DocumentCode
189853
Title
Pure software-based speech recognition for OS-less embedded systems
Author
Milinkovic, Stevan
Author_Institution
Union University, School of Computing Knez Mihailova 6, 11000 Belgrade, Serbia
fYear
2014
fDate
15-19 June 2014
Firstpage
40
Lastpage
43
Abstract
This paper discusses the implementation of a pure software-based, low cost embedded speech recognition system. The applied algorithms are memory efficient, scalable and provide sufficient accuracy of recognition. Advantages of such a system include rapid prototyping, flexibility in design modifications, and being a library, ease of integration with other applications.
Keywords
Accuracy; Acoustics; Feature extraction; Hardware; Speech; Speech processing; Speech recognition; DTW; LPC; embedded speech recognition;
fLanguage
English
Publisher
ieee
Conference_Titel
Embedded Computing (MECO), 2014 3rd Mediterranean Conference on
Conference_Location
Budva, Montenegro
Print_ISBN
978-1-4799-4827-7
Type
conf
DOI
10.1109/MECO.2014.6862730
Filename
6862730
Link To Document