An effective nonlinear fourth-order partial differential equation (PDE) - based additive denoising technique is considered in this article. The proposed image restoration approach is based on a hybrid anisotropic diffusion model with boundary conditions, which combines nonlinear second and fourth order diffusion-based components and a conventional filter kernel. The obtained PDE model is well-posed under some certain assumptions and its weak unique solution is computed by applying a consistent and fast-converging finite difference method-based numerical approximation scheme that is also proposed here. The image denoising results produced by this method are finally described.
Voice activity detection (VAD), namely determining whether a speech signal is active or inactive, and single talk detector (STD), namely detecting that only one speaker is active, are important building blocks in many speech processing applications. A speaker-localization stage (such as the steered response power (SRP)) is often concurrently implemented on the same device.In this paper, the spatial properties of the SRP are utilized for improving the performance of both the voice activity detector (VAD) and the STD. We propose to measure the entropy at the SRP output and compare with the typical entropy of noise-only frames. This feature utilizes spatial information and may therefore become advantageous in nonstationary noise environments. The STD can then be implemented by determining local minimum values of the entropy measure of the SRP.The proposed VAD was tested for a single speaker with two cases, directional background noise with changing level and with a background music source. The proposed STD was tested using real recordings of two concurrent speakers.
The difference-to-sum power ratio was proposed and used to suppress wind noise under specific acoustic conditions. In this contribution, a general formulation of the difference-to-sum power ratio associated with a mixture of speech and wind noise is proposed and analyzed. In particular, it is assumed that the complex coherence of convective turbulence can be modelled by the Corcos model. In contrast to the work in which the power ratio was first presented, the employed Corcos model holds for every possible air stream direction and takes into account the lateral coherence decay rate. The obtained expression is subsequently validated with real data for a dual microphone set-up. Finally, the difference-to-sum power ratio is exploited as a spatial feature to indicate the frame-wise presence of wind noise, obtaining improved detection performance when compared to an existing multi-channel wind noise detection approach.
In this study, we aim to classify comments as abusive or non-abusive. We develop a Hebrew corpus of user comments annotated for abusive language. Then, we investigate highly sparse n-grams representations as well as denser character n-grams representations for comment abuse classification. Since the comments in social media are usually short, we also investigate four dimension reduction methods, which produce word vectors that collapse similar words into groups. We show that the character n-grams representations outperform all the other representation for the task of identifying abusive comments.
This paper considers a problem of lossy compression of generalized Gaussian (GG) sources (i.e., sources with the probability density functions proportional to $\mathrm {e}^{-\frac {|x|^{\mathrm {S}}}{2}}$, $s \gt 0$) with an $\ell _{r}, r \gt 0$, distortion measure.It is shown that an optimal reconstruction distribution always exists and properties of this distribution are studied. In particular, it is shown that if $s \leq r-1$ then an optimal reconstruction must have unbounded support and for $s \gt r$ an optimal reconstruction must have bounded support. Further, it is shown that Shannon’s lower bound is achievable if and only if $r = s \in (0,1$] $\cup \{2\}$, or in other words when the GG distribution is self-decomposable. Finally, conditions are shown under which an optimal reconstruction is discrete with finitely many mass points.
We consider a state-dependent parallel Gaussian channel with independent states and a common cognitive helper, in which two transmitters wish to send independent information to their corresponding receivers over two parallel subchannels. Each channel is corrupted by independent additive Gaussian state. The states are not known to the transmitters nor to the receivers, but known to a helper in a noncausal manner. The helper's goal is to assist a reliable communication by mitigating the state. Outer and inner bounds are derived and segments of the capacity region is characterized for various channel parameters.
Presented here is a new approach for analysis of the so-called holey photonic crystals—a class of electro-optical components, in which periodicity of air holes in dielectric media is used for confinement of light. This class includes several kinds of microstructured fibers, semiconductor lasers etc. Accurate evaluation of optical characteristics of those devices is usually a complicated problem due to the large dimensions and the fine structure of their refractive index distribution. Furthermore, usually, only numerical solutions for this class of optical components are available. The overwhelming majority of the physical models, suitable for analysis of holey photonic devices, proceed from the “natural” assumption: the devices are considered as arrays of air holes, surrounded by dielectric material. In this work we propose another model. Namely, we treat them as arrays of dielectric spots (waveguides), embedded in the air (cladding material). This model allows utilization of the extended coupled-mode theory (a relatively new approach designed for analysis of infinite arrays of coupled waveguides and previously considered inapplicable to holey optical components) for calculations of the latter. In this sense, we present a new method for analysis of holey photonic crystals. On the one hand, our method allows analytical evaluation of some optical characteristics of holey optical components (such as the number of photonic bands and bandwidth). On the other hand, accurate numerical computation of the photonic band structure of the holey photonic devices, incorporating a large number of holes, can be done with this technique on a timescale of several minutes.
