Data processing: speech signal processing – linguistics – language – Speech signal processing – Application
Reexamination Certificate
2002-02-05
2003-08-05
Banks-Harold, Marsha D. (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Application
C704S260000, C704S275000, C704S243000
Reexamination Certificate
active
06604077
ABSTRACT:
TECHNICAL FIELD
This invention relates to speech recognition in general and, more particularly, the present invention provides remotely accessible automatic speech recognition and text-to-speech services via a packet network.
BACKGROUND OF THE INVENTION
Techniques for accomplishing automatic speech recognition (ASR) are well known. Among known ASR techniques are those that use grammars. A grammar is a representation of the language or phrases expected to be used or spoken in a given context. In one sense, then, ASR grammars typically constrain the speech recognizer to a vocabulary that is a subset of the universe of potentially-spoken words; and grammars may include sub-grammars. An ASR grammar rule can then be used to represent the set of “phrases” or combinations of words from one or more grammars or subgrammars that may be expected in a given context. “Grammar” may also refer generally to a statistical language model (where a model represents phrases), such as those used in language understanding systems.
Products and services that utilize some form of automatic speech recognition (“ASR”) methodology have been recently introduced commercially. For example, AT&T has developed a grammar-based ASR engine called WATSON that enables development of complex ASR services. Desirable attributes of complex ASR services that would utilize such ASR technology include high accuracy in recognition; robustness to enable recognition where speakers have differing accents or dialects, and/or in the presence of background noise; ability to handle large vocabularies; and natural language understanding. In order to achieve these attributes for complex ASR services, ASR techniques and engines typically require computer-based systems having significant processing capability in order to achieve the desired speech recognition capability. Processing capability as used herein refers to processor speed, memory, disk space, as well as access to application databases. Such requirements have restricted the development of complex ASR services that are available at one's desktop, because the processing requirements exceed the capabilities of most desktop systems, which are typically based on personal computer (PC) technology.
Typically, ASR services include the counterpart Text-to-Speech (“TTS”) engine for generating a response to a user of the service. The TTS engines are known in the art and utilize and data obtained from recognizing a user's speech to formulate and articulate an audible response from the service. In this manner, the user can carry on a conversation with the ASR/TTS server much as the user would with a person.
Packet networks are general-purpose data networks that are well-suited for sending stored data of various types, including speech or audio. The Internet, the largest and most renowned of the existing packet networks, connects over 4 million computers in some 140 countries. The Internet's global and exponential growth is common knowledge today.
Typically, one accesses a packet network, such as the Internet, through a client software program executing on a computer, such as a PC, and so packet networks are inherently client/server oriented. One way of accessing information over a packet network is through use of a Web browser (such as the Netscape Navigator, available from Netscape Communications, Inc., and the Internet Explorer, available from Microsoft Corp.) that enables a client to interact with Web servers. Web servers and the information available therein are typically identified and addressed through a Uniform Resource Locator (URL)-compatible address. URL addressing is widely used in Internet and intranet applications and is well known to those skilled in the art (an “intranet” is a packet network modeled in functionality based upon the Internet and is used, e.g., by companies locally or internally). Many of these networks use the well-known Internet Protocol for communication. Further, in addition to land-line networks, wireless networks also employ packet technologies and use microbrowsers and the like to enable hand-held wireless devices to communicate with wireless web servers.
What is desired is a way of enabling ASR services that may be made available to users at a location, such as at their desktop or from a wireless device, that is remote from the system hosting the ASR engine and TTS engine.
SUMMARY OF THE INVENTION
A system and method of operating an automatic speech recognition service using a client-server architecture is used to make ASR/TTS services accessible at a client location remote from the location of the main ASR and TTS engines. In accordance with the present invention, using client-server communications over a packet network, such as the Internet, the ASR server receives a grammar from the client, receives information representing speech from the client, performs speech recognition, and the TTS server returns information based upon the recognized speech to the client. The ASR server and TTS server may reside on a single computer device or multiple devices within a network. Accordingly, it is immaterial to the present invention regarding the physical location within a network where the ASR and TTS processes operate.
According to an embodiment of the invention, the invention relates to a method of operating an automatic speech recognition service remotely accessible by a client over a packet network. The method comprises receiving from the client over the packet network a grammar identifier associated with one of a plurality of grammars, based on the grammar identifier, selecting a grammar from the plurality of grammars, receiving from the client over the packet network information representing speech, recognizing the received speech information by applying an automatic speech recognition algorithm in accordance with the selected grammar, and sending information based on the recognized speech over the packet network to the client.
The present invention enables a dialogue between a human user and a computer device having an ASR and TTS engine over a packet network. The packet network may be associated with a land-line network or a wireless network. For example, packet networks may provide communication of data between a wireless device and a wireless network node, or between two nodes in a wireless network. Accordingly, the present invention may be applicable to an entire network or a portion of a network employing packet technology.
Alternative embodiments of the present invention include a variety of ways to obtain access to the desired grammar and use of compression or feature extraction as a processing step at the ASR client prior to transferring speech information to the ASR server.
REFERENCES:
patent: 5475792 (1995-12-01), Stanford et al.
patent: 5673362 (1997-09-01), Matsumoto
patent: 5890123 (1999-03-01), Brown et al.
patent: 6078886 (2000-06-01), Dragosh et al.
patent: 6366886 (2002-04-01), Dragosh et al.
Wyard, P J et al.; Spoken language systems-beyond prompt and response; BT Technology Journal; vol. 1, No. 1; Jan. 1996; pp. 187-205.
Dragosh Pamela Leigh
Roe David Bjorn
Sharp Robert Douglas
AT&T Corp.
Azad Abul K.
Banks-Harold Marsha D.
Isaacson Thomas M.
LandOfFree
System and method for providing remote automatic speech... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for providing remote automatic speech..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for providing remote automatic speech... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3082288