SCORE STABILIZATION FOR SPEECH CLASSIFICATION
Number of patents in Portfolio can not be more than 2000
United States of America Patent
Stats
-
N/A
Issued Date -
Jul 27, 2017
app pub date -
Jan 21, 2016
filing date -
Jan 21, 2016
priority date (Note) -
Abandoned
status (Latency Note)
Importance
Abstract
A method for stabilizing speaker recognition scores, comprising using one or more hardware processors for the following actions: Receiving supervectors from a Gaussian Mixture model analysis performed by a speaker recognition system, where the supervectors represent speech signals acquired by a microphone. Performing a principal component analysis of a covariance matrix of the supervectors, thereby producing eigenvalues and eigenvectors of the covariance matrix. Removing some of the eigenvectors associated with a number of highest value eigenvalues from the supervectors, thereby producing stabilized supervectors. Sending the stabilized supervectors to the speaker recognition system to compute stabilized speaker recognition scores.
First Claim
Other Claims data not available