Performance In Noisy Conditions Research Articles

Estimation of CO 2 saturation maps from synthetic seismic data using a deep-learning method with a multi-scale approach

In the context of Carbon Capture and Storage (CCS), monitoring the behaviour of CO 2 within subsurface reservoirs is pivotal for efficient and safe storage operations. This study proposes a new approach to enhance the accuracy of predicting CO 2 saturation maps directly from shot gathers, using a combination of deep learning (DL) and feature extraction methods. We employ a U-Net model, specifically tailored to solve regression tasks, and two-dimensional continuous wavelet transform (2D-CWT) to analyse shot gathers at different scales. We introduce a novel approach using multi-channel input data for the DL model, combining shot gathers and CWT images. The DL model was trained and tested using both single channel and multichannel input. The single channel datasets included shot gathers and CWT images at the first scale, while the multi-channel datasets integrated shot gathers with either two or four scales of CWT. We conducted a sensitivity analysis on the number of training epochs to compare the performance of the model with multichannel and single channel input. Additionally, a Transfer Learning approach was implemented to improve performance in noisy conditions by leveraging knowledge from a pre-trained model on noise-free images. Monte Carlo dropout was used to predict pixel-scale variability and provided a prediction variability with a standard deviation ranging from 0.019 to 0.194, enhancing the robustness of CO 2 saturation estimates. Our results suggest that combining feature extraction using 2D-CWT with the U-Net architecture improves the prediction performance of CO 2 saturation in a synthetic CCS model. Additionally, using multi-channel input instead of single channel is more efficient as it requires less training and prediction time to achieve comparable results. This confirms the effectiveness of 2D-CWT as a pre-processing technique, contributing valuable insights and advances in the field of data-driven geophysical inversion problems.

Long-term variable Q transform: A novel time-frequency transform algorithm for synthetic speech detection

The advanced speech synthetic techniques pose great threats to current speech-based biometric systems, which causes urgent demand for efficient countermeasures. However, existing detection features cannot generalize well for synthetic speeches created by unknown methods. Based on the analysis of different synthetic techniques, it is found that temporal long-term relation and high-frequency information are useful to capture more generalized and robust artifacts of synthetic speeches. To leverage the aforementioned clues, a novel time-frequency transform algorithm known as long-term variable Q transform (L-VQT) is proposed through time-frequency analysis. The core idea of L-VQT is that the variation of discrete frequencies is designed to follow the power function, which aims to perform long temporal windows at low frequencies through small frequency bandwidths. Besides, sufficient high-frequency information can be obtained by properly varying the exponent of the power function. Then, the log power spectrum feature based on L-VQT is sent to a modified Densely Connected Convolutional Network (DenseNet), referred to as Light DenseNet, which is used as a single system to obtain the detection result. Extensive experiments are conducted on ASVspoof 2019 logical access corpus. The proposed detection scheme achieves better generalization capability when compared with other state-of-the-art methods. It also retains a more robust performance in noisy conditions.

Unsupervised feature selection and NMF de-noising for robust Speech Emotion Recognition

Speech feature fusion is the most commonly used phenomenon for improving the accuracy in Speech Emotion Recognition (SER). But in this, there is a disadvantage of increasing the complexity in SER system in terms of processing time. Besides this, some of the features could be redundant and do not contribute for SER and lead to incorrect emotion prediction and reduction in SER accuracy. To overcome this problem, in this paper, unsupervised feature selection is applied to the feature set with the combination of INTERSPEECH 2010 paralinguistic features, Gammatone Cepstral Coefficients (GTCC) and Power Normalized Cepstral Coefficients (PNCC). The Feature Selection with Adaptive Structure Learning (FSASL), Unsupervised Feature Selection with Ordinal Locality (UFSOL) and a novel Subset Feature Selection (SuFS) algorithm are the feature dimension reduction techniques used to acquire better SER performance in this work. The proposed SER system is analyzed in both clean and noisy environments. The EMO-DB and IEMOCAP emotion databases are considered for evaluating the proposed SER performance. For noise analysis, the clean speech is corrupted with different noises of Aurora noise database and white Gaussian noise at different Signal to Noise Ratio (SNR) levels from −5dB to 20 dB. Support Vector Machine (SVM) classifier with linear and Radial Basis Function (RBF) kernels using 10-fold cross-validation and hold-out validation is used in this analysis with classification accuracy and computation time as the performance metrics. The results show that the proposed SER system outperforms the baseline SER system as well as many of the existing literature works both in clean and noisy conditions. For SNR levels >15 dB, the proposed SER system in presence of different noises performs same as the SER in clean environments. Whereas, for lower SNRs <15 dB the performance is likely to be reduced. Therefore, to overcome this drawback and improve the SER performance in noisy conditions, a dense Non-Negative Matrix Factorization (denseNMF) method is adopted for de-noising the noisy speech signal prior to SER achieving noise robustness.

Microwave tomography with phaseless data on the calcaneus by means of artificial neural networks.

The aim of this study is to use a multilayer perceptron (MLP) artificial neural network (ANN) for phaseless imaging the human heel (modeled as a bilayer dielectric media: bone and surrounding tissue) and the calcaneus cross-section size and location using a two-dimensional (2D) microwave tomographic array. Computer simulations were performed over 2D dielectric maps inspired by computed tomography (CT) images of human heels for training and testing the MLP. A morphometric analysis was performed to account for the scatterer shape influence on the results. A robustness analysis was also conducted in order to study the MLP performance in noisy conditions. The standard deviations of the relative percentage errors on estimating the dielectric properties of the calcaneus bone were relatively high. Regarding the calcaneus surrounding tissue, the dielectric parameters estimations are better, with relative percentage error standard deviations up to ≈ 15%. The location and size of the calcaneus are always properly estimated with absolute error standard deviations up to ≈ 3 mm. Microwave tomography of the calcaneus using phaseless data. Simulations were inspired in Computed Tomography images from real heels (above). Inverse problem was solved using Multilayer Perceptron Artificial Neural Network (below).

Robust Cochlear-Model-Based Speech Recognition

Accurate speech recognition can provide a natural interface for human–computer interaction. Recognition rates of the modern speech recognition systems are highly dependent on background noise levels and a choice of acoustic feature extraction method can have a significant impact on system performance. This paper presents a robust speech recognition system based on a front-end motivated by human cochlear processing of audio signals. In the proposed front-end, cochlear behavior is first emulated by the filtering operations of the gammatone filterbank and subsequently by the Inner Hair cell (IHC) processing stage. Experimental results using a continuous density Hidden Markov Model (HMM) recognizer with the proposed Gammatone Hair Cell (GHC) coefficients are lower for clean speech conditions, but demonstrate significant improvement in performance in noisy conditions compared to standard Mel-Frequency Cepstral Coefficients (MFCC) baseline.

Noise-Robust Sound-Event Classification System with Texture Analysis

Sound-event classification has emerged as an important field of research in recent years. In particular, investigations using sound data are being conducted in various industrial fields. However, sound-event classification tasks have become more difficult and challenging with the increase in noise levels. In this study, we propose a noise-robust system for the classification of sound data. In this method, we first convert one-dimensional sound signals into two-dimensional gray-level images using normalization, and then extract the texture images by means of the dominant neighborhood structure (DNS) technique. Finally, we experimentally validate the noise-robust approach by using four classifiers (convolutional neural network (CNN), support vector machine (SVM), k-nearest neighbors(k-NN), and C4.5). The experimental results showed superior classification performance in noisy conditions compared with other methods. The F1 score exceeds 98.80% in railway data, and 96.57% in livestock data. Besides, the proposed method can be implemented in a cost-efficient manner (for instance, use of a low-cost microphone) while maintaining high level of accuracy in noisy environments. This approach can be used either as a standalone solution or as a supplement to the known methods to obtain a more accurate solution.

Fluctuating Hearing Loss in the Only Hearing Ear: Cochlear Implantation in the Contralateral Deaf Side.

