Talking Head System In Indonesian Language With Affective Facial Expression Synthesis

Setyati, Endang (2017) Talking Head System In Indonesian Language With Affective Facial Expression Synthesis. Doctoral thesis, Institut Teknologi Sepuluh Nopember.

Text 2211301005-Disertation.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Dalam membuat persepsi dan tampilan dari sebuah antarmuka manusia-komputer alami (Natural Human-Computer Interface) dibutuhkan integrasi dari audio yang realistis dan informasi visual. Penelitian ini mengusulkan Talking Head System dalam bahasa Indonesia (Talking Head System in Indonesian Language) dengan sintesis dari ekspresi wajah berafeksi, berdasarkan 12 model viseme Indonesia. Model Viseme Indonesia ini didapatkan dari hasil pemetaan 49 fonem (termasuk silent) ke viseme Indonesia.
Sistem ini mengkonversikan input teks dari kalimat bahasa Indonesia yang berafeksi ke ucapan (speech) dengan sinkronisasi animasi dari bentuk mulut, gerakan bibir dan ekspresi emosional. Talking Head juga didasarkan pada sebuah model kepala manusia 3D generik. Untuk mengekspresikan visualisasi 3D wajah berafeksi dan pembangkitan urutan viseme (viseme chain generation) yang alami digunakan Inesh2 Corpora, Trigram Hidden Markov Model dan Algoritma Viterbi.
Sedangkan untuk mendapatkan 3D model ekspresi wajah berafeksi didasarkan pada pengklasifikasian parameter afeksi/emosi dengan menggunakan facial motion capture. Akhir dari sistem ini, diharapkan dapat menghasilkan visualisasi 3D animasi dengan ucapan kalimat bahasa Indonesia berafeksi.

======================================================================================

Natural Human-Computer Interface requires integration of realistic audio and visual information for perception and display. In this research, Talking Head System in Indonesian Language is proposed to affective facial expressions synthesis, based on 12 models viseme Indonesia. Model Viseme Indonesia is obtained from the mapping of 49 phonemes (including silent) to viseme Indonesia. The system converts text from affective Indonesian sentences to specch with synchronized animation of mouth shape, lip movements and emotion expressions. The Talking Head is based on generic 3D human head model. To express emotion and viseme chain generation using Inesh2 Corpora, Trigram Hidden Markov Model and Viterbi Algorithm.
To get a 3D model of facial expressions based on classification parameters berafeksi affections / emotions using facial motion capture. The final system will produce animated 3D visualization of speech of sentences Indonesian with affection.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	Talking Head System in Indonesian Language; Viseme Chain Generation; Inesh2 Corpora; Trigram Hidden Markov Model; Algoritma Viterbi
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms
Divisions:	Faculty of Industrial Technology > Electrical Engineering > 20001-(S3) PhD Thesis
Depositing User:	Anis Wulandari
Date Deposited:	15 May 2017 08:02
Last Modified:	27 Dec 2017 01:50
URI:	http://repository.its.ac.id/id/eprint/41268

Actions (login required)

View Item