Loading...
Search for:
cepstrum
0.022 seconds
Speech modeling and voiced/unvoiced/mixed/silence speech segmentation with fractionally gaussian noise based models
, Article Proceedings - IEEE International Conference on Acoustics, Speech, and Signal Processing, Montreal, Que, 17 May 2004 through 21 May 2004 ; Volume 1 , 2004 , Pages I613-I616 ; 15206149 (ISSN) ; Shamsollahi, M. B ; Sharif University of Technology
2004
Abstract
The ARMA filtered fractionally differenced Gaussian Noise (FdGn) model and a new AR Filtered FdGn Added up model are applied to speech signal and performance of their parameters on speech Unvoiced/Voiced/Mixed/Silence classification is evaluated against Zero Crossing Rate (ZCR) feature. For parameter estimation of AR filtered FdGn model two methods were applied: iterative Maximum Likelihood (ML) method of Tewfik and a new computationally efficient Linear Minimum Square Error (LMSE) algorithm Also for parameters estimation of new Added up model two approaches were implemented: an Expectation-Maximization (EM) based approach and an iterative MSE approach. The described models and methods were...
A unimodal person authentication system based on signing sound
, Article Proceedings - IEEE-EMBS International Conference on Biomedical and Health Informatics: Global Grand Challenge of Health Informatics, BHI 2012 ; 2012 , Pages 152-154 ; 9781457721779 (ISBN) ; Maghooli, K ; Afdideh, F ; Azimi, H ; Sharif University of Technology
IEEE
2012
Abstract
Person authentication based on only the name, password or person identification number is not secured enough. In recent years researchers have focused on human physiological and behavioral parameters, because these parameters are more unique and human-specific than traditional ones. This approach of person authentication is usually called biometric authentication. Signature is the most commonly used behavioral biometric which is investigated in two ways of online and offline by researchers. In online procedure, the temporal indices of signature such as signing velocity, and acceleration are involved to increase the accuracy relative to offline methods and to recognize counterfeit signatures....
Speech Enhancement Using Deep Neural Networks in Non-stationary Noise Environment
, M.Sc. Thesis Sharif University of Technology ; Ghaem Maghami, Shahrokh (Supervisor)
Abstract
Before performing any operation on a speech signal, it is necessary to properly remove the environmental noise existing on it. Noise Canceling Operation on Speech Signal is called speech enhancement. Up to now, many studies have been conducted on various ways to enhance the speech signal. Among the existing methods, statistical methods have proven to be superior to others. In all noise removal methods, the main challenge is that most noises are non-stationary. Since most of the noises in the environment are non-stationary, we are still looking for the better ways to remove them. With the advent of deep neural networks and their successful results in areas such as machine learning, a method...