Method and system for network-based speech recognition

Details Method and system for network-based speech recognition Method and system for network-based speech recognition

1999-10-04

2002-09-17

Banks-Harold, Marsha D. (Department: 2741)

Data processing: speech signal processing, linguistics, language

Speech signal processing

Recognition

C704S270100, C434S185000

Reexamination Certificate

active

06453290

ABSTRACT:

FIELD OF THE INVENTION
The present invention pertains to the field of speech recognition, including, more particularly, speech recognition in real-time over the internet.
BACKGROUND
In known systems and methods employing speech recognition, the speech recognition is performed entirely on a user's processing device, or client. As shown in
FIG. 1
, in known environments, a client
10
for a user
20
contains and processes all the applications required for speech recognition activities. These applications include an audio input application
30
for retrieving audio information from an audio device, e.g., a microphone, a speech recognition engine
40
, for processing the input audio speech information and making determinations of what was spoken, and a specific speech recognition application program for coordinating one or more speech recognition activities
50
.
In these known systems and methods, the audio input application
30
, the speech recognition engine
40
, and the speech recognition application program
50
must all be installed on the client
10
. The speech recognition engine installation itself is generally very large and complex, and, thus, cannot be supported by many clients
10
. Further, many users
20
do not want to be bothered with such a large, difficult installation, and, thus, will simply forgo speech recognition activities rather than initiate and maintain the necessary speech recognition engine installation.
Further, in these known systems and methods, the user
20
is thereafter limited to the speech recognition application program
50
installed on his or her client
10
. An update or new speech recognition application program will require a new installation on every client
10
using the speech recognition application program
50
. This program update or replacement can be troublesome, time consuming, and expensive, causing many users
20
to choose to forgo use of speech recognition activities, ultimately resulting in the loss of goodwill and business for these applications.
Thus, it is desirable to have a system and method supporting speech recognition where clients are exposed to minimal, uncomplicated installations. It is also desirable to have a system and method for speech recognition over the internet, so that anyone with a computing device, or client, and access to an internet connection, may have access to speech recognition activities and processing. It is further advantageous to have a system and method for speech recognition that performs in real-time, or approximate real-time, in order that feed-back is reported to users without detectable delays.
SUMMARY OF THE INVENTION
The invention comprises a system and method for speech recognition processing in approximate real-time, over the internet.
A system for supporting speech recognition processing, e.g., for language learning activities, comprises one or more clients and a server. Each client of the system comprises the capability to input audio speech from a user, and store the audio speech into one or more buffers. Each client also comprises the capability to encode, or otherwise compress, the buffers of received audio speech into a smaller data representation of the original input audio speech. Each client can encode a buffer of a portion of the inputted audio speech before all of the user's audio speech is input to the client. Each client can also package an encoded buffer of audio speech into one or more packets for transmission over the internet, and thereafter transmit the packets to the server, for speech recognition processing.
The server comprises the capability to receive packets of encoded audio speech from one or more clients at a time. The server can decode each of the received audio speech packets as they are received, and store the resultant original, or raw, audio speech into one or more buffers associated with the client transmitting the packets of encoded audio speech. The server evaluates the resultant original audio speech from each of the clients, and thereafter transmits a respective feedback response to each of the clients, to be provided to their user.
Other and further objects, features, aspects and advantages of the present invention are found in the following detailed description of the embodiments and the accompanying drawings.


REFERENCES:
patent: 5737531 (1998-04-01), Ehley
patent: 5960399 (1999-09-01), Barclay et al.
patent: 6055498 (2000-04-01), Neumeyer et al.
patent: 6199076 (2001-03-01), Logan et al.
patent: 6219339 (2001-04-01), Doshi et al.
patent: 6301258 (2001-10-01), Katseff et al.
patent: 6397253 (2002-05-01), Quinlan et al.

Affiliated with

Also associated with

Rating

Method and system for network-based speech recognition does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Method and system for network-based speech recognition, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for network-based speech recognition will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2898006