Data processing: speech signal processing – linguistics – language – Speech signal processing – Application
Reexamination Certificate
1999-03-26
2002-12-31
Dorvil, Richemond (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Application
C704S255000
Reexamination Certificate
active
06502075
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention relates generally to applications for automatic speech recognition (ASR) and, in particular, to automated telephone operator platforms that use ASR for information retrieval, playback and navigation.
2. Description of the Related Art
Conventional voice mail systems have the capability of recording a user's voice for playback and announcement purposes. Thus, for example, when a given called party is not available, the system may be controlled to play a prompt in the called party's actual voice. This operation typically is accomplished by storing a user-recorded audio file and then retrieving that file upon a given occurrence. While the above-described playback technique is desirable, in practice many users do not take the time to record their voice information. Indeed, in some environments, such as the use of such systems in hospitals or hotels, users do not require access to the system for extended periods and, thus, they have no incentive to record their voice information into the system. In these cases, playback of user identifying information can only be accomplished in one of two ways. Either the operator has to undertake the task of recording the user names individually, which may be quite labor intensive, or the system is forced to default to some other means of announcing the user's name. Typically, the default operation is performed by a text-to-speech processor, which provides unacceptable results. In particular, text-to-speech processors cannot provide accurate pronouncements of given names, especially with respect to names of foreign origin.
The present invention addresses this problem.
BRIEF SUMMARY OF THE INVENTION
It is an object of the present invention to provide an automated telephone operator platform (e.g., an auto attendant) that has a natural names library associated therewith. The natural names library comprises a large database of audio files representing common objects (e.g., names, product identifiers, colors, etc.) that are expected to be recognized by a telephone-based ASR application associated with or executing on the platform. Preferably, individual audio files are stored in a compressed format. Whenever the ASR application requires playback of an object file (e.g., the user's name), the system first determines whether the user has recorded the necessary voice sample. If so, the user's actual voice sample is output. If, however, the user's actual voice sample is not available or has not been recorded, the ASR application accesses the database to retreive objects files from which it then builds an appropriate output string.
Thus, for examaple, if a calling party desires to speak with “Mary Smith” and that person is not available, the voice mail system may respond “your call to ‘Mary Smith’ cannot be completed, if you'd like to leave a message, press
1
.” In that message, if the ‘Mary Smith’ prompt does not exist in the user's own voice, the system retrieves ‘Mary’ and ‘Smith’, two separate object entries in the common names library, concatenates the objects, and outputs ‘Mary Smith’. In the event the available library objects are not available or are insufficient, a text-to-speech processor may be used as a back-up announcement mechanism.
The natural names library may be centrally-located and accessed by a set of attendant platforms via a data network such as the Internet, an intranet, or the like. The library may be accessed on a call-by-call basis, or a subset of objects in the library may be downloaded to a given attendant platform. As the number of entries in the database grows larger and larger, centralized management and administration of the libary is preferred.
Thus, according to the preferred embodiment of the invention, individual component objects of a desired output are retreived from the names library and reassembled (typically by concatenation) into an output. The names library has particular utility as a front end or back end to a private branch exchange to provide directory assistance, voice mail routing, and the like.
The foregoing has outlined some of the more pertinent objects and features of the present invention. These objects should be construed to be merely illustrative of some of the more prominent features and applications of the invention. Many other beneficial results can be attained by applying the disclosed invention in a different manner or modifying the invention as will be described. Accordingly, other objects and a fuller understanding of the invention may be had by referring to the following Detailed Description of the Preferred Embodiment.
REFERENCES:
patent: 5210520 (1993-05-01), Housley
patent: 5483579 (1996-01-01), Stoged
patent: 5651055 (1997-07-01), Argade
patent: 5794249 (1998-08-01), Orsolini et al.
patent: 5875448 (1999-02-01), Boys et al.
patent: 5892814 (1999-04-01), Brisebois et al.
patent: 6021181 (2000-02-01), Miner et al.
patent: 6041300 (2000-03-01), Ittycheriah et al.
patent: 6075844 (2000-06-01), Goldberg et al.
patent: 6088428 (2000-07-01), Trandal et al.
patent: 6173266 (2001-01-01), Marx et al.
Foster Peter J.
Terry Kim S.
Dorvil Richemond
Koninklijke Philips Electronics , N.V.
Piotrowski Daniel J.
LandOfFree
Auto attendant having natural names database library does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Auto attendant having natural names database library, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Auto attendant having natural names database library will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2990478