Text

Segmentasi Continuous Speech dengan Menggunakan Local Adaptive Thresholding dalam Metode Improved Blocking Block Area

---

ABSTRAK
Continuous speech merupakan bentuk ucapan alamiah manusia yang bersifat kontinu tanpa
adanya batas yang jelas antar kata. Melalui ucapan tersebut memungkinkan manusia dapat
memberikan perintah berupa suara ke komputer. Akan tetapi, agar komputer dapat
mengenalinya diperlukan pengenalan continuous speech. Pada pengenalan tersebut
dibutuhkan proses segmentasi continuous speech untuk memotong kalimat pada batas tiap
kata. Segmentasi menjadi tahapan penting karena ucapan akan dikenali dari segmentsegment kata yang dihasilkan proses ini. Proses segmentasi pada penelitian ini dilakukan
dengan teknik local adaptive thresholding serta metode Improved Blocking Block Area.
Penelitian ini bertujuan melakukan perbandingan kinerja untuk lima metode local adaptive
thresholding (Niblack, Sauvola, Bradley, Guanglei Xiong dan Bernsen) dalam melakukan
segmentasi continuous speech sehingga diperoleh metode terbaik dan nilai parameter
optimum. Berdasarkan hasil penelitian diperoleh metode Niblack sebagai metode terbaik
dalam melakukan segmentasi continuous speech Bahasa Indonesia dengan akurasi mencapai
95% serta nilai parameter optimum metode tersebut adalah window=75 dan k=0.2.
Kata Kunci: Continuous Speech, Pengenalan Continuous Speech, Segmentasi Continuous
Speech, Local Adaptive Thresholding, Improved Blocking Block Area

ABSTRACT
Continuous speech is a form of natural human speech that is continuous without clear
boundaries between words. Through speech, it allows humans to give voice commands to
computers. However, the computer can recognize it requires continuous speech recognition.
On recognition, a continuous speech segmentation process is needed to cut the sentence at
the boundary of each word. Segmentation becomes an important step because speech will
be recognized by the word segments produced by this process. The segmentation process in
this study was carried out using local adaptive thresholding techniques and the Improved
Blocking Block Area method. This study aims to conduct performance comparisons for five
local adaptive thresholding methods (Niblack, Sauvola, Bradley, Guanglei Xiong and
Bernsen) in continuous speech segmentation so that the best method and optimum parameter
values are obtained. Based on the results of the study obtained the Niblack method is the
best method of segmenting Bahasa Indonesia continuous speech with accuracy reaching 95%
and the optimum parameter value for the method is window=75 and k=0.2.
Key Word: Continuous Speech, Continuous Speech Recognition, Continuous Speech
Segmentation, Local Adaptive Thresholding, Improved Blocking Block Area

---

Ketersediaan

748F19III

748 F 19-ii

Perpustakaan FSM Undip (Referensi)

Tersedia

Informasi Detail

Judul Seri

ILMU KOMPUTER/INFORMATIKA

No. Panggil

748 F 19-ii

Penerbit

: ., 2019

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

530.44

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

-

Info Detail Spesifik

-

Pernyataan Tanggungjawab

Roihan Auliya Ulfattah

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar