Subband Information Research Articles

A Feature Integration Network for Multi-Channel Speech Enhancement.

Multi-channel speech enhancement has become an active area of research, demonstrating excellent performance in recovering desired speech signals from noisy environments. Recent approaches have increasingly focused on leveraging spectral information from multi-channel inputs, yielding promising results. In this study, we propose a novel feature integration network that not only captures spectral information but also refines it through shifted-window-based self-attention, enhancing the quality and precision of the feature extraction. Our network consists of blocks containing a full- and sub-band LSTM module for capturing spectral information, and a global-local attention fusion module for refining this information. The full- and sub-band LSTM module integrates both full-band and sub-band information through two LSTM layers, while the global-local attention fusion module learns global and local attention in a dual-branch architecture. To further enhance the feature integration, we fuse the outputs of these branches using a spatial attention module. The model is trained to predict the complex ratio mask (CRM), thereby improving the quality of the enhanced signal. We conducted an ablation study to assess the contribution of each module, with each showing a significant impact on performance. Additionally, our model was trained on the SPA-DNS dataset using a circular microphone array and the Libri-wham dataset with a linear microphone array, achieving competitive results compared to state-of-the-art models.

Subjective and objective quality evaluation for industrial images

Recently, the demand for ever-better image processing technique continues to grow in field of industrial scenario monitoring and industrial process inspection. The subjective and objective quality evaluation of industrial images are vital for advancing the development of industrial visual perception and enhancing the quality of industrial image/video processing applications. However, the scarcity of publicly available industrial image databases with reliable subjective scores restricts the development of industrial image quality evaluation (IIQE). In preparation for a vacancy, this article first establishes two industrial image databases (i.e., industrial scenario image dataset (ISID) and industrial process image dataset (IPID)) for assessing IIQE metrics. Furthermore, in order to avoid overwhelming industrial image nuances due to the wavelet subband summation, we then present a novel industrial application subband information fidelity standard (SIFS) evaluation method using the channel capacity of visual signals in wavelet domain. Specifically, we first build a visual signals channel model based on perception process from human eyes to brain. Second, we compute and compare the channel capacity for reference and distorted images to measure the information fidelity in each wavelet subband. Third, we sum over the subbands for information fidelity ratio to obtain the overall quality score. Finally, we fairly compare some up-to-date and our proposed image quality evaluation (IQE) methods in two novelty industrial datasets respectively. Our ISID and IPID datasets are capable of evaluating most IQE metrics comprehensively and paves the way for further research on IIQE. Our SIFS model show a remarkable performance comparing with other up-to-date IQE methods.

Vehicle detection in varied weather conditions using enhanced deep YOLO with complex wavelet

Traffic congestion is prevalent in many major and medium-sized cities throughout different countries in contemporary society. In traffic images, various multi-sized vehicles are tightly clustered together and obstructed from one another. Identifying vehicles in such instances is crucial for urban traffic surveillance, safety monitoring, and legal concerns but it also presents major challenges. The remarkable detection accuracy and efficiency of deep learning-based systems have led to their recent and extensive use in vehicle identification. There are significant advanced YOLO models with different backbone architectures and frameworks developed for vehicle detection. Yet, the performance of YOLO variants are facing the challenges of handling false detection against occluded and densely sophisticated scenarios. The proposed model is developed to address such types of limitations, for example; dynamic illumination, noisy images, and scale sensitivity to improve the vehicle detection rate in different traffic scenarios and varying weather conditions. The proposed study employs an improved YOLOv4 to identify moving vehicles in different lighting conditions including daylight, cloudy, rainy, and night. For hybridization, three techniques are utilized such as the Multiscale Retinex, Dual tree complex wavelet transform (DTCWT), and Pulse Coupled Neural Networks (PCNN). The DTCWT is employed for multiscale decomposition and to denoise the complex high frequency subband information, then the denoised subbands are reconstructed into a denoised image. The Multiscale retinex is utilized to reduce the halo artifacts on high-contrast edges and maintain the balance with dynamic range compression and color reproduction. The synchronizing pulse burst property of PCNN is used to detect the isolated noisy pixels and modify the detected noisy pixels. From the results it is worth noting that the developed model surpasses state-of-the-art methods in sunny, night, cloudy, and rainy modes. The proposed method using the DTCWT technique can detect the vehicles with mAP of 91.09% and 35FPS.

