Illllllllllllllllllllllllllllllllllllllllllllllllll
US006687673B2
(12) United States Patent (io) Patent No.: US 6,687,673 B2
Mann (45) Date of Patent: Feb. 3,2004
(54) SPEECH RECOGNITION SYSTEM
(75) Inventor: Robert James Mann, Romsey (GB)
(73) Assignee: International Business Machines
Corporation, Armonk, NY (US)
( * ) Notice: Subject to any disclaimer, the term ol this
patent is extended or adjusted under 35
U.S.C. 154(b) by 0 days. GB
(21) Appl. No.: 10/369,780
(22) Filed: Feb. 19, 2003
(65) Prior Publication Data
US 2003/0167166 Al Sep. 4, 2003
Related U.S. Application Data
(63) Continuation of application No. 09/552,847, filed on Apr. 20, 2000, now Pat. No. 6,629,071.
(30) Foreign Application Priority Data
Sep. 4, 1999 (GB) 9920844
(51) Int. CI.7 G10L 15/22; G10L 15/06
(52) U.S. CI 704/244; 704/254
(58) Field of Search 704/243, 244,
704/251, 252, 254, 255, 270, 275
(56) References Cited
U.S. PATENT DOCUMENTS
5,027,406 A * 6/1991 Roberts et al 704/244
5,651,056 A * 7/1997 Eting et al.
5,890,117 A * 3/1999 Silverman 704/260
A method ol performing speech recognition to determine a particular alphabetic character can include receiving acoustic spoken input including a single alphabetic character and a word association with the single character such that the first character ol the word is intended to be the same as the single alphabetic character. The acoustic input can be processed using a speech vocabulary recognition system to recognize the single alphabetic character and the word. An attempt to match the single alphabetic character with the first character ol the word can be made and, responsive to an unsuccessful attempt at matching the single character with the first character of the word, information relating to the unsuccessful attempt can be stored into a history of unsuccessful attempts. The stored history information can be used to determine the alphabetic character for the spoken input.
3 Claims, 3 Drawing Sheets
