Implementing mobile interface based voice recognition system

Abstract

Recently, as supply of smart phone is widely spreading, various voice applications for user's convenience are under development. However, since Google Android-based smart phone delivered by Korean manufacturer processes voice recognition through Google server, it has a weakness to take long time to be processed and need activation of internet. This paper implemented Android-based voice recognition system using continuous HMM without having to use Google server. As the result of evaluation for proposed voice recognition system and Google's voice recognition system, the proposed system showed similar voice recognition performance but its processing speed is proved to be better than Google's. © 2012 Springer-Verlag.