Objective To investigate the hearing performance of adult patients presenting unilateral deafness with contralateral fluctuating hearing loss who received a cochlear implant on the deaf side. Study Design Case series with chart review. Setting University tertiary referral center. Subjects and Methods Preoperatively and at 6 and 12 months postoperatively, 23 patients underwent pure tone audiometry and speech audiometry with disyllabic and monosyllabic words in a quiet environment and sentences in quiet and noisy (signal-to-noise ratio +10 dB SPL) environments under best-aided conditions. The Abbreviated Profile of Hearing Aid Benefit (APHAB) inventory was evaluated preoperatively and at 6 and 12 months postoperatively. Results No difference was found between pre- and postoperative tests for disyllabic and monosyllabic words. For sentences in quiet and noisy environments, a difference between pre- and postoperative performance was present at 1 year ( P = .002 and P = .02, respectively). In a noisy environment, a difference was present at 6 and 12 months postoperatively as compared with the preoperative value (mean ± SD: 6 months: 42% ± 7.1% vs 61% ± 6.5%, P = .016). A significant improvement in the APHAB score was found at 6 and 12 months postimplantation (Friedman's 2-way analysis of variance by ranks, P < .001). The number of years of hearing deprivation of the deaf ear was not correlated with performance. Conclusion When incapacitating fluctuating hearing loss occurs in patients presenting a contralateral deaf ear, a cochlear implant is indicated in the latter ear, significantly improving performance in noisy conditions and allowing a better quality of communication to be achieved.

A comparative study of noise estimation algorithms for nonlinear compensation in robust speech recognition

Nonlinear compensation models make use of a nonlinear mismatch function, which characterizes the joint effects of additive and convolutional noise, to realize noise-robust speech recognition. Representative compensation models consist of vector Taylor series (VTS), data-driven parallel model combination (DPMC), and unscented transform (UT). The noise parameters of the compensation models, often estimated in the maximum likelihood (ML) sense, are known to play an important role on the system performance in noisy conditions. In this paper, we conduct a systematic comparison between two popular approaches for estimating the noise parameters. The first approach employs the Gauss-Newton method in a generalized EM framework to iteratively maximizing the EM auxiliary function. The second approach views the compensation models from a generative perspective, giving rise to an EM algorithm, analogous to the ML estimation for factor analysis (EM-FA). We demonstrate a close connection between these two approaches: they belong to the family of gradient-based methods except with different convergence rates. Note that the convergence property can be crucial to the noise estimation since model compensation may be frequently carried out in changing noisy environments for retaining desired performance. Furthermore, we present an in-depth discussion on the advantages and limitations of the two approaches, and illustrate how to extend these approaches to allow for adaptive training. The investigated noise estimation approaches are evaluated on several tasks. The first is to fit a GMM model to artificially corrupted samples, and then speech recognition are performed on the Aurora 2 and Aurora 4 tasks.

Stream fusion for multi-stream automatic speech recognition

Multi-stream automatic speech recognition (MS-ASR) has been confirmed to boost the recognition performance in noisy conditions. In this system, the generation and the fusion of the streams are the essential parts and need to be designed in such a way to reduce the effect of noise on the final decision. This paper shows how to improve the performance of the MS-ASR by targeting two questions; (1) How many streams are to be combined, and (2) how to combine them. First, we propose a novel approach based on stream reliability to select the number of streams to be fused. Second, a fusion method based on Parallel Hidden Markov Models is introduced. Applying the method on two datasets (TIMIT and RATS) with different noises, we show an improvement of MS-ASR.