Deep Neural Networks (DNN), contain multiple convolutional and several fully connected layers, require considerable hardware resources to train in a reasonable time. Multiple CPUs, GPUs or FPGAs are usually combined to reduce the training time of a DNN. However, many individuals or small organizations do not possess the resources to obtain multiple hardware units. The contribution of this work is two-fold. First, we present an implementation of a distributed DNN training system that uses multiple small (wimpy) nodes to accelerate the training process. The nodes are mobile smartphone devices, with variable hardware specifications. All DNN training tasks are performed on the small nodes, coordinated by a centralized server. Second, we propose a novel method to mitigate issues arising from the variability in hardware resources. We demonstrate that the method allows training a DNN to high accuracy on known image recognition datasets with multiple small different nodes. The proposed method factors in the contribution from each node according to its run time on a specific training task, relative to the other nodes. In addition, we discuss practical challenges that arise from small node system and suggest several solutions.
A common pre-possessing task in machine learning is to complete missing data entries in order to form a full dataset. In case the dimension of the input data is high, it is often the case that the rows and columns are correlated. In this work, we construct a multi-scale model that is based on the the dual row-column geometry of the dataset and apply it to imputation, which is carried out within the model construction. Experimental results demonstrate the efficiency of our approach on a publicly available dataset.
CMOS Silicon Photomultiplier (SiPM) composed of a mosaic array of SPADs (single photon avalanche diode in Geiger Mode) combined in parallel, is the building block of optical radars based on a LIDAR (light detection and ranging). An open essential design parameter is the required number of sub-pixels for adequate detection of a packet of n photons, taking into consideration that each sub-pixel, composed of a single SPAD, can detect only the first photon. This study evaluates this design parameter based on a stochastic approach, where the random number of incident photons as well as the detection probability of each SPAD is taken into consideration. An expression for Signal to Noise Ratio (SNR) is developed, yielding the optimal number of sub-pixels required for a practical implementation of a LIDAR.
We present a model of a one-dimensional chain of two-level artificial atoms driven simultaneously with a dc field and a quantum light in the strong coupling regime. It is shown that the entanglement of the electron and photon (dressing of the atoms with light) dramatically changes the scenario of the Bloch oscillations (BO) compared with standard solution of the Bloch-Zener model. We considered the mutual influence of dressing and BO and show that the quantum properties of light become controllable via an adiabatic dc field tuning. The obtained results open new ways in quantum state engineering, nano-photonic spectroscopy and nano-antennas.
Capturing of videos from the television (TV) screens or from the theater screens by using the mobile cameras and its illegal distribution through video-sharing websites like YouTube, Dailymotion, Metacafe, etc. is a well-known challenge faced by the film industry. The video-sharing websites like YouTube does not encourage the illegal distribution of videos (without proper consent from the content owner). Currently, the YouTube has a facility to remove an illegally distributed video content from its video repository based on the request from the content owner. In general, the removal of an illegally distributed video may take a few days, hence during this period, the video may be downloaded by many of the people. The downloaded videos may be again distributed over the internet through different modes. This paper proposed a new technique which will classify a given video into normal video or screen captured video and it can be incorporated with video-sharing websites to prevent the illegal distribution of screen captured videos. The proposed scheme uses a support vector machine model which is trained using no-reference image quality measures. As far as our knowledge is concerned, there is no related work in this area.
In this paper we propose a switching controller for a class of MIMO bilinear systems with constant delays in both the state and the input. The motivation to consider such a class of systems is that, in the scalar case, it has shown to be suitable for modelling and controller design of some turbulent flow control systems.
A proposed voltage-dependent-capacitor tuning method for WPT systems was studied analytically, by simulation and verified experimentally. The circuit implementation applies common ferroelectric ceramic capacitors as voltage dependent elements. The experimental results on a mockup of autonomous mini-submarine charging, suggest that commercial ceramic capacitors are a viable option for tuning WPT.