Detection of cardiac abnormalities from 12-lead ecg using complex wavelet sub-band features

Aim of the study. This research endeavours to optimize cardiac anomaly detection by introducing a method focused on selecting the most effective Daubechis wavelet families. The principal aim is to differentiate between cardiac states that are normal and abnormal by utilizing longer electrocardiogram (ECG) signal events based on the Apnea ECG dataset. Apnea ECG is often used to detect sleep apnea, a sleep disorder characterized by repeated interruptions in breathing during sleep. By using machine learning methods, such as Principal Component Analysis (PCA) and different classifiers, the goal is to improve the precision of cardiac irregularity identification. Used method. To extract important statistical and sub-band information from lengthy ECG signal episodes, the study uses a novel method that combines discrete wavelet transform with Principal Component Analysis (PCA) for dimension reduction. The methodology focuses on successfully categorizing ECG signals by utilizing several classifiers, including multilayer perceptron (MLP) neural network, Ensemble Subspace K-Nearest Neighbour(KNN), and Ensemble Bagged Trees, together with varied Daubechis wavelet families (db2, db3, db4, db5, db6). Brief Description of Results. The results emphasize the importance of the chosen Daubechis wavelet family, db5, and its superiority in ECG representation. The method distinguishes normal and abnormal ECG signals well on the Physionet Apnea ECG database. The Neural Network-based method accurately recognizes 100% of healthy signals and 97.8% of problematic ones with 98.6% accuracy. Findings. The Ensemble Subspace K-Nearest Neighbour (KNN) and Ensemble Bagged Trees methods got 87.1% accuracy and 0.89 and 0.87 AOC curve values on this dataset, showing that the method works. Precision values of 0.96, 0.86, and 0.86 for MLP Neural Network, KNN Subspace, and Ensemble Bagged Trees confirm their robustness. These findings suggest wavelet families and machine learning can improve cardiac abnormality detection and categorization.

EEG-based Emotion Recognition Using Sub-Band Time-Delay Correlations.

Emotion recognition is a challenging task with many potential applications in psychology, psychiatry, and human-computer interaction (HCI). The use of time-delay information in the controlled time-delay stability (cTDS) algorithm can help to capture the temporal dynamics of EEG signals, including sub-band information and bi-directional coupling that can aid in emotion recognition and identification of specific connectivity patterns between brain rhythms. Incorporating EEG frequency bands can be used to design better emotion recognition systems. This paper evaluates the cTDS algorithm for binary classification tasks of arousal and valence using EEG sub-band signals. This method achieved a high accuracy of 91.1% for arousal and 91.7% for valence based on one electrode recording site at Fp1. The cTDS algorithm is a promising approach to analyzing brain network interactions. It can be particularly applicable to arousal and valence classification tasks, especially within a complex, multimodal feature space associated with understanding psychiatric disorders and HCI applications.

Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations?

It has been suggested that perceiving blurry images in addition to sharp images contributes to the development of robust human visual processing. To computationally investigate the effect of exposure to blurry images, we trained convolutional neural networks (CNNs) on ImageNet object recognition with a variety of combinations of sharp and blurred images. In agreement with recent reports, mixed training on blurred and sharp images (B+S training) brings CNNs closer to humans with respect to robust object recognition against a change in image blur. B+S training also slightly reduces the texture bias of CNNs in recognition of shape-texture cue conflict images, but the effect is not strong enough to achieve human-level shape bias. Other tests also suggest that B+S training cannot produce robust human-like object recognition based on global configuration features. Using representational similarity analysis and zero-shot transfer learning, we also show that B+S-Net does not facilitate blur-robust object recognition through separate specialized sub-networks, one network for sharp images and another for blurry images, but through a single network analyzing image features common across sharp and blurry images. However, blur training alone does not automatically create a mechanism like the human brain in which sub-band information is integrated into a common representation. Our analysis suggests that experience with blurred images may help the human brain recognize objects in blurred images, but that alone does not lead to robust, human-like object recognition.

A Reliable Online Dictionary Learning Denoising Strategy for Noisy Microseismic Data

Improving the quality of microseismic recordings is a critical step in microseismic data processing. We introduce a wavelet-weighted online dictionary learning (WWODL) denoising strategy for noisy microseismic recordings. We develop an adaptive parameters estimation approach for tunable Q-factor wavelet transform, which provides accurate periodic and non-stationary subband information from microseismic data for online dictionary learning (ODL). A sliding time window is employed to divide the obtained subbands into a series of patches of equal length, which are then assembled into a matrix and fed into the ODL. The subband kurtosis information is weighted to the constraint function of the ODL, further enhancing the sparse coding ability for each subband. Shortening the oscillation duration, an improved ODL is developed with a faster convergence speed in calculating sparse coefficients. Our results confirm that the WWODL can suppress high-frequency, low-frequency, and shared-bandwidth noises and has a minimal impact on the first arrival. The time consumption and signal-to-noise ratio of the WWODL are average 1/8.675 and 56.82% higher than ODL, respectively.

