ETRI-Knowledge Sharing Plaform

KOREAN
논문 검색
Type SCI
Year ~ Keyword

Detail

Conference Paper Effective Voice Data Processing Techniques for Speech Recognition in Client/Server Environment
Cited - time in scopus Share share facebook twitter linkedin kakaostory
Authors
MiKyoung Kim, JungHee Choi, HyunSoon Shin, SangHo Lee
Issue Date
2015-01
Citation
International Conference on Small and Medium Business (ICSMB) 2015, pp.284-287
Publisher
중소기업융합학회
Language
English
Type
Conference Paper
Abstract
In this paper, we propose a processing technique for performing audio data communication effectively in a client/server environment by reducing the amount of audio data transmitting for speech recognition. In general, the speech recognition system using HMM algorithm uses 2-dimensional 8-bytes array data with float type, yet it has a lot of overhead in a client/server. In order to make up for the problem, we use 1-dimensional 4-bytes array data with integer type using vector quantization and DHMM algorithms.
KSP Keywords
Array data, Audio data, Data Processing Techniques, Data transmitting, Speech recognition system, Vector Quantization(VQ), Voice Data, data communication