Unlock instant, AI-driven research and patent intelligence for your innovation.

Background learning of speaker voices

A speaker and speaker model technology, applied in speech analysis, speech recognition, measuring devices, etc., can solve problems such as system difficulty and achieve fast and simple registration

Inactive Publication Date: 2006-01-11
KONINKLIJKE PHILIPS ELECTRONICS NV
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

It also makes the system difficult for casual users who are not familiar with registration

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Background learning of speaker voices
  • Background learning of speaker voices
  • Background learning of speaker voices

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] figure 1 A block diagram of a speaker recognition system according to the present invention is shown. The system consists of three main units that execute consecutively in time: background learning 110 , speaker registration 120 and speaker recognition 130 . Background learning includes speech data acquisition 112, followed by blind clustering of speech utterances based on speaker characteristics. The goal of blind utterance clustering is to group unknown utterances when no initial information is available about speaker identity or even about speaker group size. The details of this part will be described below. Once the clusters are generated, speaker model 116 ensures that the utterances in each of these clusters are used to train models each belonging to a possible speaker. The model is best trained using traditional Gaussian Mixture Model (GMM) techniques, where a set of M clusters is defined by the GMM's {λ 1 c ,λ 21 c ,...,λ M c}express. Those familiar wi...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

A speaker identification system includes a speaker model generator 110 for generating a plurality of speaker models. To this end, the generator records training utterances from a plurality of speakers in the background, without prior knowledge of the speakers who spoke the utterances. The generator performs a blind clustering of the training utterances based on a predetermined criterion. For each of the clusters a corresponding speaker model is trained. A speaker identifier 130 identifies a speaker determining a most likely one of the speaker models for an utterance received from the speaker. The speaker associated with the most likely speaker model is identified as the speaker of the test utterance.

Description

technical field [0001] The present invention relates to automatic speaker identification by receiving a test utterance; determining the most probable one of a plurality of speaker models for the test utterance; and determining the speaker associated with the most probable speech model as the speaker of the test utterance. Background technique [0002] Speaker recognition is becoming increasingly important. Traditional speaker recognition is used for security purposes, such as verifying a speaker's identity based on vocal characteristics. As more and more voice-activated applications are developed for CE devices, speaker recognition can also play an important role in simplifying the interaction with CE devices in the future. [0003] In the task of traditional speaker identification (speaker ID), the customer's registration data is used to train a speaker-specific model. Sub-word units, such as phonemes or diphones, are often modeled using Hidden Markov Models (HMMs). In o...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G10L17/00G10L15/06G01L17/00G10L15/07G10L17/04
CPCG10L15/07G10L17/04
Inventor C·-S·黄Y·-C·楚W·-H·蔡J·-M·程
Owner KONINKLIJKE PHILIPS ELECTRONICS NV