000146121 001__ 146121
000146121 005__ 20181203035844.0
000146121 037__ $$aREP_WORK
000146121 245__ $$aNeural Network based Regression for Robust Overlapping Speech Recognition using Microphone Arrays
000146121 269__ $$a2008
000146121 260__ $$bIDIAP$$c2008
000146121 336__ $$aReports
000146121 500__ $$aSubmitted for publication
000146121 520__ $$aThis paper investigates a neural network based acoustic feature mapping to extract robust features for automatic speech recognition (ASR) of overlapping speech. In our preliminary studies, we trained neural networks to learn the mapping from log mel filter bank energies (MFBEs) extracted from the distant microphone recordings, including multiple overlapping speakers, to log MFBEs extracted from the clean speech signal. In this paper, we explore the mapping of higher order mel-filterbank cepstral coefficients (MFCC) to lower order coefficients. We also investigate the mapping of features from both target and interfering distant sound sources to the clean target features. This is achieved by using the microphone array to extract features from both the direction of the target and interfering sound sources. We demonstrate the effectiveness of the proposed approach through extensive evaluations on the MONC corpus, which includes both non-overlapping single speaker and overlapping multi-speaker conditions.
000146121 700__ $$0242359$$aLi, Weifeng$$g188567
000146121 700__ $$0243992$$aDines, John$$g192380
000146121 700__ $$0243959$$aMagimai.-Doss, Mathew$$g127186
000146121 700__ $$0243348$$aBourlard, Hervé$$g117014
000146121 8564_ $$uhttp://publications.idiap.ch/downloads/reports/2008/li-idiap-rr-08-09.pdf$$zURL
000146121 8564_ $$s188691$$uhttps://infoscience.epfl.ch/record/146121/files/li-idiap-rr-08-09.pdf$$zn/a
000146121 909C0 $$0252189$$pLIDIAP$$xU10381
000146121 909CO $$ooai:infoscience.tind.io:146121$$pSTI$$preport$$qGLOBAL_SET
000146121 937__ $$aLIDIAP-REPORT-2008-010
000146121 970__ $$ali:rr08-09/LIDIAP
000146121 973__ $$aEPFL$$sPUBLISHED
000146121 980__ $$aREPORT