Convolutive Mixtures Research Articles

Dimma: Semi-Supervised Low-Light Image Enhancement with Adaptive Dimming.

Enhancing low-light images with natural colors poses a challenge due to camera processing variations and limited access to ground-truth lighting conditions. To address this, we propose Dimma, a semi-supervised approach that aligns with any camera using a small set of image pairs captured under extreme lighting conditions. Our method employs a convolutional mixture density network to replicate camera-specific noise present in dark images. We enhance results further by introducing a conditional UNet architecture based on user-provided lightness values. Trained on just a few real image pairs, Dimma achieves competitive results compared to fully supervised state-of-the-art methods trained on large datasets.

SA-UCBSS: Sparsity-Based Adaptive Underdetermined Convolutive Blind Source Separation

Blind source separation is the process of separating unknown source signals based only on the received mixing signals when the source signal cannot be accurately known. Traditional methods are mainly based on linear approximate transformation models that often have large systematic errors in high reverberation environments, resulting in poor separation performance. To avoid these limitations, a sparsity-based adaptive underdetermined convolutive blind source separation (SA-UCBSS) algorithm is proposed by unifying a convolutive approximate transformation model and an Lp(0<p≤1)-norm regularisation penalty. A crosstalk canceller is designed for room impulse response pretreatment with p-norm optimisation to eliminate audible echoes, and a sparse optimisation scheme is designed for the parameter initialisation to accelerate the convergence of the iterative recovery process. Subsequently, a convolutive approximate transformation model is built to achieve a more accurate estimation of the mixing system. The SA-UCBSS algorithm is proposed to achieve better source separation by adaptively tuning the value of parameter p. Experimental results including BSS of the synthesised and live-recorded speech convolutive mixtures showed that the SA-UCBSS algorithm had superior separating properties and better robustness than several popular algorithms.

Independent vector analysis: Model, applications, challenges

This paper overviews an appealing unsupervised learning method named independent vector analysis (IVA) for its promising applications, such as in audio/speech signal separation, medical signal processing, remote sensing, video/image processing, wireless communication processing, and so on. As a useful data-driven technique in blind source separation (BSS) field, IVA has played an increasingly vital role in dealing with the problems of convolutive mixture separation, multivariate latent variable analysis and multivariate data fusion. IVA extends the conventional independent component analysis (ICA) to multidimensional components, which can result in more available information utilization. Compared with ICA mechanism, IVA is not only to utilize the statistical independence of multivariate signals but also the statistical inner dependency of each multivariate signal. With this generalization, IVA can manipulate some prominent ill-pose issues faced in sensor receiving models and has the advantage to overcome the inherent random permutation ambiguity problem in joint BSS. Motivated by the flexible and versatile technology superiorities of IVA, this paper concentrates on reviewing the IVA model in details, associated methods briefly, and its potential applications as well as prospects. Moreover, some significant open problems about IVA challenges are also discussed in this paper.

A Unifying View on Blind Source Separation of Convolutive Mixtures Based on Independent Component Analysis

In many daily-life scenarios, acoustic sources recorded in an enclosure can only be observed with other interfering sources. Hence, convolutive Blind Source Separation (BSS) is a central problem in audio signal processing. Methods based on Independent Component Analysis (ICA) are especially important in this field as they require only few and weak assumptions and allow for blindness regarding the original source signals and the acoustic propagation path. Most of the currently used algorithms belong to one of the following three families: Frequency Domain ICA (FD-ICA), Independent Vector Analysis (IVA), and TRIple- N Independent component analysis for CONvolutive mixtures (TRINICON). While the relation between ICA, FD-ICA and IVA becomes apparent due to their construction, the relation to TRINICON is not well established yet. This paper fills this gap by providing an in-depth treatment of the common building blocks of these algorithms and their differences, and thus provides a common framework for all considered algorithms.

Nonlinear Signal Estimation Using Semi-Blind Mutually Referenced Equalizers in Convolutive Mixture

Nonlinear Signal Estimation Using Semi-Blind Mutually Referenced Equalizers in Convolutive Mixture

CNN Classification of Multi-Scale Ensemble OCT for Macular Image Analysis

Computer-Aided Diagnosis (CAD) of retinal pathology is a dynamic medical analysis area. The CAD system in the optical coherence tomography (OCT) is important for the monitoring of ocular diseases because of the heavy utilization of the retinal OCT imaging process. The Multi-Scale Expert Convolution Mixture (MCME) is designed to classify the normal retina. OCT is becoming one of the most popular non-invasive evaluation approaches for retinal eye disease. The amount of OCT is growing and the automation of OCT image analysis is becoming increasingly necessary. The surrogate-aided classification approach is to automatically classify retinal OCT images because of the Convolution Neural Network (CNN). The methods to classify OCT images and macular OCT classification are done by using CNN. Maculopathy is a combined collection of diseases to facilitate the effect of the inner region of the retina identified as the macula. Central Serous Choric Retinopathy (CSCR) and macular edema are the main two types of maculopathies. Numerous researches have focused on the detection of these macular disorders with OCT. It is used to overcome retinal diseases.

A probabilistic deep learning model to distinguish cusps and cores in dwarf galaxies

ABSTRACT Numerical simulations within a cold dark matter (DM) cosmology form haloes whose density profiles have a steep inner slope (‘cusp’), yet observations of galaxies often point towards a flat central ‘core’. We develop a convolutional mixture density neural network model to derive a probability density function (PDF) of the inner density slopes of DM haloes. We train the network on simulated dwarf galaxies from the NIHAO and AURIGA projects, which include both DM cusps and cores: line-of-sight velocities and 2D spatial distributions of their stars are used as inputs to obtain a PDF representing the probability of predicting a specific inner slope. The model recovers accurately the expected DM profiles: $\sim 82{{\ \rm per\ cent}}$ of the galaxies have a derived inner slope within ±0.1 of their true value, while $\sim 98{{\ \rm per\ cent}}$ within ±0.3. We apply our model to four Local Group dwarf spheroidal galaxies and find results consistent with those obtained with the Jeans modelling based code GravSphere: the Fornax dSph has a strong indication of possessing a central DM core, Carina and Sextans have cusps (although the latter with large uncertainties), while Sculptor shows a double peaked PDF indicating that a cusp is preferred, but a core cannot be ruled out. Our results show that simulation-based inference with neural networks provide a innovative and complementary method for the determination of the inner matter density profiles in galaxies, which in turn can help constrain the properties of the elusive DM.

Probability density function forecasting of electricity price: Deep gabor convolutional mixture network

This paper establishes a framework for probabilistic electricity price forecasting in the electrical markets using an end-to-end and direct approach. In the proposed method, firstly, a deep Gabor filter-oriented convolutional network is designed as the strong deep structure. Then, the designed deep Gabor network is developed as a deep mixture network to predict a set of probability density functions (PDFs) in the next hours. To do so, a probabilistic loss function is formulated. The results are validated by the actual electricity market data in California independent system operator (CAISO) and show high level of accuracy and reliability in comparison with other state-of-the-arts methods

Deconvolution of the Functional Ultrasound Response in the Mouse Visual Pathway Using Block-Term Decomposition

Functional ultrasound (fUS) indirectly measures brain activity by detecting changes in cerebral blood volume following neural activation. Conventional approaches model such functional neuroimaging data as the convolution between an impulse response, known as the hemodynamic response function (HRF), and a binarized representation of the input signal based on the stimulus onsets, the so-called experimental paradigm (EP). However, the EP may not characterize the whole complexity of the activity-inducing signals that evoke the hemodynamic changes. Furthermore, the HRF is known to vary across brain areas and stimuli. To achieve an adaptable framework that can capture such dynamics of the brain function, we model the multivariate fUS time-series as convolutive mixtures and apply block-term decomposition on a set of lagged fUS autocorrelation matrices, revealing both the region-specific HRFs and the source signals that induce the hemodynamic responses. We test our approach on two mouse-based fUS experiments. In the first experiment, we present a single type of visual stimulus to the mouse, and deconvolve the fUS signal measured within the mouse brain’s lateral geniculate nucleus, superior colliculus and visual cortex. We show that the proposed method is able to recover back the time instants at which the stimulus was displayed, and we validate the estimated region-specific HRFs based on prior studies. In the second experiment, we alter the location of the visual stimulus displayed to the mouse, and aim at differentiating the various stimulus locations over time by identifying them as separate sources.

Accelerated Conjugate Gradient for Second-Order Blind Signal Separation

This paper proposes a new adaptive algorithm for the second-order blind signal separation (BSS) problem with convolutive mixtures by utilising a combination of an accelerated gradient and a conjugate gradient method. For each iteration of the adaptive algorithm, the search point and the search direction are obtained based on the current and the previous iterations. The algorithm efficiently calculates the step size for the accelerated conjugate gradient algorithm in each iteration. Simulation results show that the proposed accelerated conjugate gradient algorithm with optimal step size converges faster than the accelerated descent algorithm and the steepest descent algorithm with optimal step size while having lower computational complexity. In particular, the number of iterations required for convergence of the accelerated conjugate gradient algorithm is significantly lower than the accelerated descent algorithm and the steepest descent algorithm. In addition, the proposed system achieves improvement in terms of the signal to interference ratio and signal to noise ratio for the dominant speech outputs.

One to multiple mapping dual learning: Learning multiple signals from one mixture

Single channel blind source separation (SCBSS) refers to separating multiple sources from a mixture collected by a single sensor. Existing methods for SCBSS have limited performance in separating multiple sources and generalization. To address these problems, an algorithm is proposed in this paper to separate multiple sources from a mixture by designing a parallel dual generative adversarial network (PDualGAN) that can build the relationship between a mixture and the corresponding multiple sources to achieve one-to-multiple cross-domain mapping. This algorithm can be applied to a variety of mixtures including both instantaneous and convolutive mixtures. In addition, new datasets for single channel source separation are created which include the mixtures and corresponding sources for this study. Experiments were performed on four different datasets including both one-dimensional and two-dimensional signals. Experimental results show that the proposed algorithm outperforms state-of-the-art algorithms, measured with peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), source-to-distortion ratio (SDR), source-to-interferences ratio (SIR), relative root mean squared error (RRMSE) and correlation.

Independent Vector Analysis for Blind Deconvolving of Digital Modulated Communication Signals

For the purpose of overcoming the random permutation ambiguity of the frequency-domain-independent component analysis (FDICA) for blind separation of convolutive mixtures, this paper proposes an independent vector analysis (IVA) detection receiver for blindly deconvolving the convolutive mixtures of digitally modulated signals for wireless communications. The foundation of IVA is through jointly carrying out separation work for different frequency bin data fusion, and the dependencies of frequency bins are exploited in solving the random permutation problem of separation signals. In addition, IVA uses multivariate prior distributions instead of the univariate distribution used in FDICA. Multivariate prior distribution is employed to preserve the interfrequency dependencies for individual sources, which can give rise to separation performance enhancement. Simulation results and analysis corroborate the effectiveness of the proposed detection method.

Measuring Dependence for Permutation Alignment in Convolutive Blind Source Separation

This brief proposes an effective implementation for addressing permutation ambiguity issue of convolutive blind source separation in frequency domain. Generally, signal envelope and power ratio as common inter-frequency dependence measures are utilized to group bin-wise separated signals for convolutive mixtures, where the new measure of permutation alignment method is represented by the activation matrix of bin-wise spectrum which is based on nonnegative matrix factorization (NMF). Meanwhile, canonical correlation analysis (CCA) rather than the maximum sum of correlation coefficient among different bins, is applied for verifying correlation determinations. In addition, the influence of bin distance and separation quality at each bin are explored to optimize permutation result. Simulation results demonstrate the effectiveness of the proposed technique in real recorded convolutive mixtures.

