Text Independent Speaker Identification Model Using Finite Doubly Truncated Gaussian Distribution and Hierarchical Clustering
نویسندگان
چکیده
In Speaker Identification the goal is to determine which one of a group of a known voice with best matches with the one of the input voices. Modelling the speaker voices is an important consideration for many applications. In developing the model, it is customary to consider that the voice of the individual speaker is characterized with finite component Gaussian mixture model. However, the Mel frequency-cepstral coefficient associated with the voice spectrum of the individual speaker is having the finite range and may be asymmetrically distributed. This motivated to generalise the Speaker Identification model with Finite Doubly Truncated Gaussian Mixture Model. The number of components in the mixture model is determined by using Hierarchical clustering algorithm. The model parameters are estimated using EM algorithm. The Speaker Identification is carried by maximizing the likelihood function of the individual speaker. The efficiency of the proposed model is studied through accuracy measures with experimentation on 100 speaker’s database. This model performs much better than the existing earlier algorithms in Speaker Identification.
منابع مشابه
Inverted Mel Feature Set based Text-Independent Speaker Identification using Finite Doubly Truncated Gaussian Mixture Model
This paper provides an efficient approach for text-independent speaker identification using the Inverted Mel-frequency Cepstral Coefficients as feature set and Finite Doubly Truncated Gaussian Mixture as Model (FDTGMM). Over the years, Mel-Frequency Cepstral Coefficients (MFCC), modeled on the human auditory system, has been used as a standard acoustic feature set for speech related application...
متن کاملImage Segmentation Method Based On Finite Doubly Truncated Bivariate Gaussian Mixture Model with Hierarchical Clustering
Image segmentation is one of the most important area of image retrieval. In colour image segmentation the feature vector of each image region is ’n’ dimension different from grey level image. In this paper a new image segmentation algorithm is developed and analyzed using the finite mixture of doubly truncated bivariate Gaussian distribution by integrating with the hierarchical clustering. The ...
متن کاملFace Recognition Algorithm Based on Doubly Truncated Gaussian Mixture Model Using Hierarchical Clustering Algorithm
A robust and efficient face recognition system was developed and evaluated. The each individual face is characterized by 2D-DCT coefficients which follows a finite mixture of doubly truncated Gaussian distribution. In modelling the features vector of the face the number of components (in the mixture model) are determined by hierarchical clustering. The model parameters are estimated using EM al...
متن کاملEfficient Text-Independent Speaker Identification using Optimized Hierarchical Mixture Clustering
Conventional Speaker Identification(SI) Systems uses individual Gaussian Mixture Models(GMM) for every speaker. If this method used for the large population Speaker identification systems, then during identification, likelihood computations between an unknown speaker's test feature vectors and speaker models has become a time-consuming process. This approach also increases the computationa...
متن کاملFace recognition system based on Doubly truncated multivariate Gaussian Mixture Model
A face recognition algorithm based on doubly truncated multivariate Gaussian mixture model with DCT is introduced. The truncation on the feature vector with a significant influence on improving the recognition rate of the system using EM algorithm with K-means or hierarchical clustering is implemented. The characteristic model parameters are estimated. The EM algorithm containing the updated eq...
متن کامل