Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
2000-05-26
2004-08-10
Chawan, Vijay (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S260000, C379S088010, C379S088220, C379S093240
Reexamination Certificate
active
06775651
ABSTRACT:
CROSS REFERENCE TO RELATED APPLICATIONS
(Not Applicable)
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
(Not Applicable)
BACKGROUND OF THE INVENTION
1. Technical Field
This invention relates to the field of speech recognition software, and more particularly, to a system for transcribing telephone answering machine voice mail messages.
2. Description of the Related Art
An ever growing number of people have begun using personal computers as a source of voice mail services. By alleviating the need and expense of a separate telephone answering machine, the use of a personal computer to record voice mail messages over a telephone line allows consumers to save money. This trend is likely to continue due to the impressive amount of computing power presently available to consumers in affordable multimedia personal computers. Moreover, the components enabling personal computers to provide voice mail services, such as sound cards and modems, have become standard equipment on most high-speed multimedia personal computers.
Beyond the savings afforded to consumers, computer voice mail systems provide consumers with increased flexibility over their telephone answering machine counterparts. The increasing speed and storage capacity of personal computers enables these machines to record longer messages and store far more messages than conventional telephone answering machines. Moreover, voice mail messages left by callers can be recorded in any of a variety of standardized multimedia or audio file types such as Wave or MP3 files. Such digital files can be manipulated, copied, stored, or transmitted.
Despite the many advantages of using a personal computer for voice mail, however, there exist disadvantages. One such disadvantage is that although the storage capacity of modern personal computers may seem limitless, audio files themselves can be quite large. Thus, storing or archiving old voice mail messages may consume far more storage capacity than what is available in any particular personal computer. In a network context, where storage capacity is often obtained at a premium, the large size of voice mail audio files can become even more problematic.
Another disadvantage inherent to computer based voice mail systems is that the large size of audio files can hinder rapid transmission of the files over networks and can cause network congestion. Such congestion often results in decreased network performance or even a network service outage. Further contributing to the problem is that compression of a Wave or MP3 file typically does not result in a substantial enough reduction of the file size. Thus, a detailed voice mail message of three to four minutes in length, saved as a Wave or MP3 file, can take much longer than the three or four minute playing time to upload and transmit the audio file via a conventional 28.8 kpbs modem connection.
Another disadvantage, inherent to all voice mail systems, is that undoubtedly an occasion will arise in which the user would find a textual transcription of the voice mail message convenient. Such is the case when a voice mail contains directions to a location. Whether the voice mail message containing the directions is left on a conventional telephone answering machine or on a computer based voice mail system, the user must transcribe the voice mail message manually to obtain an accurate transcription of the voice mail message.
Another known technology, referred to as speech recognition, is the process by which an acoustic signal received by microphone is converted to a set of text words by a computer. These recognized words may then be used in a variety of computer software applications for purposes such as document preparation, data entry, and command and control. Recently, speech recognition has been applied to recording technology. Specifically, voice recorders have been designed to record audio input which subsequently can be supplied to a speech recognition engine for conversion to text. Still, in order to convert recorded audio to text, first the speech recognition engine must be trained to recognize the speaker supplying the originally recorded audio input.
Speaker Recognition is the process of automatically recognizing who is speaking on the basis of individual information included in speech signals. Speaker Recognition can be divided into Speaker Identification and Speaker Verification. Speaker Identification determines which registered speaker provides a given utterance from amongst a set of known speakers. By comparison, Speaker Verification accepts or rejects the identity claim of a speaker—is the speaker the person they say they are? Speaker Recognition technology has been applied to the problem of using a speaker's voice to control access to restricted services, for example, phone access to banking, database services, shopping or voice mail, and access to secure equipment. Both technologies require users to “enroll” in the system, that is, to give examples of their speech to a system so that it can characterize (or learn) their voice patterns. Speaker Recognition methods can be divided into text-dependent and text-independent methods.
Paramount to text-independent speaker identification systems is the extraction of features from a given utterance which uniquely belong to a speaker and do not change with time. Specifically, when collecting enrollment data in a speech recognition system, the features of a speaker's speech can be extracted and associated with a known speaker and stored in a database along with a reference, for example a name or identifier associated with the known speaker. Typically, during feature extraction, a speaker-independent phoneme detector can recognize a phoneme that is distinctive from speaker to speaker. The enrollment data subsequently can be retrieved using the reference and compared with features extracted from an unknown speaker voice. If the features extracted from the unknown speaker voice favorably compare with the features of the retrieved enrollment data, the unknown speaker can be identified as the speaker who had provided the retrieved enrollment data.
Notwithstanding advances in Speaker Recognition technology, voice mail systems have yet to incorporate Speaker Recognition technology beyond access control. Moreover, although both computer based voice mail systems and speech recognition systems employing Speaker recognition technology exist, there has yet to be a union of the two technologies to better serve the user. Accurate and efficient transcription of voice mail messages based on Speaker Recognition technology would greatly enhance the usefulness of a computer based voice mail system. As a result, there has arisen a need for a system of transcribing computer voice mail messages.
SUMMARY OF THE INVENTION
The invention disclosed herein for transcribing computer voice mail messages in accordance with the inventive arrangements satisfies the long-felt need of the prior art by using a speech recognition system equipped with Speaker Recognition technology in conjunction with a computer based voice mail system. The invention can receive or import a voice mail message stored in an audio file from a computer voice mail system. After importation of the voice mail message, the system can identify the speaker of the voice mail message. Using enrollment data corresponding to the identified speaker, the system can convert the voice mail message to text, or transcribe, the audio contained in the audio file. Finally, the text can be stored in a text file. Thus, the resulting text file is much smaller in size than the imported audio file from which the text was converted. The decreased file size is especially beneficial for saving storage space and reducing the resources needed to transmit the file. Moreover, the resulting text file can be made available to the user in a variety of forms including, but not limited to displaying the text on a video display terminal, printing the text, transmitting the text file, or storing the text file for use at a later time.
The invention concerns a me
Lewis James R.
Ortega Kerry A.
Akerman & Senterfitt
Chawan Vijay
LandOfFree
Method of transcribing text from computer voice mail 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 of transcribing text from computer voice mail, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of transcribing text from computer voice mail will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3301331