Some content of this application is unavailable at the moment.
If this situation persist, please contact us atFeedback&Contact
1. (WO2001033552) METHOD AND MEANS FOR A ROBUST FEATURE EXTRACTION FOR SPEECH RECOGNITION
Latest bibliographic data on file with the International Bureau   

Pub. No.: WO/2001/033552 International Application No.: PCT/EP2000/009607
Publication Date: 10.05.2001 International Filing Date: 30.09.2000
Chapter 2 Demand Filed: 22.05.2001
IPC:
G10L 15/02 (2006.01) ,G10L 15/20 (2006.01) ,G10L 21/0208 (2013.01)
G PHYSICS
10
MUSICAL INSTRUMENTS; ACOUSTICS
L
SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15
Speech recognition
02
Feature extraction for speech recognition; Selection of recognition unit
G PHYSICS
10
MUSICAL INSTRUMENTS; ACOUSTICS
L
SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15
Speech recognition
20
Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise or of stress induced speech
G PHYSICS
10
MUSICAL INSTRUMENTS; ACOUSTICS
L
SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
21
Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
02
Speech enhancement, e.g. noise reduction or echo cancellation
0208
Noise filtering
Applicants:
TELEFONAKTIEBOLAGET LM ERICSSON (publ) [SE/SE]; S-126 25 Stockholm, SE
Inventors:
BRÜCKNER, Raymond; DE
HIRSCH, Hans-Günter; DE
KLISCH, Rainer; DE
SPRINGER, Volker; DE
Agent:
TONSCHEIDT, Andreas; Ericsson Eurolab Deutschland GmbH Ericsson Allee 1 52134 Herzogenrath, DE
GÜNTER Schmalz; Ericsson Eurolab Deutschland GmbH Nordostpark 12 D-90411 Nümberg, DE
Priority Data:
99203613.729.10.1999EP
Title (EN) METHOD AND MEANS FOR A ROBUST FEATURE EXTRACTION FOR SPEECH RECOGNITION
(FR) PROCEDE ET MOYENS D'EXTRACTION DE CARACTERISTIQUES ROBUSTES POUR RECONNAISSANCE VOCALE
Abstract:
(EN) The present invention relates to a method and an apparatus for a robust feature extraction for speech recognition in a noisy environment, wherein the speech signal is segmented and is characterized by spectral components. The speech signal is splitted into a number of short term spectral components in L subbands, with L = 1,2,... and a noise spectrum from segments that only contain noise is estimated. Then a spectral subtraction of the estimated noise spectrum from the corresponding short term spectrum is performed and a probability for each short-term spectrum component to contain noise is calculated. Finally these spectral components of each short-term spectrum, having a low probability to contain speech are interpolated in order to smooth those short-term spectra that only contain noise. With the interpolation the spectral components containing noise are interpolated by reliable spectral speech components that could be found in the neighborhood.
(FR) L'invention concerne un procédé et un appareil d'extraction de caractéristiques robustes pour la reconnaissance vocale dans un milieu bruyant. Le signal vocal, segmenté, se caractérise par des composants spectraux. Le signal vocal est divisé en plusieurs composants spectraux à court terme dans L sous-bandes, avec L = 1, 2, ... et à partir des segments qui ne contiennent que du bruit on estime un spectre de bruit. Ensuite, on effectue, à partir du spectre à court terme correspondant, la soustraction spectrale du spectre du bruit estimé et on calcule une probabilité pour chaque composant de spectre à court terme. Enfin, ces composants de spectre à court terme, dont la probabilité de contenir de la parole est faible, sont interpolés afin de lisser les spectres à court terme ne contenant que du bruit. Avec l'interpolation, les composants spectraux contenant du bruit sont interpolés par des composants spectraux vocaux fiables se trouvant éventuellement à proximité.
front page image
Designated States: AE, AG, AL, AM, AT, AU, AZ, BA, BB, BG, BR, BY, BZ, CA, CH, CN, CR, CU, CZ, DE, DK, DM, DZ, EE, ES, FI, GB, GD, GE, GH, GM, HR, HU, ID, IL, IN, IS, JP, KE, KG, KP, KR, KZ, LC, LK, LR, LS, LT, LU, LV, MA, MD, MG, MK, MN, MW, MX, MZ, NO, NZ, PL, PT, RO, RU, SD, SE, SG, SI, SK, SL, TJ, TM, TR, TT, TZ, UA, UG, UZ, VN, YU, ZA, ZW
African Regional Intellectual Property Organization (ARIPO) (GH, GM, KE, LS, MW, MZ, SD, SL, SZ, TZ, UG, ZW)
Eurasian Patent Office (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM)
European Patent Office (EPO) (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE)
African Intellectual Property Organization (BF, BJ, CF, CG, CI, CM, GA, GN, GW, ML, MR, NE, SN, TD, TG)
Publication Language: English (EN)
Filing Language: English (EN)