APLIKASI PENGENALAN SUARA PEMBICARA MENGGUNAKAN HIDDEN MARKOV MODEL (HMM)

Perdana, Rezdy Anugrah (2014) APLIKASI PENGENALAN SUARA PEMBICARA MENGGUNAKAN HIDDEN MARKOV MODEL (HMM). S1 thesis, Universitas Pendidikan Indonesia.

[img]
Preview
Text
S_KOM_0608616_Title.pdf

Download (182kB) | Preview
[img]
Preview
Text
S_KOM_0608616_Abstract.pdf

Download (111kB) | Preview
[img]
Preview
Text
S_KOM_0608616_Table_of_content.pdf

Download (203kB) | Preview
[img]
Preview
Text
S_KOM_0608616_Chapter1.pdf

Download (151kB) | Preview
[img] Text
S_KOM_0608616_Chapter2.pdf
Restricted to Staf Perpustakaan

Download (752kB)
[img]
Preview
Text
S_KOM_0608616_Chapter3.pdf

Download (177kB) | Preview
[img] Text
S_KOM_0608616_Chapter4.pdf
Restricted to Staf Perpustakaan

Download (889kB)
[img]
Preview
Text
S_KOM_0608616_Chapter5.pdf

Download (201kB) | Preview
[img]
Preview
Text
S_KOM_0608616_Bibliography.pdf

Download (188kB) | Preview
Official URL: http://repository.upi.edu

Abstract

Sistem Speaker Recognition merupakan sebuah sistem yang dapat mengenali pembicara hanya melalui suaranya. Terdapat dua jenis speaker recognition, yaitu speaker identification dan speaker verification. Speaker recognition digunakan untuk mengidentifikasi siapa yang sedang berbicara sedangkan speaker verification digunakan untuk mengesahkan apakah orang yang sedang berbicara merupakan target speaker. Dalam skripsi ini digunakan metode ekstraksi ciri Mel-Frequency Cepstral Coefficients (MFCC) dan pengenalan pola menggunakan Hidden Markov Models. Eksperimen dilakukan dengan 8 orang responden yang mengucapkan kata “Pendidikan” sebanyak 20 kali secara berturut-turut. 10 data ucapan digunakan sebagai data training sedangkan 10 sisanya digunakan sebagai data uji. Hasil eksperimen tersebut menunjukkan tingkat akurasi dari sistem adalah sebesar 71.25%. Kata Kunci: Speaker Recognition, Speaker Verification, Speaker Identification, MFCC, Hidden Markov Models Speaker Recognition System is a system that can recognize people only through his voice. There are two types of speaker recognition, the speaker identification and speaker verification. Speaker recognition is used to identify who is speaking while the speaker verification is used to validate whether the person who is speaking is the target speaker. This study used Mel-Frequency Cepstral Coefficients (MFCC) feature extraction method and pattern recognition using Hidden Markov Models. Experiments performed with 8 respondents who say the word "Pendidikan" as many as 20 times in a row. 10 are used as training data, while the remaining 10 are used as test data. The experimental results demonstrate the accuracy of the system is equal to 71.25%. Keywords: Speaker Recognition, Speaker Verification, Speaker Identification, MFCC, Hidden Markov Models

Item Type: Skripsi,Tesis,Disertasi (S1)
Uncontrolled Keywords: Pengenalan suara,hidden markov model
Subjects: Q Science > Q Science (General)
Divisions: Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Program Studi Pendidikan Ilmu Komputer
Depositing User: Riki N Library ICT
Date Deposited: 19 Aug 2014 03:08
Last Modified: 19 Aug 2014 03:08
URI: http://repository.upi.edu/id/eprint/7352

Actions (login required)

View Item View Item