Wyniki wyszukiwania - BazTech

1

Speaker Identification using Data-Driven Score Classification

Gan H., Mporas I., Safavi S., Sotudeh R.

Image Processing & Communications

|

2016

|

Vol. 21, no. 2

55--64

EN

We present a comparative evaluation of different classification algorithms for a fusion engine that is used in a speaker identity selection task. The fusion engine combines the scores from a number of classifiers, which uses the GMM-UBM approach to match speaker identity. The performances of the evaluated classification algorithms were examined in both the text-dependent and text-independent operation modes. The experimental results indicated a significant improvement in terms of speaker identification accuracy, which was approximately 7% and 14.5% for the text-dependent and the text-independent scenarios, respectively. We suggest the use of fusion with a discriminative algorithm such as a Support Vector Machine in a real-world speaker identification application where the text-independent scenario predominates based on the findings.

2

Szczegóły implementacyjne algorytmów do rozpoznawania mowy

Ośka J., Wojtuń J., Piotrowski Z., Bernat M.

Elektronika : konstrukcje, technologie, zastosowania

|

2015

|

Vol. 56, nr 2

40-44

PL

W artykule zaprezentowano i porównano algorytmy do rozpoznawania mowy w kontekście ich późniejszej implementacji na platformie sprzętowej DSK OMAP. Głównym zadaniem było dogłębne porównanie dwóch klasycznych metod wykorzystywanych w rozpoznawaniu mowy GMM vs HMM (ang. GMM Gaussian Mixtures Models, ang. HMM – Hidden Markov Models). W artykule jest również opisana i porównana metoda ulepszonych mikstur gaussowskich GMM-UBM (ang. GMM UBM – Gaussian Mixtures Model Universal Background Model). Parametryzacja sygnału w oparciu o współczynniki MFCC oraz LPCC (ang. Mel Frequency Cepstral Coefficients, ang. Linear Prediction Cepstral Coefficients) została opisana [1]. Analizowany model składał się ze zbioru 10-elementowego reprezentującego cyfry mowy polskiej 0-9. Badania zostały przeprowadzone na zbiorze 3000 nagrań, które zostały przygotowane przez nasz zespół. Porównanie wyników wykonano dla rozłącznych zbiorów uczących oraz trenujących. Każda z opisywanych metod klasyfikacji operuje na tych samych danych wejściowych. Daje to możliwość miarodajnego porównania jakości tych klasyfikatorów jako skutecznych narzędzi do rozpoznawania izolowanych fraz głosowych.

EN

This paper presents and compares the speech recognition algorithms in the context of their subsequent implementation on the hardware platform OMAP DSK. The main task was to compare two classical methods used in speech recognition systems GMM vs HMM (GMM – Gaussian Mixtures Models, HMM – Hidden Markov Models). In the article improved Gaussian Mixtures Model called GMM-UBM (Gaussian Mixtures Model Universal Background Model) were described and compared. Preprocessing of the input signal using MFCC and LPCC coefficients (Mel Frequency Cepstral Coefficients, Linear Prediction Cepstral Coefficients) were described [1]. Analyzed data model consists set of 10-elements that represents Polish language digits 0-9. Research is done on a set of 3000 records prepared by our team with disjoint sets of learners and trainees. Methods are compared on the same input data. The same set of input data allows for reliable comparison of these classifiers to choose effective classifier for identifying isolated voice phases.

3

An improved ant colony optimization algorithm and its application to text-independent speaker verification system

Aghdam M. H.

Journal of Artificial Intelligence and Soft Computing Research

|

2012

|

Vol. 2, No. 4

301--315

EN

With the growing trend toward remote security verification procedures for telephone banking, biometric security measures and similar applications, automatic speaker verification (ASV) has received a lot of attention in recent years. The complexity of ASV system and its verification time depends on the number of feature vectors, their dimensionality, the complexity of the speaker models and the number of speakers. In this paper, we concentrate on optimizing dimensionality of feature space by selecting relevant features. At present there are several methods for feature selection in ASV systems. To improve performance of ASV system we present another method that is based on ant colony optimization (ACO) algorithm. After feature selection phase, feature vectors are applied to a Gaussian mixture model universal background model (GMM-UBM) which is a text-independent speaker verification model. The performance of proposed algorithm is compared to the performance of genetic algorithm on the task of feature selection in TIMIT corpora. The results of experiments indicate that with the optimized feature set, the performance of the ASV system is improved. Moreover, the speed of verification is significantly increased since by use of ACO, number of features is reduced over 80% which consequently decrease the complexity of our ASV system.