Solution that integrates voice enrollment with other types of recognition operations performed by a speech recognition engine using a layered grammar stack
First Claim
1. A computer system for enabling use of a single voice recognition engine for both command recognition and user speech enrollment, said computer system comprising:
- a user interface to receive speech input, the speech input comprising at least a portion of a phrase desired to be added to a voice-enrolled grammar;
at least one processor; and
a computer-readable medium encoded with at least one computer program for performing enrollment and recognition by applying speech input to a turn-based speech recognition engine in one turn that, when executed by the at least one processor, controls the computer system to;
process the speech input during a speech recognition turn beginning at a first layer of an ordered stack of grammars,wherein the stack comprises a higher layer, the higher layer comprising an application grammar,a lower layer, the lower layer comprising the enrollment grammar; and
a layer between the higher layer and the lower layer comprising a confusable grammar layer, andeach layer in the stack includes an exit criterion;
wherein the processing in one turn is progressively performed from the higher layer in the stack toward the lower layer in the stack, until a given layer in the stack is reached at which the exit criterion is met;
in response to determining that the exit criterion for the given layer is satisfied, end the speech recognition turn, return a speech recognition result based upon the given layer, and ignore any lower layers of the ordered stack,wherein the processing comprises conditionally adding the at least a portion of the phrase to the voice-enrolled grammar based on the processing completing for the lower layer of the stack without satisfying an exit criterion for any layer of the stack for which processing was performed.
2 Assignments
0 Petitions
Accused Products
Abstract
The present invention can include a speech enrollment system including an ordered stack of grammars and a recognition engine. The ordered stack of grammars can include an application grammars layer, a confusable grammar layer, a personal grammar layer, a phrase enrolled grammar layer, and an enrollment grammar layer. The recognition engine can return recognition results for speech input by processing the input using the ordered stack of grammars. The processing can occur from the topmost layer in the stack to the bottommost layer in the stack. Each layer in the stack can includes exit criteria based upon a defined condition. When the exit criteria is satisfied, a result can be returned based upon that layer and lower layers of the ordered stack can be ignored.
57 Citations
21 Claims
-
1. A computer system for enabling use of a single voice recognition engine for both command recognition and user speech enrollment, said computer system comprising:
-
a user interface to receive speech input, the speech input comprising at least a portion of a phrase desired to be added to a voice-enrolled grammar; at least one processor; and a computer-readable medium encoded with at least one computer program for performing enrollment and recognition by applying speech input to a turn-based speech recognition engine in one turn that, when executed by the at least one processor, controls the computer system to; process the speech input during a speech recognition turn beginning at a first layer of an ordered stack of grammars, wherein the stack comprises a higher layer, the higher layer comprising an application grammar, a lower layer, the lower layer comprising the enrollment grammar; and a layer between the higher layer and the lower layer comprising a confusable grammar layer, and each layer in the stack includes an exit criterion; wherein the processing in one turn is progressively performed from the higher layer in the stack toward the lower layer in the stack, until a given layer in the stack is reached at which the exit criterion is met; in response to determining that the exit criterion for the given layer is satisfied, end the speech recognition turn, return a speech recognition result based upon the given layer, and ignore any lower layers of the ordered stack, wherein the processing comprises conditionally adding the at least a portion of the phrase to the voice-enrolled grammar based on the processing completing for the lower layer of the stack without satisfying an exit criterion for any layer of the stack for which processing was performed. - View Dependent Claims (2, 3, 4, 5)
-
-
6. A method for creating voice-enrolled grammars comprising:
-
receiving speech input, the speech input comprising at least a portion of a phrase desired to be added to a voice-enrolled grammar; using entries in an application grammar and at least one processor to determine whether the speech input matches an entry in the application grammar with a sufficient confidence; when a sufficient confidence is determined, returning a result that indicates a recognition match; when an insufficient confidence is determined, using entries in at least one of an application grammar and a personal grammar to determine whether the speech input matches an entry in the application grammar with a sufficient clash value; when a sufficient clash value is determined, returning a result that indicates a clash with an existing grammar entry; when an insufficient clash value is determined, performing a voice enrollment consistency detection operation for determining acoustic quality; and if the voice enrollment consistency detection operation indicates a sufficient level of consistency, voice enrolling the speech input by adding the at least a portion of the phrase to the voice-enrolled grammar, thereby rejecting the speech input if the speech input is recognized by the application grammar, or if the speech input is confusable with an entry in either the application grammar or the user grammar, or if the acoustic quality of the speech input is inadequate for matching, wherein the above steps are performed by a turn based speech recognition engine in a single turn; and wherein the method utilizes an ordered stack of grammars consisting of a higher layer, a lower layer, and a layer between the higher layer and the lower layer, wherein the higher layer comprises the application grammar, the lower layer comprises the voice-enrolled grammar, and the layer between the higher layer and the lower layer comprises a confusable grammar layer, wherein the layers of the ordered stack are processed in order from the topmost layer in the stack to the bottommost layer in the stack, wherein each layer in the stack includes an exit criterion based upon a defined condition, such that when the exit criterion is satisfied, a result is returned based upon that layer, and lower layers of the ordered stack are ignored. - View Dependent Claims (7, 8, 9)
-
-
10. A method of utilizing a layered grammar stack to integrate voice enrollment operations with other types of recognition operations of a speech recognition engine comprising:
-
establishing an ordered stack of grammars comprising a higher layer, a lower layer, and a layer between the higher layer and the lower layer, each layer including an exit criterion, wherein the higher layer in the stack includes at least one application grammar, the lower layer in the stack includes an enrollment grammar, and the layer between the higher layer and the lower layer includes a confusable grammar layer, the enrollment grammar comprising a mapping of speech inputs to user-defined phrases, the enrollment grammar for allowing a user to define a set of speech recognition commands, the application grammar comprising a mapping of speech inputs to a plurality of application-specific voice commands; receiving speech input, the speech input comprising at least a portion of a phrase desired to be added to a voice-enrolled grammar; and processing the speech input during a speech recognition turn beginning at a first layer of the ordered stack of grammars, wherein the processing in one turn is progressively performed from the topmost layer in the stack to the bottommost layer in the stack, until a given layer in the stack is reached at which the exit criterion is met; in response to determining that the exit criterion is satisfied, ending the speech recognition turn, returning a speech recognition result based upon the given layer, and ignoring lower layers of the ordered stack, wherein the processing comprises conditionally adding the at least a portion of the phrase to the voice-enrolled grammar based on the processing completing for the lower layer of the stack without satisfying an exit criterion for any layer of the stack for which processing was performed. - View Dependent Claims (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21)
-
Specification