Application of DTCWT Decomposition and Partial Differential Equation Denoising Methods in Remote Sensing Image Big Data Denoising and Reconstruction

The precision of the traditional satellite remote sensing image denoising model cannot deal well with some precise production scenes. To solve this problem, this research proposes an improved remote sensing image processing model, in which the dual tree complex wavelet transform (DTCWT) method is used to conduct multiscale decomposition of the impact, and the fourth-order differential equation is used to denoise the decomposed complex high-frequency subband information, and then the denoised subbands are reconstructed into the denoised image. Through these two advanced signal-processing methods, the quality of reconstructed signals is improved and the noise content of various types is greatly reduced. The experimental results show that the normalized root mean square error of the denoising model designed in this study after training convergence is 0.02. When the noise variance is 0.030, the structure similarity, peak signal to noise ratio, and normalized signal to noise ratio are 0.74, 25.3, and 0.76, respectively, which are better than all other comparison models. The experimental data prove that the satellite remote sensing image data denoising model designed in this study has better denoising performance, and has certain application potential in high-precision satellite remote sensing image big data processing.

Modeling Sub-Band Information Through Discrete Wavelet Transform to Improve Intelligibility Assessment of Dysarthric Speech

Modeling Sub-Band Information Through Discrete Wavelet Transform to Improve Intelligibility Assessment of Dysarthric Speech

Deep Dehazing Network for Remote Sensing Image with Non-Uniform Haze

The haze in remote sensing images can cause the decline of image quality and bring many obstacles to the applications of remote sensing images. Considering the non-uniform distribution of haze in remote sensing images, we propose a single remote sensing image dehazing method based on the encoder–decoder architecture, which combines both wavelet transform and deep learning technology. To address the clarity issue of remote sensing images with non-uniform haze, we preliminary process the input image by the dehazing method based on the atmospheric scattering model, and extract the first-order low-frequency sub-band information of its 2D stationary wavelet transform as an additional channel. Meanwhile, we establish a large-scale hazy remote sensing image dataset to train and test the proposed method. Extensive experiments show that the proposed method obtains greater advantages over typical traditional methods and deep learning methods qualitatively. For the quantitative aspects, we take the average of four typical deep learning methods with superior performance as a comparison object using 500 random test images, and the peak-signal-to-noise ratio (PSNR) value using the proposed method is improved by 3.5029 dB, and the structural similarity (SSIM) value is improved by 0.0295, respectively. Based on the above, the effectiveness of the proposed method for the problem of remote sensing non-uniform dehazing is verified comprehensively.

Jointly Leveraging Decorrelation and Sparsity for Improved Feedback Cancellation in Hearing Aids.

We propose a new adaptive feedback cancellation (AFC) system in hearing aids (HAs) based on a well-posed optimization criterion that jointly considers both decorrelation of the signals and sparsity of the underlying channel. We show that the least squares criterion on subband errors regularized by a p-norm-like diversity measure can be used to simultaneously decorrelate the speech signals and exploit sparsity of the acoustic feedback path impulse response. Compared with traditional subband adaptive filters that are not appropriate for incorporating sparsity due to shorter sub-filters, our proposed framework is suitable for promoting sparse characteristics, as the update rule utilizing subband information actually operates in the fullband. Simulation results show that the normalized misalignment, added stable gain, and other objective metrics of the AFC are significantly improved by choosing a proper sparsity promoting factor and a suitable number of subbands. More importantly, the results indicate that the benefits of subband decomposition and sparsity promoting are complementary and additive for AFC in HAs.

