Malay spoken word segmentation using magnitude sum function / Fara Ezwana Dardihi

Speech recognition or voice recognition is the identification of spoken words by a machine. The spoken words are digitized by getting the input through a microphone and matched the patterns produced by the speaker against coded database in order to identify the words. In the speech recognition the segmentation of speech is important. Speech segmentation is a method of separating the speech into some isolated sub-words with optimal boundaries. The aim of this research is to apply the segmentation techniques to Malay speeches. In this research, Malay digit speeches were recorded and segmented using magnitude sum function. The segmented speeches can be used on Malay speech recognition on other application that related to speech recognition for example spoken document retrieval system that mainly for indexing continuous Malay speeches and its transcribed document.