The main goal of this paper is to propose a novel voice activity detector (VAD) approach for Speaker Verification applications which is both faster and robust to noise environment. The idea is to use Support Vector Machine (SVM) parameters estimated using speech and nonspeech of a long utterance to compute a function decision of the proposed VAD approach. Speaker Verification results on TIMIT d...

Currently, new trends in wireless communications are demanding reliable human-machine interaction in real-life environments. However, there are obstacles inhibiting automatic speech recognition systems (ASR) working in noisy environments. The main difficulty is the degradation suffered by ASR systems due to a mismatch between training and test conditions. This paper shows an improved voice acti...

Handsfree interfaces provide a nice solution to add-on devices in car platforms. However, the amount of acoustic disturbances existing in automotive environments usually prevents satisfactory results. In most of the cases, noise reduction techniques involving a voice activity detector (VAD) are required. Through this paper, a robust microphone array processing technique for speech detection und...

This contribution addresses the problem of additive noise reduction using one-channel noise suppression system. A new implementation effective method is suggested and evaluated. The method consists of two independent parts. The noise estimation part is based on the noise matched filter producing an estimation of background noise without the need of a voice activity detector. The noise reduction...

In this paper, a variable threshold voice activity detector (VAD) is developed to control the operation of a two-sensor adaptive noise canceller (ANC). The VAD prohibits the reference input of the ANC from containing some strength of actual speech signal during adaptation periods. The novelty of this approach resides in using the residual output from the noise canceller to control the decisions...

In this communication we first review the human speech production process and feature extraction approaches commonly used in a speaker verification system. Mel Frequency Cepstral Coefficients (MFCCs), delta (regression) features and Cepstral Mean Subtraction (CMS) are covered. A recently proposed feature set, termed Maximum Auto-Correlation Values (MACVs), which utilizes information from the so...

This paper addresses the problem of the mismatch between a silence model and background noises which often occurs in a telephone speech recognition system (SRS) application. At first, the use of parallel model combination (PMC) methods is studied with the respect to this application. Secondly, the effective adaptation of a silence model to various background noises is confirmed. Finally, an ori...

This work presents a novel approach to speech enhancement by exploiting the bimodality of speech and the correlation that exists between audio and visual speech features. For speech enhancement, a visually-derived Wiener filter is developed. This obtains clean speech statistics from visual features by modelling their joint density and making a maximum a posteriori estimate of clean audio from v...

This study presents a dual channel decoder capable of modeling cross-speaker dependencies for segmentation and classification of fillers and feedbacks in conversational speech found in the DEAL corpus. For the same number of Gaussians per state, we have shown improvement in terms of average F-score for the successive addition of 1) increased frame rate from 10 ms to 50 ms 2) Joint Maximum Cross...

This paper presents a new method for estimating the nonstationary noise power spectral density given a noisy signal. The method is based on averaging the noisy speech power spectrum using time and frequency dependent smoothing factors. These factors are adjusted based on signal-presence probability in individual frequency bins. Signal presence is determined by computing the ratio of the noisy s...