Method and apparatus using decision trees to generate and score multiple pronunciations for a spelled word

(19)

(11)

EP 0 953 970 A3

(12)	EUROPEAN PATENT APPLICATION

(88)	Date of publication A3:
	19.01.2000 Bulletin 2000/03

(43)	Date of publication A2:
	03.11.1999 Bulletin 1999/44

(21)	Application number: 99303390.1

(22)	Date of filing: 29.04.1999

(51)	International Patent Classification (IPC)⁷: G10L 5/04

(84)	Designated Contracting States:
	AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE
	Designated Extension States:
	AL LT LV MK RO SI

(30)

Priority:

29.04.1998 US 67764
29.04.1998 US 69308
30.04.1998 US 70300

(71)	Applicant: MATSUSHITA ELECTRIC INDUSTRIAL CO., LTD.
	Kadoma-shi, Osaka 571-8501 (JP)

(72)	Inventors:
	Kuhn, Roland Santa Barbara, California 93110 (US) Junqua, Jean-Claude Santa Barbara, California 93111 (US) Contolini, Matteo Santa Barbara, California 93109 (US)

(74)	Representative: Senior, Alan Murray
	J.A. KEMP & CO., 14 South Square, Gray's Inn London WC1R 5LX London WC1R 5LX (GB)

(54)	Method and apparatus using decision trees to generate and score multiple pronunciations for a spelled word

(57) The mixed decision tree includes a network of yes-no questions about adjacent letters in a spelled word sequence and also about adjacent phonemes in the phoneme sequence corresponding to the spelled word sequence. Leaf nodes of the mixed decision tree provide information about which phonetic transcriptions are most probable. Using the mixed trees, scores are developed for each of a plurality of possible pronunciations, and these scores can be used to select the best pronunciation as well as to rank pronunciations in order of probability. The pronunciations generated by the system can be used in speech synthesis and speech recognition applications as well as lexicography applications.

Search report