
Learning. Deep Learning is a branch of Machine Learning that has excellent
performance. The most commonly used class of Deep Learning algorithms for spatial
pattern recognition analysis is Convolutional Neural Networks. This study used CNN
to train the SIBI alphabet image and managed to get the best accuracy of 87.62% for
training accuracy and 92.07% validation accuracy. Testing using confusion matrix got
an accuracy of 92.9%. The architecture that gets the best accuracy in this study
consists of four Convolution Layer, four Pooling Layer and Fully Connecter Layer.
Based on this research, it is concluded that the model successfully recognizes the SBI
alphabet image well and gets the best accuracy in Phase V trials at the 50th training.
Keywords: Sign Language, SIBI, CNN

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

1907111410

Penerbit

Pekanbaru : Universitas Riau – Fakultas Teknik – Teknik Informatika., 2023

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

1907111410

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

TEKNIK ELEKTRO

Info Detail Spesifik

-

Pernyataan Tanggungjawab

DAUS

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar