Single Channel Speech Enhancement by De-noising Using Stationary Wavelet Transform

A method of single channel speech enhancement is proposed by de-noising using stationary wavelet transform. The approach developed herein processes multi-resolution wavelet coefficients individually and then recovery signal is reconstructed. The time invariant characteristics of stationary wavelet transform is particularly useful in speech de-noising. Experimental results show that the proposed speech enhancement by de-noising algorithm is possible to achieve an excellent balance between suppresses noise effectively and preserves as many target characteristics of original signal as possible. This de-noising algorithm offers a superior performance to speech signal noise suppress.

Predicting Wavelet-Transformed Stock Prices Using a Vanishing Gradient Resilient Optimized Gated Recurrent Unit with a Time Lag

The development of accurate prediction models continues to be highly beneficial in myriad disciplines. Deep learning models have performed well in stock price prediction and give high accuracy. However, these models are largely affected by the vanishing gradient problem escalated by some activation functions. This study proposes the use of the Vanishing Gradient Resilient Optimized Gated Recurrent Unit (OGRU) model with a scaled mean Approximation Coefficient (AC) time lag which should counter slow convergence, vanishing gradient and large error metrics. This study employed the Rectified Linear Unit (ReLU), Hyperbolic Tangent (Tanh), Sigmoid and Exponential Linear Unit (ELU) activation functions. Real-life datasets including the daily Apple and 5-minute Netflix closing stock prices were used, and they were decomposed using the Stationary Wavelet Transform (SWT). The decomposed series formed a decomposed data model which was compared to an undecomposed data model with similar hyperparameters and different default lags. The Apple daily dataset performed well with a Default_1 lag, using an undecomposed data model and the ReLU, attaining 0.01312, 0.00854 and 3.67 minutes for RMSE, MAE and runtime. The Netflix data performed best with the MeanAC_42 lag, using decomposed data model and the ELU achieving 0.00620, 0.00487 and 3.01 minutes for the same metrics.

Crack identification in functionally graded material framed structures using stationary wavelet transform and neural network

In this paper,an integrated procedure is proposed to identify cracks in a portal framed structure made of functionally graded material(FGM)using stationary wavelet transform(SWT)and neural network(NN).Material properties of the structure vary along the thickness of beam elements by the power law of volumn distribution.Cracks are assumed to be open and are modeled by double massless springs with stiffness calculated from their depth.The dynamic stiffness method(DSM)is developed to calculate the mode shapes of a cracked frame structure based on shape functions obtained as a general solution of vibration in multiple cracked FGM Timoshenko beams.The SWT of mode shapes is examined for localization of potential cracks in the frame structure and utilized as the input data of NN for crack depth identification.The integrated procedure proposed is shown to be very effective for accurately assessing crack locations and depths in FGM structures,even with noisy measured mode shapes and a limited amount of measured data.

Terahertz Spectroscopic Characterization and Thickness Evaluation of Internal Delamination Defects in GFRP Composites

The use of terahertz time-domain spectroscopy(THz-TDS)for the nondestructive testing and evaluation(NDT&E)of materials and structural systems has attracted significant attention over the past two decades due to its superior spatial resolution and capabilities of detecting and characterizing defects and structural damage in non-conducting materials.In this study,the THz-TDS system is used to detect,localize and evaluate hidden multi-delamination defects(i.e.,a three-level multi-delamination system)in multilayered GFRP composite laminates.To obtain accurate results,a wavelet shrinkage de-noising algorithm is used to remove the noise from the measured time-of-flight(TOF)signals.The thickness and location of each delamination defect in the z-direction(i.e.,through-the-thickness direction)are calculated from the de-noised TOF signals considering the interaction between the pulsed THz waves and the different interfaces in the GFRP composite laminates.A comparison between the actual and the measured thickness values of the delamination defects before and after the wavelet shrinkage denoising process indicates that the latter provides better results with less than 3.712%relative error,while the relative error of the non-de-noised signals reaches 16.388%.Also,the power and absorbance levels of the THz waves at every interface with different refractive indices in the GFRP composite laminates are evaluated based on analytical and experimental approaches.The present study provides an adequate theoretical analysis that could help NDT&E specialists to estimate the maximum thickness of GFRP composite materials and/or structures with different interfaces that can be evaluated by the THz-TDS.Also,the accuracy of the obtained results highlights the capabilities of the THz-TDS for the NDT&E of multilayered GFRP composite laminates.

Multiscale anisotropic diffusion for ringing artifact suppression in geophysical deconvolution data

Ringing artifact degradations always appear in the deconvolution of geophysical data. To address this problem, we propose a postprocessing approach to suppress ringing artifacts that uses a novel anisotropic diffusion based on a stationary wavelet transform （SWT） algorithm. In this paper, we discuss the ringing artifact suppression problem and analyze the characteristics of the deconvolu- tion ringing artifact. The deconvolution data containing ringing artifacts are decomposed into different SWT sub- bands for analysis, and a new multiscale adaptive aniso- tropic filter is developed to suppress these degradations. Finally, we demonstrate the performance of the proposed method and describe the experiments in detail.

Multiscale classification and its application to process monitoring

Multiscale classification has potential advantages for monitoring industrial processes generally driven by events in different time and frequency domains.In this study, we adopt stationary wavelet transform for multiscale analysis and propose an applicable scale selection method to obtain the most discriminative scale features.Then using the multiscale features, we construct two classifiers:(1) a supported vector machine(SVM) classifier based on classification distance, and(2) a Bayes classifier based on probability estimation.For the SVM classifier, we use 4-fold cross-validation and grid-search to obtain the optimal parameters.For the Bayes classifier, we introduce dimension reduction techniques including kernel Fisher discriminant analysis(KFDA) and principal component analysis(PCA) to investigate their influence on classification accuracy.We tested the classifiers with two simulated benchmark processes:the continuous stirred tank reactor(CSTR) process and the Tennessee Eastman(TE) process.We also tested them on a real polypropylene production process.The performance comparison among the classifiers in different scales and scale combinations showed that when datasets present typical scale features, the multiscale classifier had higher classification accuracy than conventional single scale classifiers.We also found that dimension reduction can generally contribute to a better classification in our tests.

Edge preserving super-resolution algorithm using multi-stage cascaded joint bilateral filter

Super-resolution(SR)algorithms address the inabilities of poor imaging devices,there by producing high quality images with enhanced resolution.We propose a new SR approach which produces sharp high resolution(HR)image using its low resolution(LR)counterparts.The proposed method uses geometric duality for spatially adapting covariance-based interpolation(CBI).To preserve edge information,a multi-stage cascaded joint bilateral filter(MSCJBF)is proposed as an intermediary stage.These edges are incorporated in the high frequency subbands obtained by the stationary wavelet transform(SWT),through nearest neighbor interpolation(NNI)method.Prior to the NNI process,the high frequency subbands undergo two-lobed lanczos interpolation to achieve the desired resolution enhancement.The quantitative and qualitative analysis for various test images prove the superiority of our method.