The Estimating Optimal Number of Gaussian Mixtures Based on Incremental k-means for Speaker Identification

نویسندگان

  • Younjeong Lee
  • Ki Yong Lee
  • Joohun Lee
چکیده

Gaussian mixture model (GMM) is generally used to estimate the speaker model from speech for speaker identification. In this paper, we propose the method that estimates the optimal number of Gaussian mixtures based on incremental k-means for speaker identification. In the proposed method, the initialization with the optimal number of mixtures is done by adding dynamically the number of mixtures one by one until the mutual relationship between any two mixtures becomes dependent. The effectiveness of the proposed method is proven by two experiments. Keyword: Gaussian mixture model, Incremental k-means algorithm, Mutual relationship, Speaker identification.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

On the Determination of Optimal Model Order for GMM-Based Text-Independent Speaker Identification

Gaussian mixture models (GMMs) are recently employed to provide a robust technique for speaker identification. The determination of the appropriate number of Gaussian components in a model for adequate speaker representation is a crucial but difficult problem. This number is in fact speaker dependent. Therefore, assuming a fixed number of Gaussian components for all speakers is not justified. I...

متن کامل

Approximate Gaussian Mixtures for Large Scale Vocabularies

We introduce a clustering method that combines the flexibility of Gaussian mixtures with the scaling properties needed to construct visual vocabularies for image retrieval. It is a variant of expectationmaximization that can converge rapidly while dynamically estimating the number of components. We employ approximate nearest neighbor search to speed-up the E-step and exploit its iterative natur...

متن کامل

Text Independent Speaker Identification Using Automatic Acoustic Segmentation

This paper describes an acoustic class dependent technique for text independent speaker identification on very short utterances. The technique is based on maximum likelihood estimation of a Gaussian mixture model representation of speaker identity. Gaussian mixtures are noted for their robustness as a parametric model and their ability to form smooth estimates of rather arbitrary underlying den...

متن کامل

Estimating the second virial coefficients of some real gas mixtures and related thermodynamic views

Using the Gaussian 2003 software and MP2 /6 – 311+ G method for the C2H4 : O2, CO:Cl2 andCO2:CO2 pairs and MP2/6-311++G** method for the CO2:H2O pair and B3lyp/6-31G methodfor the O2:O2 pair the optimized interaction energies between two considered pair molecules ofstudied gases(C2H4:O2, CO:Cl2, CO2:H2O, O2:O2 and CO2:CO2 pairs) as a function of thedistances between the centers of two considere...

متن کامل

A hybrid DEA-based K-means and invasive weed optimization for facility location problem

In this paper, instead of the classical approach to the multi-criteria location selection problem, a new approach was presented based on selecting a portfolio of locations. First, the indices affecting the selection of maintenance stations were collected. The K-means model was used for clustering the maintenance stations. The optimal number of clusters was calculated through the Silhou...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2006