Perancangan pembelajaran sistem isyarat bahasa Indonesia menggunakan YOLOv8 berbasis web

Lahitani, Anissa Tri (2025) Perancangan pembelajaran sistem isyarat bahasa Indonesia menggunakan YOLOv8 berbasis web. Sarjana thesis, UIN Sunan Gunung Djati Bandung.

Preview	Text (COVER) 1_cover.pdf Download (134kB) | Preview
Preview	Text (ABSTRAK) 2_abstrak.pdf Download (298kB) | Preview
Preview	Text (SK BEBAS PLAGIARISM) 3_skbebasplagiarism.pdf Download (554kB) | Preview
Preview	Text (DAFTAR ISI) 4_daftarisi.pdf Download (288kB) | Preview
Preview	Text (BAB I) 5_bab1.pdf Download (566kB) | Preview
	Text (BAB II) 6_bab2.pdf Restricted to Registered users only Download (884kB) | Request a copy
	Text (BAB III) 7_bab3.pdf Restricted to Registered users only Download (2MB) | Request a copy
	Text (BAB IV) 8_bab4.pdf Restricted to Registered users only Download (2MB) | Request a copy
	Text (BAB V) 9_bab5.pdf Restricted to Registered users only Download (246kB) | Request a copy
	Text (DAFTAR PUSTAKA) 10_daftarpustaka.pdf Restricted to Registered users only Download (259kB) | Request a copy
	Text (LAMPIRAN) 11_lampiran.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

INDONESIA: Bahasa isyarat merupakan sarana komunikasi utama bagi penyandang tunarungu dan tunawicara, salah satunya Sistem Isyarat Bahasa Indonesia (SIBI) yang diakui secara resmi di Indonesia. Meskipun teknologi computer vision dan deep learning telah berkembang pesat, pemanfaatannya untuk media pembelajaran bahasa isyarat masih sangat terbatas. Penelitian ini menggunakan metode Agile SCRUM untuk pengembangan aplikasi web dan Cross-Industry Standard Process for Data Mining (CRISP-DM) untuk membangun model machine learning. Dataset citra gestur huruf dan angka SIBI dikumpulkan, diproses, dan dilatih menggunakan model YOLOv8-cls. Model kemudian diintegrasikan ke sistem berbasis NestJS (backend), React.js (frontend), dan MongoDB (basis data). Solusi yang diusulkan adalah sistem pembelajaran SIBI berbasis web dengan fitur video pembelajaran, latihan gestur interaktif menggunakan kamera, kuis evaluasi, dan papan peringkat, yang mampu memberikan umpan balik langsung kepada pengguna berdasarkan hasil klasifikasi gestur. Hasil pengujian menunjukkan kinerja model yang tinggi, dengan precision dan recall sebesar 93,08%, F1-Score 93,06%, dan Top-1 Accuracy 90,27%. Temuan ini membuktikan bahwa integrasi teknologi computer vision dan deep learning menggunakan YOLOv8 dapat menghasilkan media pembelajaran bahasa isyarat yang efektif, inklusif, dan mudah diakses secara daring. ENGLISH: Sign language is the primary means of communication for deaf and mute individuals, one of which is the Indonesian Sign Language System (SIBI) officially recognized in Indonesia. Although computer vision and deep learning technologies have developed rapidly, their use for sign language learning media is still very limited. This study uses the Agile SCRUM method for web application development and the Cross-Industry Standard Process for Data Mining (CRISP-DM) to build a machine learning model. A SIBI letter and number gesture image dataset is collected, processed, and drilled using the YOLOv8-cls model. The model is then integrated into a NestJS (backend), React.js (frontend), and MongoDB (database)-based system. The proposed solution is a web-based SIBI learning system with learning video features, interactive gesture exercises using a camera, evaluation quizzes, and a leaderboard, which is able to provide direct feedback to users based on gesture classification results. Test results demonstrated high model performance, with a precision and recall of 93.08%, an F1-Score of 93.06%, and a Top-1 Accuracy of 90.27%. These findings demonstrate that the integration of computer vision and deep learning technologies using YOLOv8 can produce effective, inclusive, and accessible sign language learning media.

Actions (login required)

View Item