Implementasi CNN-LSTM dalam Sistem Pengenalan Bahasa Isyarat Indonesia Berbasis Suara

Hidayat, Enryco (2025) Implementasi CNN-LSTM dalam Sistem Pengenalan Bahasa Isyarat Indonesia Berbasis Suara. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 21081010023.-cover.pdf Download (1MB)
	Text (Bab 1) 21081010023.-bab1.pdf Download (91kB)
	Text (Bab 2) 21081010023.-bab2.pdf Restricted to Repository staff only until 8 December 2027. Download (446kB)
	Text (Bab 3) 21081010023.-bab3.pdf Restricted to Repository staff only until 8 December 2027. Download (1MB)
	Text (Bab 4) 21081010023.-bab4.pdf Restricted to Repository staff only until 8 December 2027. Download (3MB)
	Text (Bab 5) 21081010023.-bab5.pdf Download (78kB)
	Text (Daftar pustaka) 21081010023.-daftarpustaka.pdf Download (87kB)
	Text (Lampiran) 21081010023.-lampiran.pdf Restricted to Repository staff only Download (59kB)

Abstract

This research aims to develop a voice-to-text and sign language translation system using the Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) architecture. The system is designed to facilitate communication between hearing-impaired individuals and the general public by automatically converting voice input into real-time sign language representations. Audio data were recorded in 8-second WAV format, processed through normalization, Mel-Spectrogram feature extraction, and augmentation before being trained with the CNN–LSTM model. The training achieved 99.99% accuracy and 100% validation accuracy, with precision, recall, and f1-score all reaching 1.00 across all classes. The model was implemented using a Flask backend and a simple web interface that displays the prediction results in both text and sign language images. Based on testing with seven sample labels, the system achieved confidence scores between 90%–100% and similarity scores between 99%–100%, demonstrating strong generalization and stability. These results confirm that integrating CNN–LSTM with Speech Recognition and Similarity Score produces a highly accurate, efficient, and real-time web-based voice-to-sign language translation system.

Item Type:	Thesis (Undergraduate)
Contributors:	Contribution Contributors NIDN/NIDK Email Thesis advisor Idhom, Mohammad NIDN0010038305 idhom@upnjatim.ac.id Thesis advisor Nurlaili, Afina Lina NIDN0013129303 afina.lina.if@upnjatim.ac.id
Subjects:	Q Science > QA Mathematics > QA76.6 Computer Programming
Divisions:	Faculty of Computer Science > Departemen of Informatics
Depositing User:	Enryco Hidayat
Date Deposited:	08 Dec 2025 04:46
Last Modified:	08 Dec 2025 04:46
URI:	https://repository.upnjatim.ac.id/id/eprint/48208

Actions (login required)

View Item