Implementasi Sistem Pengenalan Karakter Braille Berbasis Android

Muhammad, Muhammad (2023) Implementasi Sistem Pengenalan Karakter Braille Berbasis Android. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 07111940000131-UnderGraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (2MB) | Request a copy

Abstract

Indonesia merupakan salah satu negara dengan penyandang disabilitas terbesar di Dunia.
Menurut estimasi Kementerian Kesehatan RI, jumlah tunanetra di Indonesia adalah 1,5 % dari
seluruh penduduk[9]. Membaca braille merupakan kemampuan untuk membantu
memudahkan peserta didik mempelajari materi pelajaran, dan sebagai salah satu sarana dalam
membantu pembacaan braille dibutuhkan system digital untuk membaca tulisan braille secara
otomatis. Oleh karena itu pada tugas akhir ini peneliti mengimplementasi sistem dengan
machine learning dan computer vision yang direalisasikan pada smartphone android, dimana
pada aplikasi ini tulisa braille akan ditangkap oleh kamera, yang kemudian diproses oleh
pemrosesan citra dan selanjutnya akan dikenali oleh algoritma CNN (convolusional neural
network). Hasil pengenalan kemudian ditampilkan pada layar smartphone berupa teks dan
suara dengan menggunakan google API text to speech algorithm. Pada penelitian tugas akhir
ini aplikasi telah diuji untuk membaca kata dengan batasan 10 huruf dan mendapat rata rata
akurasi 91.5% untuk pengujian sebanyak 10 kali, dengan skenario pengujian menggunakan
citra data latih sebesar 260 citra perhuruf. Sistem ini diharapkan mampu membantu dunia
pendidikan khususnya pada sekolah luar biasa dan orang awam yang ingin mempelajari huruf
braille.=======================================================================================================================================================
Indonesia is one of the countries with the largest population of people with disabilities
in the world. According to the estimation of the Indonesian Ministry of Health, the number of
visually impaired individuals in Indonesia is 1.5% of the total population[9]. Reading Braille
is a skill that helps facilitate students in learning their subjects. As one of the means to aid
Braille reading, a digital system is required for automatic Braille writing recognition.
Therefore, in this final project, the researcher implements a system using machine learning
and computer vision, realized on an Android smartphone. In this application, Braille text is
captured by the camera, then processed through image processing, and subsequently
recognized by a Convolutional Neural Network (CNN) algorithm. The recognition results are
displayed on the smartphone screen as text and voice using Google API text-to-speech
algorithm.In this final research project, the application has been tested to read words with a
maximum of 10 letters, achieving an average accuracy of 91.5% in 10 test runs. The testing
scenario used a training data set consisting of 260 images per letter. The system is expected to
assist the field of education, especially in special schools and for the general public who want
to learn Braille letters.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Braille, Android, Convolusional neural network, Computer vision, Tunanetra Braille, Android, Convolusional neural network, Computer vision, Visually Impaired.
Subjects:	H Social Sciences > HV Social pathology. Social and public welfare > HV1669 Braille Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA76.774.A53 Android
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Muhammad
Date Deposited:	04 Aug 2023 02:07
Last Modified:	04 Aug 2023 02:07
URI:	http://repository.its.ac.id/id/eprint/100204

Actions (login required)

View Item