IMPLEMENTASI HIDDEN MARKOV MODEL UNTUK APLIKASI PENGENALAN SUARA DAN UCAPAN SEBAGAI SISTEM PENGAMANAN UNTUK PERANGKAT KOMPUTER/LAPTOP MENGGUNAKAN LINEAR PREDICTIVE CODING (LPC)

Nasrullah Nasrullah
Ernawati Ernawati
Endina Putri Purwandari

Jurnal Teknologi Informasi Cyberku
Vol 11 No 1 (2015): Jurnal Teknologi Informasi CyberKU Vol.11 no 1

The security of computer devices is very necessary to keep the data in it can be securely maintained. In this research, security application for windows login process that uses user's voice and speech is built. The method used is Hidden Markov Model (HMM) method to identify matches user speech that is to say words that have meaning. Later it also used Linear Predictive Coding (LPC) as a feature extraction to recognize the user voice characteristic who uttered these words. Applications that built with both methods is a desktop application for the training and testing of user's voice and speech and the startup application for login process on computer devices. In this research, LPC methods implementation produces the same speech recognition to a different user by 100%, speech recognition that is different from the same user by 100%. Whereas HMM method application generates voice recognition based on the number of training by 84%, voice recognition based on the number of words spoken by 100%. For voice and speech recognition with a different tone of voice by the same user with average as many as 94%. This application is built with Delphi 7 programming language.