Experimental studies of differences in the distribution parameters of multidimensional spatial-subband object vectors on video images during subband information processing

The variant of representation of various objects on video images in the form of samples of space-subband vectors is resulted. The technique of estimation of differences of parameters of distribution of multidimensional space-subband vectors of objects image is given. As indicators of closeness of two regions of localization of space-subband vectors of different objects, the vector of modules of difference of eigenvalues of corresponding spectral matrices of different objects, received from corresponding subband covariance matrices of objects, is used. Experimental estimations of differences in localization regions of different objects of Copter type using their video images are given. It is shown that when there is an obvious external similarity of objects, the areas of their localization differ significantly in size and orientation. It is shown that the localization region parameters of object vectors are practically invariant to object rotations on the image. The possibility of using spatially sub-band vectors and their estimates to build the decisive rules for the recognition of different objects on video images is shown.

Iterative spatial domain 2-D signal decomposition for effectual image up-scaling

Image up-scaling employs various polynomial interpolation schemes for their reduced computational complexity and suitability for various real-time applications. However, they give blurring artifacts in up-scaled images due to the loss of high frequency (HF) information. Likewise, most of the other edge directed and transform domain interpolation schemes available in the literature though produce lesser blurring as compared to polynomial interpolation schemes but are computationally more complex. To overcome these problems, an iterative spatial domain 2-D signal decomposition technique is proposed. It is meant for extracting the very high frequency (VHF) information from a low resolution (LR) image. The VHF information is obtained by performing the signal decomposition for an estimated number of iterations. Subsequently, the superimposition of this VHF extract with the low resolution image prior to image up-scaling reduces the blurring in its up-scaled counterpart. Since the degradation of higher order sub-band information such as HF and VHF is more than the low and medium frequency information during an up-scaling process, restoration of the most degraded VHF sub-band information would produce much lesser blurring. Simulation results reveal that the proposed scheme gives better performance than many of the existing schemes in terms of objective and subjective measures.

Replay attack detection using variable-frequency resolution phase and magnitude features

Replay attacks pose the most severe threat to automatic speaker verification systems among various spoofing attacks. In this paper, we propose a novel feature extraction method that leverages both the phase-based and magnitude-based features. The proposed method fully utilizes the subband information and the complementary information from the phase and magnitude spectra. First, we conduct a discriminative performance analysis on full frequency bands via the F-ratio method. Then, variable-frequency resolution features are extracted via several techniques to capture highly discriminative information on frequency bands. Finally, complementary information from the phase and magnitude domains are fused to achieve higher performance. The results on the ASVspoof 2017 database demonstrate that our proposed frequency adaptive features attain relative error reduction rates of 83.4% and 62.3% on the development and evaluation datasets, respectively, compared to the baseline method.

An algorithm for three-dimensional plumonary parenchymal segmentation by integrating surfacelet transform with pulse coupled neural network

In order to overcome the difficulty in lung parenchymal segmentation due to the factors such as lung disease and bronchial interference, a segmentation algorithm for three-dimensional lung parenchymal is presented based on the integration of surfacelet transform and pulse coupled neural network (PCNN). First, the three-dimensional computed tomography of lungs is decomposed into surfacelet transform domain to obtain multi-scale and multi-directional sub-band information. The edge features are then enhanced by filtering sub-band coefficients using local modified Laplacian operator. Second, surfacelet inverse transform is implemented and the reconstructed image is fed back to the input of PCNN. Finally, iteration process of the PCNN is carried out to obtain final segmentation result. The proposed algorithm is validated on the samples of public dataset. The experimental results demonstrate that the proposed algorithm has superior performance over that of the three-dimensional surfacelet transform edge detection algorithm, the three-dimensional region growing algorithm, and the three-dimensional U-NET algorithm. It can effectively suppress the interference coming from lung lesions and bronchial, and obtain a complete structure of lung parenchyma.

Infrared and Visible Image Fusion Based on a Latent Low-Rank Representation Nested With Multiscale Geometric Transform

To solve the problems of low image contrast and low feature representation in infrared and visible image fusion, an image fusion algorithm based on latent low-rank representation (LatLRR) and non-subsampled shearlet transform (NSST) methods is proposed. First, infrared and visible images are decomposed into base subbands, saliency subbands and sparse noise subbands by the LatLRR model. Then, the base subbands are decomposed into low-frequency and high-frequency coefficients by NSST, and a feature extraction algorithm based on VGGNet and a logical weighting algorithm based on filtering are proposed to merge the coefficients. An adaptive threshold algorithm based on the regional energy ratio is proposed to fuse the saliency subbands. Finally, the fused base subbands are reconstructed, the sparse noise subbands are discarded, and a fused image is obtained by combining the subband information after fusion. Experimental results show that for the fused image produced, the algorithm performs well in both subjective and objective evaluation.