New approach for T-wave end detection on electrocardiogram: Performance in noisy conditions

BackgroundThe detection of T-wave end points on electrocardiogram (ECG) is a basic procedure for ECG processing and analysis. Several methods have been proposed and tested, featuring high accuracy and percentages of correct detection. Nevertheless, their performance in noisy conditions remains an open problem.MethodsA new approach and algorithm for T-wave end location based on the computation of Trapezium's areas is proposed and validated (in terms of accuracy and repeatability), using signals from the Physionet QT Database. The performance of the proposed algorithm in noisy conditions has been tested and compared with one of the most used approaches for estimating the T-wave end point: the method based on the threshold on the first derivative.ResultsThe results indicated that the proposed approach based on Trapezium's areas outperformed the baseline method with respect to accuracy and repeatability. Also, the proposed method is more robust to wideband noise.ConclusionsThe trapezium-based approach has a good performance in noisy conditions and does not rely on any empirical threshold. It is very adequate for use in scenarios where the levels of broadband noise are significant.

WiMAX OFDM Adaptive Modulation with Broadband Wireless Technology HSDPA

This paper works comprises on the analysis of IEEE 802.16 WiMAX standards and the performance analysis of its physical layer using the OFDM adaptive modulation and it uses the concept of cyclic prefix that adds additional bits at the transmitter end. The signal is transmitted through the channel and it is received at the receiver end, then the receiver removes these additional bits in order to minimize the inter symbol interference, to improve the bit error rate and to reduce the power spectrum. An Additive white Gaussian noise (AWGN) channel model is used with Rayleigh fading to resemble the real world scenario. The transmitter and receiver are assumed to be in fixed positions and the channel coding is not used. The transmitted and received signals are considered to in perfect sync with each other in time and frequency. Losses from feeder cables, joints, connectors, jumpers are not taken in consideration. In our research work, we investigated the physical layer performance on the basis of bit error rate, signal to noise ratio, power spectral density and error probability. These parameters are discussed in two different models. The first model is a simple OFDM communication model without the cyclic prefix, while the second model includes cyclic prefix. The study reviewed scientific articles that provide in-depth knowledge that WiMAX provides increased throughput and better protection against multi path propagation and self interference as compared to HSDPA, furthermore, in this paper different aspects of the model were investigated and with MATLAB simulations it has been demonstrated that lower modulation schemes provide better performance in noisy conditions. Higher modulation schemes provide higher data rates, but encounter high BER in noisy and low SNR conditions.

Formula omitted]-Gaussian mixture modelling for speaker recognition

Gaussian mixture model is the conventional approach employed in speaker recognition tasks. Although it is efficient to model specific speaking characteristics of a speaker, especially in quiet environments, its performance in noisy conditions is still far from the human cognitive process. Recently, a new method of α -integration of stochastic models has been proposed based on psychophysical experiments that suggests α -integration is used in a human brain. In this paper, we proposed a method to extend the conventional GMM to the α -integrated GMM ( α -GMM) to model personal speaking traits. Model parameters were re-estimated recursively based on a given data set. The experiments showed that the new approach significantly outperforms the traditional method, especially on telephony speech.

Combined speech enhancement and auditory modelling for robust distributed speech recognition

The performance of automatic speech recognition (ASR) systems in the presence of noise is an area that has attracted a lot of research interest. Additive noise from interfering noise sources, and convolutional noise arising from transmission channel characteristics both contribute to a degradation of performance in ASR systems. This paper addresses the problem of robustness of speech recognition systems in the first of these conditions, namely additive noise. In particular, the paper examines the use of the auditory model of Li et al. [Li, Q., Soong, F.K., Siohan, O., 2000. A high-performance auditory feature for robust speech recognition. In: Proc. 6th Internat. Conf. on Spoken Language Processing (ICSLP), Vol. III. pp. 51–54] as a front-end for a HMM-based speech recognition system. The choice of this particular auditory model is motivated by the results of a previous study by Flynn and Jones [Flynn, R., Jones, E., 2006. A comparative study of auditory-based front-ends for robust speech recognition using the Aurora 2 database. In: Proc. IET Irish Signals and Systems Conf., Dublin, Ireland. pp. 111–116] in which this auditory model was found to exhibit superior performance for the task of robust speech recognition using the Aurora 2 database [Hirsch, H.G., Pearce, D., 2000. The Aurora experimental framework for the performance evaluation of speech recognition systems under noisy conditions. In: Proc. ISCA ITRW ASR2000, Paris, France. pp. 181–188]. In the speech recognition system described here, the input speech is pre-processed using an algorithm for speech enhancement. A number of different methods for the enhancement of speech, combined with the auditory front-end of Li et al., are evaluated for the purpose of robust connected digit recognition. The ETSI basic [ETSI ES 201 108 Ver. 1.1.3, 2003. Speech processing, transmission and quality aspects (STQ); distributed speech recognition; front-end feature extraction algorithm; compression algorithms] and advanced [ETSI ES 202 050 Ver. 1.1.5, 2007. Speech processing, transmission and quality aspects (STQ); distributed speech recognition; advanced front-end feature extraction algorithm; compression algorithms] front-ends proposed for DSR are used as a baseline for comparison. In addition to their effects on speech recognition performance, the speech enhancement algorithms are also assessed using perceptual speech quality tests, in order to examine if a correlation exists between perceived speech quality and recognition performance. Results indicate that the combination of speech enhancement pre-processing and the auditory model front-end provides an improvement in recognition performance in noisy conditions over the ETSI front-ends.

