A wavelet-based forward BSS algorithm for acoustic noise reduction and speech enhancement

Ghribi, Khadidja; Djendi, Mohamed; Berkani, Daoued

doi:10.1016/j.apacoust.2015.11.011

Cited by 27 publications

(7 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Attackers may balance attack performance and stealthiness by varying the recording distance. They may improve the inference accuracy at a given distance through signal enhancement algorithms [21].…”

Section: B Robustness Of Keystroke Inferencementioning

confidence: 99%

Eavesdropping on Controller Acoustic Emanation for Keystroke Inference Attack in Virtual Reality

Luo,

Nguyen,

Farooq

et al. 2024

Proceedings 2024 Network and Distributed System Security Symposium

View full text Add to dashboard Cite

Understanding the vulnerability of virtual reality (VR) is crucial for protecting sensitive data and building user trust in VR ecosystems. Previous attacks have demonstrated the feasibility of inferring VR keystrokes inside head-mounted displays (HMDs) by recording side-channel signals generated during user-HMD interactions. However, these attacks are heavily constrained by the physical layout or victim pose in the attack scenario since the recording device must be strictly positioned and oriented in a particular way with respect to the victim. In this paper, we unveil a placement-flexible keystroke inference attack in VR by eavesdropping the clicking sounds of the moving hand controller during keystrokes. The malicious recording smartphone can be placed anywhere surrounding the victim, making the attack more flexible and practical to deploy in VR environments. As the first acoustic attack in VR, our system, Heimdall, overcomes unique challenges unaddressed by previous acoustic attacks on physical keyboards and touchscreens. These challenges include differentiating sounds in a 3D space, adaptive mapping between keystroke sound and key in varying recording placement, and handling occasional hand rotations. Experiments with 30 participants show that Heimdall achieves key inference accuracy of 96.51% and top-5 accuracy of 85.14%-91.22% for inferring passwords with 4-8 characters. Heimdall is also robust under various practical impacts such as smartphone-user placement, attack environments, hardware models, and victim conditions.

show abstract

Section: B Robustness Of Keystroke Inferencementioning

confidence: 99%

Eavesdropping on Controller Acoustic Emanation for Keystroke Inference Attack in Virtual Reality

Luo,

Nguyen,

Farooq

et al. 2024

Proceedings 2024 Network and Distributed System Security Symposium

View full text Add to dashboard Cite

show abstract

“…Numerous research works highlighted the problem of speech enhancement on a simple problem of mixing and unmixing signals with convolutive and instantaneous noisy observations [35][36][37]. In the last decade, a novel research direction has proven the efficacy of the wavelet domain as a novel effective mean that can ameliorates the speech enhancement approaches, and numerous algorithms and methods have been proposed for the same aim [38,39]. In this chapter, we propose a novel speech enhancement technique based on Lifting Wavelet Transform (LWT) and Artifitial Neural Network (ANN) and also uses MMSE Estimate of Spectral Amplitude [40].…”

Section: Introductionmentioning

confidence: 99%

Speech Enhancement Based on LWT and Artificial Neural Network and Using MMSE Estimate of Spectral Amplitude

Talbi¹,

Baazaoui²,

Bouhlel³

2021

Deep Learning Applications

View full text Add to dashboard Cite

In this chapter, we will detail a new speech enhancement technique based on Lifting Wavelet Transform (LWT) and Artifitial Neural Network (ANN). This technique also uses the MMSE Estimate of Spectral Amplitude. It consists at the first step in applying the LWTto the noisy speech signal in order to obtain two noisy details coefficients, cD1 and cD2 and one approximation coefficient, cA2. After that, cD1 and cD2 are denoised by soft thresholding and for their thresholding, we need to use suitable thresholds, thrj,1≤j≤2. Those thresholds, thrj,1≤j≤2, are determined by using an Artificial Neural Network (ANN). The soft thresholding of those coefficients, cD1 and cD2, is performed in order to obtain two denoised coefficients, cDd1 and cDd2 . Then the denoising technique based on MMSE Estimate of Spectral Amplitude is applied to the noisy approximation cA2 in order to obtain a denoised coefficient, cAd2. Finally, the enhanced speech signal is obtained from the application of the inverse of LWT, LWT−1 to cDd1, cDd2 and cAd2. The performance of the proposed speech enhancement technique is justified by the computations of the Signal to Noise Ratio (SNR), Segmental SNR (SSNR) and Perceptual Evaluation of Speech Quality (PESQ).

show abstract

“…To provide improved extraction of speech from noise, the Teager energy (TE) operator was also utilized in [31], [37], and [38]. With the use of a two microphone system, the blind source separation (BSS) technique was employed to separate the speech and noise wavelet coefficients in [40], [41]. To reduce speech distortion, various custom threshold functions with continuous derivative characteristic were proposed in [36]− [41] to replace the hard/soft threshold function.…”

Section: Introductionmentioning

confidence: 99%

“…This is in contrast to the need for special psychoacoustic or perceptual wavelets in [35]- [38] and more sophisticated custom threshold functions by the subband-level thresholding algorithms in [32], [33], [35], [37], and [38] to avoid excessive distortion in the enhanced speech. • Unlike its BSS counterparts in [40], [41], the proposed adaptive wavelet thresholding algorithm requires only one single microphone. This paper is organized as follows.…”

Section: Introductionmentioning

confidence: 99%

Performance Improvement of Adaptive Wavelet Thresholding for Speech Enhancement Using Generalized Gaussian Priors and Frame-Wise Context Modeling

2020

View full text Add to dashboard Cite

This work aims at developing an adaptive wavelet thresholding algorithm for speech enhancement with significant performance improvement over other wavelet-based counterparts. This is accomplished through the formulation of the optimum threshold for noise reduction, based on the generalized Gaussian priors to fully characterize the statistics of speech and noise wavelet coefficients. In addition, through the frame-wise context modeling which enables tracking of the statistical characteristics of each individual coefficient on the frame-wise basis, the optimum threshold is accurate and adaptive at both the coefficient level and frame level. The frame-wise context model is formulated by virtue of the context subspace projection of the wavelet coefficients, with the context index employed as the invariant correspondence between successive frame parameters, thereby enabling the frame-wise tracking at the coefficient level. Simulation results show significant improvement over the wavelet-based speech enhancement algorithms in terms of the segmental signal-to-noise ratio improvement by as much as 226%, the perceptual evaluation of speech quality by 36%, the short-time objective intelligibility by 17.8% and the cepstral distance by 33.3%. When benchmarked with the well-established short-time-Fouriertransform-based counterparts, the proposed wavelet thresholding algorithm offers favorable and more robust performances, particularly under non-stationary noise conditions, with no adverse musical noise effect.INDEX TERMS context modeling, speech enhancement, wavelet thresholding.

show abstract

A wavelet-based forward BSS algorithm for acoustic noise reduction and speech enhancement

Cited by 27 publications

References 35 publications

Eavesdropping on Controller Acoustic Emanation for Keystroke Inference Attack in Virtual Reality

Eavesdropping on Controller Acoustic Emanation for Keystroke Inference Attack in Virtual Reality

Speech Enhancement Based on LWT and Artificial Neural Network and Using MMSE Estimate of Spectral Amplitude

Performance Improvement of Adaptive Wavelet Thresholding for Speech Enhancement Using Generalized Gaussian Priors and Frame-Wise Context Modeling

Contact Info

Product

Resources

About