Pujiani, Ni Komang Intan Tri and Miftahuddin, Yusup (2022) Sistem Automatic Speech Recognition Menggunakan PCA dan VQ Untuk Deteksi Kemiripan Kata Bahasa Sunda. In: Prosiding Diseminasi Fakultas Teknik Industri Semester Ganjil 2021/2022, Bandung.

Text Sistem Automatic Speech Recognition Menggunakan PCA dan VQ Untuk Deteksi Kemiripan Kata Bahasa Sunda.pdf Download (2MB)

Abstract

Voice Recognition technology can be used to recognize what someone is saying. In this research technology will be used in speech recognition to detect when someone is pronouncing a word incorrectly due to a high degree of similarity. Therefore, we need a system capable of recognizing speech with the help of technology (speech recognition). The system detects the voice by taking input in *.wav format, extracting the voice using the Principal Component Analysis (PCA) approach, and then identifying it using the Vector Quantization (VQ) method. Testing was performed with codebook sizes of 32, 64, 128, 256, 512 dan 1024 in this study. The test was run in 8 sets of Sundanese switch a high degree of similarity. Codebook usage accuracy may vary depending on test results. Using code size 128 had the highest average accuracy rate for each word pair, 79.8%. Keywords: Voice Recognition, Principal Component Analysis, Vector Quantization

Actions (login required)

View Item