The aim of this paper is to use visual speech information to create Wiener filters for audio speech enhancement. Wiener filters require estimates of both clean speech statistics and noisy speech statistics. Noisy speech statistics are obtained from the noisy input audio while obtaining clean speech statistics is more difficult and is a major problem in the creation of Wiener filters for speech ...

We present an improved method on combining temporal and spectral processing approaches for multichannel determined blind sources separation. The separation task is performed by applying the spectral processing on a mixed speech, using sources’ excitation characteristics. The performance of the proposed method is investigated by separating two sources from a stereo recording mixture extracted fr...

Model-based approaches to achieve Single Channel Source Separation (SCSS) have been reasonably successful at separating two sources. However, most of the currently used model-based approaches require pre-trained speaker specific models in order to perform the separation. Often, insufficient or no prior training data may be available to develop such speaker specific models, necessitating the use...

We propose a joint filtering and factorization algorithm to recover latent structure from noisy speech. We incorporate the minimum variance distortionless response (MVDR) formulation within the non-negative matrix factorization (NMF) framework to derive a single, unified cost function for both filtering and factorization. Minimizing this cost function jointly optimizes three quantities – a filt...

Neurology 2012;79:1194–1195 The selection of antiepileptic drugs (AEDs) is based primarily on efficacy; however, in some conditions, efficacy is often equivalent for several AEDs. Hence other factors such as pharmacokinetics, cost, adverse effects (AEs), age, and sex should also be considered when selecting an appropriate AED.1 AEs, which are observed in over one-third of patients with epilepsy...

In this paper, we propose a new adaptive wavelet thresholding method for using in speech Enhancement. This modified version of the wavelet thresholding method updates an adaptive threshold in each frame. In the proposed method the selection of the wavelet threshold value depends on the estimates of the clean speech signal features. The evaluation results show that by using the specific features...

We propose a novel method to estimate the quality of coded speech signals. The joint probability distribution of the subjective mean opinion score (MOS) and perceptual distortion feature variables is modelled using a Gaussian mixture density. The feature variables are sifted from a large pool of candidate features using statistical data mining techniques. We study what combinations of features ...

Objective speech quality assessment algorithms provide low-cost and online monitoring of voice calls, replacing costly and timeconsuming subjective listening tests. We propose a novel approach to objective speech quality measurements using Gaussian mixture models (GMMs). A large pool of perceptual distortion features is extracted from speech files and multivariate adaptive regression splines (M...

Speech enhancement methods using spectral subtraction have the drawback of generating an annoying residual noise with musical character. In this paper a frequency domain optimal linear estimator with perceptual post filtering is proposed which incorporates the masking properties of the human auditory system to make the residual noise distortion inaudible. The performance of the proposed enhance...