Convolutive Transfer Function-Based Multichannel Nonnegative Matrix Factorization for Overdetermined Blind Source Separation

Most multichannel blind source separation (BSS) approaches rely on a spatial model to encode the transfer functions from sources to microphones and a source model to encode the source power spectral density. The rank-1 spatial model has been widely exploited in independent component analysis (ICA), independent vector analysis (IVA), and independent low-rank matrix analysis (ILRMA). The full-rank spatial model is also considered in many BSS approaches, such as full-rank spatial covariance matrix analysis (FCA), multichannel nonnegative matrix factorization (MNMF), and FastMNMF, which can improve the separation performance in the case of long reverberation times. This paper proposes a new MNMF framework based on the convolutive transfer function (CTF) for overdetermined BSS. The time-domain convolutive mixture model is approximated by a frequency-wise convolutive mixture model instead of the widely adopted frequency-wise instantaneous mixture model. The iterative projection algorithm is adopted to estimate the demixing matrix, and the multiplicative update rule is employed to estimate nonnegative matrix factorization (NMF) parameters. Finally, the source image is reconstructed using a multichannel Wiener filter. The advantages of the proposed method are twofold. First, the CTF approximation enables us to use a short window to represent long impulse responses. Second, the full-rank spatial model can be derived based on the CTF approximation and slowly time-variant source variances, and close relationships between the proposed method and ILRMA, FCA, MNMF and FastMNMF are revealed. Extensive experiments show that the proposed algorithm achieves a higher separation performance than ILRMA and FastMNMF in reverberant environments.

Blind Separation for Wireless Communication Convolutive Mixtures Based on Denoising IVA

In wireless communication systems, signal transmission through a channel can not avoid the influence of noise. When it reaches the receiver, it is accompanied by time delay and attenuation. Therefore, the observed mixture signals at the receiver are convolutional mixed signals with noise contamination. To solve the problem of traditional frequency-domain based convolutive blind signal separation methods have poor separation performance for convolutional mixed signals with noise, this paper propose a denoise-FastIVA method to separate convolutional mixed signals with noise. The basic principle is to use a wavelet transform to denoise the observation signal, reduce the effect of noise on the separation effect of the algorithm, and enhance the robustness of the fast fixed-point independent vector analysis (FastIVA) separation algorithm to noise. Simulation experiments show the effectiveness of a denoise-FastIVA, under the condition that the baseband signal of binary phase shift keying (BPSK) and binary frequency shift keying (2FSK) signal modulation signal is 10 bits respectively,the separation accuracy of linear frequency modulization (LFM) has increased from 87% to more than 94%; BPSK has risen from 83% to over 97%; 2FSK has improved from 81% to over 95%. When the SNR is greater than 10 dB, the separation similarity of denoise-FastIVA for the communication mixed signals is more than 90%, and the demixing signal can demodulate the baseband signal completely and correctly. Baseband signals of experimental BPSK and 2FSK signal modulation signals are 100 bits respectively. When the signal-to-noise ratio is greater than 5dB, the signal separated by denoise-FastIVA method has the highest similarity coefficient with the source signal, and the bit error rate (BER) of the separated BPSK signal and the separated 2FSK signal are the lowest, compared with the traditional frequency domain demixing method and FastIVA algorithm.

An Efficient Multistage Approach for Blind Source Separation of Noisy Convolutive Speech Mixture

This paper proposes a novel efficient multistage algorithm to extract source speech signals from a noisy convolutive mixture. The proposed approach comprises two stages named Blind Source Separation (BSS) and de-noising. A hybrid source prior model separates the source signals from the noisy reverberant mixture in the BSS stage. Moreover, we model the low- and high-energy components by generalized multivariate Gaussian and super-Gaussian models, respectively. We use Minimum Mean Square Error (MMSE) to reduce noise in the noisy convolutive mixture signal in the de-noising stage. Furthermore, the two proposed models investigate the performance gain. In the first model, the speech signal is separated from the observed noisy convolutive mixture in the BSS stage, followed by suppression of noise in the estimated source signals in the de-noising module. In the second approach, the noise is reduced using the MMSE filtering technique in the received noisy convolutive mixture at the de-noising stage, followed by separation of source signals from the de-noised reverberant mixture at the BSS stage. We evaluate the performance of the proposed scheme in terms of signal-to-distortion ratio (SDR) with respect to other well-known multistage BSS methods. The results show the superior performance of the proposed algorithm over the other state-of-the-art methods.

