US 6,185,530 B2Granted

Apparatus and methods for identifying potential acoustic confusibility among words in a speech recognition system

AssigneeInternational Business Machines Corporation

InventorsAbraham Ittycheriah, Stephane H. Maes

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method of determining potential acoustic confusion between at least one new word and at least a portion of existing words of a vocabulary of a speech recognition engine comprises the steps of: a user inputting the at least one new word; computing respective measures between the at least one new word and the at least a portion of existing vocabulary words, the respective measures indicative of acoustic similarity between the at least one word and the at least a portion of existing words; if at least one measure is within a threshold range, indicating results associated with the at least one measure and prompting the user to input an alternative word or additional information pertaining to the at least one new word; and if no measure is within the threshold range, adding the at least one new word to the vocabulary.

Claims29

Dependent claims

A method for adding words to a vocabulary of a speech recognition engine, the method comprising the steps of:

(a) inputting at least one new word;
(b) computing an acoustic similarity measure between the at least one new word and each word of at least a portion of existing vocabulary words, wherein the step of computing an acoustic similarity measure comprises the steps of:
computing a leaf sequence for the at least one new word;
comparing the leaf sequence for the at least one new word to respective leaf sequences associated with the at least a portion of the existing words;
and
generating respective distance measures in response to the comparisons, the respective distance measures indicative of acoustic distances between the compared leaf sequences, wherein the respective distance measures are calculated via a Kuhlback-Liebler distance metric;
(c) if at least one of the computed acoustic similarity measures is within a threshold range, identifying the at least one new word as being potentially acoustically confusing with an existing vocabulary word, indicating results associated with the at least one computed acoustic similarity measure and prompting a user to input an alternative word or additional information pertaining to the at least one new word;
and
(d) if none of the computed acoustic similarity measures are within the threshold range, adding the at least one new word to the vocabulary.

6 more elements

Computer-based apparatus for adding words to a vocabulary of a speech recognition engine, the apparatus comprising:

an input device for inputting at least one new word;
a processor, operatively coupled to the input device, for computing an acoustic similarity measure between the at least one new word and each word of at least a portion of existing vocabulary words;
and
an output device operatively coupled to the processor, wherein the output device comprises a text-to-speech system;
wherein if at least one of the computed acoustic similarity measures is within a threshold range, the processor generates output results associated with the at least one computed acoustic similarity measure to identify the at least one new word as being potentially acoustically confusing with an existing vocabulary word, wherein the processor causes speech synthesis of the results for playback to the user via the text-to-speech system and causes a prompting of the user to input an alternative word or additional information pertaining to the at least one new word;
and wherein if none of the computed acoustic similarity measures are within the threshold range, the processor causes the at least one new word to be added to the vocabulary.

3 more elements

Dependencies are inferred from the claim text.