A novel aliasing-free subband information fusion approach for wideband sparse spectral estimation

Wideband sparse spectral estimation is generally formulated as a multi-dictionary/multi-measurement (MD/MM) problem which can be solved by using group sparsity techniques. In this paper, the MD/MM problem is reformulated as a single sparse indicative vector (SIV) recovery problem at the cost of introducing an additional system error. Thus, the number of unknowns is reduced greatly. We show that the system error can be neglected under certain conditions. We then present a new subband information fusion (SIF) method to estimate the SIV by jointly utilizing all the frequency bins. With orthogonal matching pursuit (OMP) leveraging the binary property of SIV’s components, we develop a SIF-OMP algorithm to reconstruct the SIV. The numerical simulations demonstrate the performance of the proposed method.

Non‐negative matrix factorisation‐based subband decomposition for acoustic source localisation

A novel non-negative matrix factorization (NMF) based subband decomposition in frequency spatial domain for acoustic source localization using a microphone array is introduced. The proposed method decomposes source and noise subband and emphasises source dominant frequency bins for more accurate source representation. By employing NMF, delay basis vectors and their subband information in frequency spatial domain for each frame is extracted. The proposed algorithm is evaluated in both simulated noise and real noise with a speech corpus database. Experimental results clearly indicate that the algorithm performs more accurately than other conventional algorithms under both reverberant and noisy acoustic environments.

Early Quantitative Gamma-Band EEG Marker is Associated with Outcomes After Cardiac Arrest and Targeted Temperature Management.

Brain recovery after cardiac arrest (CA) is sensitive to temperature. Yet the effect of temperature management on different EEG frequency bands has not been elucidated. A novel quantitative EEG algorithm, sub-band information quantity (SIQ), was applied to evaluate EEG recovery and outcomes after CA. Twenty-four Wistar rats undergoing 7-min CA were randomly assigned to immediate hypothermia (32-34 °C), normothermia (36.5-37.5 °C), or hyperthermia (38.5-39.5 °C) (n = 8). EEG was recorded continuously for the first 8 h and then for serial 30-min epochs daily. The neurologic deficit score (NDS) at 72-h was the primary functional outcome. Another four rats without brain injury were added as a control. Better recovery of gamma-band SIQ was found in the hypothermia group (0.60 ± 0.03) compared with the normothermia group (0.40 ± 0.03) (p < 0.01) and in the normothermia group compared with the hyperthermia group (0.34 ± 0.03) (p < 0.05). The NDS was also improved in the lower temperature groups: hypothermia [median (25th, 75th), 74 (61, 74)] versus normothermia [49 (47, 61)] versus hyperthermia [43 (0, 50)] (p < 0.01). Throughout the 72-h experiment, the gamma-band SIQ showed the strongest correlation at every time point (ranging 0.520-0.788 from 30-min to 72-h post-resuscitation, all p < 0.05) whereas the delta-band SIQ had poor correlation with the 72-h NDS. No significant difference of sub-band EEG was found with temperature manipulation alone. Recovery of gamma-band SIQ-qEEG was strongly associated with functional outcomes after CA. Induced hypothermia was associated with faster recovery of gamma-band SIQ and improved functional outcomes. Targeted temperature management primarily affected gamma frequency oscillations but not delta rhythm.

A Conceptual Framework for Relative Valuation

<h3>Abstract</h3> It is suggested that experiences of perceiving blurry images in addition to sharp images contribute to the development of robust human visual processing. To computationally investigate the effect of exposure to blurry images, we trained Convolutional Neural Networks (CNNs) on ImageNet object recognition with a variety of combinations of sharp and blurry images. In agreement with related studies, mixed training on sharp and blurred images (B+S) makes the CNNs close to humans with respect to robust object recognition against a change in image blur. B+S training also reduces the texture bias of CNN in recognition of shape-texture-cue-conflict images, but the effect is not strong enough to achieve a strong shape bias comparable to what humans show. Other tests also suggest that B+S training is not sufficient to produce robust human-like object recognition based on global con-figurational features. We also show using representational similarity analysis and zero-shot transfer learning that B+S-Net does not acquire blur-robust object recognition through separate specialized sub-networks, each for sharp and blurry images, but through a single network analyzing common image features. However, blur training alone does not automatically create a mechanism like the human brain where subband information is integrated into a common representation. Our analyses suggest that experience with blurred images helps the human brain develop neural networks that robustly recognize the surrounding world, but it is not powerful enough to fill a large gap between humans and CNNs.