dc.contributor.author | So, S | |
dc.contributor.author | Paliwal, KK | |
dc.contributor.editor | Luís Oliveira | |
dc.date.accessioned | 2017-05-03T13:01:32Z | |
dc.date.available | 2017-05-03T13:01:32Z | |
dc.date.issued | 2005 | |
dc.date.modified | 2007-03-21T21:26:49Z | |
dc.identifier.refuri | http://www.interspeech2005.org/ | |
dc.identifier.uri | http://hdl.handle.net/10072/2673 | |
dc.description.abstract | In this paper, we examine a coding scheme for quantising feature vectors in a distributed speech recognition environment that is more robust to noise. It consists of a vector quantiser that operates on the logarithmic filterbank energies (LFBEs). Through the use of a perceptually-weighted Euclidean distance measure, which emphasises the LFBEs that represent the spectral peaks, the vector quantiser codebook provides /emph{a priori} knowledge of the spectral characteristics of clean speech and is used to quantise features from noise-corrupted speech. Our comparative results from the ETSI Aurora-2 recognition task show that the perceptually-weighted vector quantisation of LFBEs achieves higher recognition accuracies for noisy speech than the unweighted vector quantisation, memoryless and multi-frame GMM-based block quantisation and scalar quantisation of Mel frequency-warped cepstral coefficients. | |
dc.description.peerreviewed | Yes | |
dc.description.publicationstatus | Yes | |
dc.language | English | |
dc.language.iso | eng | |
dc.publisher | International Speech Communication Association (ISCA) | |
dc.publisher.place | Lisbon, Portugal | |
dc.relation.ispartofstudentpublication | N | |
dc.relation.ispartofconferencename | 9th European Conference on Speech Communication and Technology | |
dc.relation.ispartofconferencetitle | 9th European Conference on Speech Communication and Technology | |
dc.relation.ispartofdatefrom | 2005-09-04 | |
dc.relation.ispartofdateto | 2005-09-08 | |
dc.relation.ispartoflocation | Lisbon, Portugal | |
dc.relation.ispartofpagefrom | 941 | |
dc.relation.ispartofpageto | 944 | |
dc.rights.retention | Y | |
dc.subject.fieldofresearchcode | 280204 | |
dc.subject.fieldofresearchcode | 280206 | |
dc.title | Improved noise-robustness in distributed speech recognition via perceptually-weighted vector quantisation of filterbank energies | |
dc.type | Conference output | |
dc.type.description | E1 - Conferences | |
dc.type.code | E - Conference Publications | |
gro.faculty | Griffith Sciences, Griffith School of Engineering | |
gro.date.issued | 2005 | |
gro.hasfulltext | No Full Text | |
gro.griffith.author | Paliwal, Kuldip K. | |
gro.griffith.author | So, Stephen | |