-
公开(公告)号：JP2001109490A
公开(公告)日：2001-04-20
申请号：JP2000265326
申请日：2000-09-01
Applicant: CSELT CENTRO STUDI LAB TELECOM
Inventor： FRASCA ALESSANDRA , MICCA GIORGIO , PALME ENRICO
Abstract: PROBLEM TO BE SOLVED: To provide a voice recognition device for language in a condition that an associated voice training material cannot substantially be obtained. SOLUTION: A first step 1 and 2 is to generate a multiple language voice recognition device 2 with respect to plural known languages based on an associated voice training material 1. Then, a recognition device 5 with respect to the given language is constituted by interpolation 3 from the multiple language recognition device 2. The recognition device 5 which is thus generated can substantially be improved based on the associated voice training material 4 collected by online during subsequent use.
-
公开(公告)号：CA2317231A1
公开(公告)日：2001-03-02
申请号：CA2317231
申请日：2000-09-01
Applicant: CSELT CENTRO STUDI LAB TELECOM
Inventor： PALME ENRICO , FRASCA ALESSANDRA , MICCA GIORGIO
Abstract: A process for speech recognition. To implement a speech recognizer for a language in conditions of substantial unavailability of related speech training material the first step is, based on related speech training material, a multilingual speech recognizer for a plurality of known languages. The recognizer for such given language is then implemented by interpolation starting from the said multilingual recognizer. The recognizer generated in this fashion is susceptible of being subsequently refined based on related speech training material acquire d online during later use.
-
公开(公告)号：CA2317231C
公开(公告)日：2008-01-22
申请号：CA2317231
申请日：2000-09-01
Applicant: CSELT CENTRO STUDI LAB TELECOM
Inventor： FRASCA ALESSANDRA , MICCA GIORGIO , PALME ENRICO
Abstract: A process for speech recognition. To implement a speech recognizer for a language in conditions of substantial unavailability of related speech training material the first step is, based on related speech training material, a multilingual speech recognizer for a plurality of known languages. The recognizer for such given language is then implemented by interpolation starting from the said multilingual recognizer. The recognizer generated in this fashion is susceptible of being subsequently refined based on related speech training material acquire d online during later use.
-
公开(公告)号：IT1310154B1
公开(公告)日：2002-02-11
申请号：ITTO990734
申请日：1999-09-02
Applicant: CSELT CENTRO STUDI LAB TELECOM
Inventor： FRASCA ALESSANDRA , MICCA GIORGIO , PALME ENRICO
-
-
-
Search Results
4
records
(took 0.022 seconds)