Classifying the degree of Parkinson's disease is an important clinical necessity. Nonetheless, current methodology requires manual (and subjective) evaluation by a trained clinical expert. Recently, Machine Learning tools have been developed that can produce a classification of the presence of PD directly from the speech signal in an automated and objective fashion. However, these methods were not sufficient for the classification of the degree of the disease. In this work, we show how to apply and leverage topological information on the both the label space and the feature space of the speech signal in order to solve this problem. We address the problem by performing topological clustering (using a version of the Kohonen Self Organizing Map algorithm) of the feature space and then optimizing separate multi-class classifiers on each cluster. Using these methods, we can reliably train our system to classify new speech signal data to more than the 70% level on a 7 degree classification (where random level is 14%) which is close to the obtainable accuracy on the simple 2 class classification.
This paper presents progress in research on a novel oscillator architecture based on a current-starved ring oscillator and ring-connected transmission lines. Additional simulations of the proposed oscillator architecture and details regarding the construction and measurement of the prototype device are outlined. The results verify both the hysteretic behavior and the improved noise performance of the novel oscillator.
Electrification of Israel railroads is a successful project being realized during present time. Among the well-known advantages of electric train exploitation are high energetic efficiency, regenerative breaking, diminished pollution, flexibility of traffic control and improved transportation dynamics. However, the use of electric train is accompanied with some specific hurdles caused by significantly stochastic power flows. While in transit between stations an electric train has different phases of movement, during some consuming energy, but from time to time throughout braking processes they become generators transmitting energy back into the electric grid. This circumstance may produce voltage instability in distribution lines, which should be kept inside permissible limits to allow normal operation of electric equipment. Tap-changers on distribution transformers considered for a voltage regulation have sluggish response of ~7-8 sec. Therefore, fast changes of power stream can cause significant voltage deviations. If the voltage overcome explicit allowable level a trigger of a protection system is engendered disconnecting vehicles from a grid. The last causes extremely dangerous detriments and should be maximally prevented.Rapid and considerable changes in train power consumption, in addition may cause network frequency instability which in turn can violate electricity supply.Present article provides an original examination of the future stability of network frequency and distribution voltage which may be assumed when 420 km of Israel railroads will be electrified in the following 5–10 years. For such study special algorithm based on simulation approach was developed and applied. Predicting scenarios for frequency and distribution lines voltage control are represented below in the following text.
We consider an extension of the information bottleneck problem where underlying Markov Chain is $X-0-(Y,\ S)-0-Z$, and where $P_{X,S,Y}=P_{X}P_{S}P_{Y|X,S}$ is the joint distribution of a source X, a channel state S independent of the source, and the channel output Y of a state-dependent channel. For the case $Y=SX+N$ with $X, S$ and N Gaussian circularly symmetric, we provide an upper bound and two achievable lower bounds on the information bottleneck rate. We relate this problem to the case of an oblivious relay with channel state information. Our results show that simple symbol-by-symbol relay processing, possibly followed by “entropy coding” (data compression) yields a very effective method, virtually achieving the upper bound on a wide range of relevant system parameters.
Standard Coupled-Mode Theory (Standard CMT) is a well-known approach for analysis of coupling and propagation of guided modes in multiwaveguide systems. In order to analyze propagation of EM fields these systems, Standard CMT solves 1st order differential matrix equation (Standard CMT equation). Analytical solution for this equation currently exists only for the multiwaveguide systems with dielectric function, homogeneous along the optical axis (z-axis). Coupled-mode analysis of the devices, whose dielectric function varies along the optical axis (e.g. photonic components with integrated optical gratings) till now, is only available with numerical techniques. In this work, we propose the general Analytical solution for Standard CMT equation, including the case of dielectric function, inhomogeneous in the z-direction. This solution represents an effective analytical tool for fast and accurate analysis, design and optimization of a variety of photonic components, whose principles of operation are based on the variation of their dielectric function along the optical axis.
In recent years, autonomous marine systems have increasingly been used for civilian and military purposes. Usually, the platforms act as independent units, which limits their usefulness. The main objective of this research is to enhance the capabilities of autonomous marine vehicles either autonomous underwater vehicles (AUVs) or autonomous surface vehicles (ASVs) - by allowing cooperation between them. In this paper, we examine a multi-variable system utilizing autonomous nonlinear vehicles whose coupled dynamic effects depend on complicated hydrodynamics, generalized forces and variables, and limited communications. A simulator based on the Robotic Operation System (ROS) with GAZEBO as the simulation environment was employed. The simulator models the dynamics of the autonomous vehicles and the behavior of the sensors, and it is used for testing and evaluating the cooperation algorithms before and after performing actual trials at sea.