Energy Correlation Permutation Algorithm of Frequency-Domain Blind Source Separation Based on Frequency Bins Correction

Although blind source separation of convolutive mixtures can be efficiently solved in the frequency domain, the problem of permutation ambiguity must be solved. Thus, this paper proposes an improved permutation algorithm. Firstly, the improved algorithm uses energy correlation of the separated signal to sort the signals of each frequency bin. Then, the reliability of the sorting of the signals on the frequency bin will be judged according to the threshold. The unreliable frequency bins will be corrected in time so that we can obtain the separated signal accurately, eventually. This algorithm effectively reduces sorting errors and error propagation, thereby improving the separation effect. BSS experiments are performed on the voice signals and radar signals under different convolutive mixing models. The simulation results show that the improved algorithm has better separation performance and higher robustness than the traditional permutation algorithms, which is reflected by the increasing signal to interference ratio of separated signals.

Vision-Based Uncertainty-Aware Lane Keeping Strategy Using Deep Reinforcement Learning

Abstract Recent deep learning techniques promise high hopes for self-driving cars while there are still many issues to be addressed such as uncertainties (e.g., extreme weather conditions) in learned models. In this work, for the uncertainty-aware lane keeping, we first propose a convolutional mixture density network (CMDN) model that estimates the lateral position error, the yaw angle error, and their corresponding uncertainties from the camera vision. We then establish a vision-based uncertainty-aware lane keeping strategy in which a high-level reinforcement learning policy hierarchically modulates the reference longitudinal speed as well as the low-level lateral control. Finally, we evaluate the robustness of our strategy against the uncertainties of the learned CMDN model coming from unseen or noisy situations, as compared to the conventional lane keeping strategy without taking into account such uncertainties. Our uncertainty-aware strategy outperformed the conventional lane keeping strategy, without a lane departure in our test scenario during high-uncertainty periods with random occurrences of fog and rain situations on the road. The successfully trained deep reinforcement learning agent slows down the vehicle speed and tries to minimize the lateral error during high uncertainty situations similarly to what human drivers would do in such situations.

A RobustICA-based algorithmic system for blind separation of convolutive mixtures

In this paper, we propose a frequency-domain method employing robust independent component analysis (RICA) to address the multichannel Blind Source Separation (BSS) problem of convolutive speech mixtures in highly reverberant environments. We impose regularization processes to tackle the ill-conditioning problem of the covariance matrix and to mitigate the performance degradation. We evaluate the impact of several parameters on the performance of separation, e.g., windowing type and overlapping ratio of the frequency domain method. We then assess and compare different techniques to solve the frequency-domain permutation ambiguity. Furthermore, we develop an algorithm to separate the source signals in adverse conditions, i.e. high reverberation conditions when short observation signals are available. Finally, through extensive simulations and real-world experiments, we evaluate and demonstrate the superiority of the presented convolutive algorithmic system in comparison to other BSS algorithms, including recursive regularized ICA (RR-ICA) and independent vector analysis (IVA).

Supervised and semi-supervised deep probabilistic models for indoor positioning problems

WiFi fingerprint-based indoor localization has been a popular research topic recently. In this work, we propose two novel deep learning-based models, the convolutional mixture density recurrent neural network and the variational autoencoder-based semi-supervised learning model. The convolutional mixture density recurrent neural network is designed for indoor next location prediction, in which the advantages of convolutional neural networks, recurrent neural networks and mixture density networks are combined. Furthermore, since most of real-world WiFi fingerprint data are not labeled, we devise a variational autoencoder-based model to compute accurate user location in a semi-supervised learning manner. Finally, in order to evaluate the proposed models, we conduct the validation experiments on two real-world datasets. The final results are compared to other existing methods and verify the effectiveness of our approaches.