T-netix, Inc.

Other

Rutgers, the State University of New Jersey

Manish Sharma

Richard James Mammone

Xiaoyu Zhang

Kevin Farrell

Khaled Assaleh

Devang Naik

Han-Seng Liou

The voice print system of the present invention is a subword-based, text-dependent automatic speaker verification system that embodies the capability of user-selectable passwords with no constraints on the choice of vocabulary words or the language. An estimate of the enrollement channel and of the test channel is developed for inverse filtering of the enrollment or the test speech, respectively. Automatic blind speech segmentation allows speech to be segmented into subword units without any linguistic knowledge of the password. Subword modeling is performed using a multiple classifiers. The system also takes advantage of such concepts as multiple classifier fusion and data resampling to successfully boost the performance. Key word/key phrase spotting is used to optimally locate the password phrase. Numerous adaptation techniques increase the flexibility of the base system, and include: channel adaptation, fusion adaptation, model adaptation and threshold adaptation.

Subword-based speaker verification using multiple-classifier fusion, with channel, fusion, model and threshold adaptation

The voice print system of the present invention is a subword-based, text-dependent automatic speaker verification system that embodies the capability of user-selectable passwords with no constraints on the choice of vocabulary words or the language. Automatic blind speech segmentation allows speech to be segmented into subword units without any linguistic knowledge of the password. Subword modeling is performed using a multiple classifiers. The system also takes advantage of such concepts as multiple classifier fusion and data resampling to successfully boost the performance. Key word/key phrase spotting is used to optimally locate the password phrase. Numerous adaptation techniques increase the flexibility of the base system, and include: channel adaptation, fusion adaptation, model adaptation and threshold adaptation.

Subword-based speaker verification with multiple-classifier score fusion weight and threshold adaptation

The present invention is a reliable system and method for detecting a recorded voice, which can be employed independently or to provide protection from fraudulent use of a recording to defeat an automatic speaker recognition system. Several techniques and systems are employed either independently or in combination to verify that a detected audio sample is live and not recorded. Temporal speech characteristics of an audio sample are analyzed to determine whether a sample under examination is so similar to a previous sample as to indicate a recording. Communications channel characteristics are examined to determine whether an sample was recorded on a different channel from a predetermined communications channel. A pattern classifier is trained to distinguish between live and recorded speech. Finally, an &ldquo;audio watermark&rdquo; is used to determine whether a detected audio sample is a recording of a previous communication by an authorized user. In addition, the various techniques of the present invention may be employed in serial or parallel combination with a variety of decisionmaking schemes to provide increased performance.

System and method for detecting a recorded voice

The present invention relates to a method for segmenting speech into subword speech segments. Optimal boundary locations for each estimate of a number of segments are determined within an estimated range of the number of segments. In addition, an optimality criteria is found for each estimate of the number of segments within the range. Using the optimality criteria, the optimal number of subwords are determined. From the location of the boundaries and the optimal number of segments, data can be clustered or speech can be segmented. The method can be used in data processing systems, speaker verification, medium size vocabulary speech recognition systems, language identification systems and coarse subword level speech segmentation processes.

Blind clustering of data with application to speech processing systems

The present invention relates to a pattern recognition system which uses data fusion to combine data from a plurality of extracted features and a plurality of classifiers. Speaker patterns can be accurately verified with the combination of discriminant based and distortion based classifiers. A novel approach using a training set of a 'leave one out' data can be used for training the system with a reduced data set. Extracted features can be improved with a pole filtered method for reducing channel effects and an affine transformation for improving the correlation between training and testing data.

Growing community of inventors

Sommerset, NJ, United States of America

Manish Sharma