Continuous multiword recognition performance of young and elderly listeners in ambient noise

Hearing threshold shift due to aging is known as a dominant factor to degrade speech recognition performance in noisy conditions. On the other hand, cognitive factors of aging-relating speech recognition performance in various speech-to-noise conditions are not well established. In this study, two kinds of speech test were performed to examine how working memory load relates to speech recognition performance. One is word recognition test with high-familiarity, four-syllable Japanese words (single-word test). In this test, each word was presented to listeners; the listeners were asked to write the word down on paper with enough time to answer. In the other test, five continuous word were presented to listeners and listeners were asked to write the word down after just five words were presented (multiword test). Both tests were done in various speech-to-noise ratios under 50-dBA Hoth spectrum noise with more than 50 young and elderly subjects. The results of two experiments suggest that (1) Hearing level is related to scores of both tests. (2) Scores of single-word test are well correlated with those of multiword test. (3) Scores of multiword test are not improved as speech-to-noise ratio improves in the condition where scores of single-word test reach their ceiling.

Efficient automatic recognition of spoken digit strings

Automatic recognition of spoken digit sequences (such as credit card numbers) is now feasible even in speaker-independent applications over the telephone. However, all recognition tasks have lower performance in noisy conditions. If significant limitations are also imposed on the computational resources used for recognition, then robust speech recognition is still a significant challenge, even for a simple digit vocabulary. Since recognition of continuously spoken digits over telephone links is a very practical application, such recognition was investigated here under different conditions. Traditional hidden Markov model approaches with cepstral analysis were not used, because they are computationally intensive and have not always worked well under adverse acoustic conditions. Simpler spectral analysis was used, combined with a segmental approach. The analysis focuses on locations of spectral peaks, similar to formant tracking, but without the need to estimate peaks for all time frames. The limited nature of the vocabulary (i.e., ten digits) allows this simpler approach. High recognition accuracy is maintained despite being very efficient in both memory and computation. Recent progress will be reported.

Effect of Noise on the Detection of Rib Fractures by Residents

The authors assessed the effect of noise on the detectability of rib fractures by residents. Eight radiology residents read chest posterior-anterior radiographs of 92 subjects with rib fracture(s) and 28 normal subjects to detect rib fracture(s) according to a five-point scale of confidence, under quiet and "noisy" conditions. Each individual's attitude toward noise was measured by a multiple-choice questionnaire. The readers were divided into two groups depending on the questionnaire result: group A readers were accustomed to a quiet environment, and group B readers were accustomed to noisy environments or were unaffected by noise. Group A's performance, measured by the area (Az) under the receiver operating characteristic (ROC) curve, was better in quiet conditions when compared with their performance in noisy conditions; however, the opposite tendency was observed for group B. There was a significant individual difference of performance in response to noise. Effect of noise on the detection of rib fractures depends on an individual's attitude toward sound and noise.