In this paper we discuss speech recognition related algorithms that are suitable for implementation on embedded devices. We will provide a system that can perform speaker dependant speech recognition, via a dictionary search. The dictionary is prefilled with phrases that need to be recognized. We will describe the theoretical foundation of the used algorithms, as well as provide a practical implementation description for all relevant system components.

Keywords: DTW, Embedded, LPC, MFCC, Препознавање говора
Published on website: 22.12.2013