Sharif Digital Repository

Similarity measure is an important key in image registration. Most traditional intensity-based similarity measures (e.g., SSD, CC, MI, and CR) assume stationary image and pixel by pixel independence. Hence, perfect image registration cannot be achieved especially in presence of spatially-varying intensity distortions and outlier objects that appear in one image but not in the other. Here, we suppose that non stationary intensity distortion (such as Bias field or Outlier) has sparse representation in transformation domain. Based on this as-sumption, the zero norm (ℓ0)of the residual image between two registered images in transform domain is introduced as a new similarity measure in presence... 

Jumps are discontinuous variations in time series and with large amplitude can be considered as an extreme event. We expect the higher the jump activity to cause higher uncertainty in the stochastic behaviour of measured time series. Therefore, building statistical evidence to detect real jump seems of primary importance. In addition jump events can participate in the observed non-Gaussian feature of the increments’ (ramp up and ramp down) statistics of many time series [1]. This is the reason that most of the jump detection techniques are based on threshold values for differential of time series. There is not, however, a robust method for detection and characterisation of such discontinuous... 

Jumps are discontinuous variations in time series and with large amplitude can be considered as an extreme event. We expect the higher the jump activity to cause higher uncertainty in the stochastic behaviour of measured time series. Therefore, building statistical evidence to detect real jump seems of primary importance. In addition jump events can participate in the observed non-Gaussian feature of the increments’ (ramp up and ramp down) statistics of many time series [1]. This is the reason that most of the jump detection techniques are based on threshold values for differential of time series. There is not, however, a robust method for detection and characterisation of such discontinuous... 

In this paper we demonstrate a dynamical system simulator that runs on a single GPU. The model (running on an NVIDIA GT325M with 1GB of memory) is up to 50 times faster than a CPU version when more than 10 million adaptive Hopf oscillators have been simulated. The simulation shows that the oscillators tune to the correct frequencies for both discrete and continuous spectra. Due to its dynamic nature the system is also capable to track non-stationary spectra. With the help of this model the frequency spectrum of an ECG signal (as a non-stationary signal) obtained and was showed that frequency domain representation of signal (i.e. FFT) is the same as one MATLAB generates  

Similarity measure is an important part of image registration. The main challenge of similarity measure is lack of robustness to different distortions. A well-known distortion is spatially-varying intensity distortion. Its main characteristic is correlation among pixels. Most traditional intensity based similarity measures (e.g., SSD, MI) assume stationary image and pixel to pixel independence. Hence, these similarity measures are not robust against spatially-varying intensity distortion. Here, we suppose that non-stationary intensity distortion has a sparse representation in transform domain, i.e. its distribution has high peak at origin and a long tail. We use two viewpoints of Maximum... 

This paper presents a new approach to sensor placement strategy in emitter localization problem which is based on Time Difference of Arrival (TDOA) measurements. The advantage of this method is its flexibility and capability of positioning sensors in constrained or non-stationary situations in which the positions of the sensors are restricted to certain portions of the space and/or needed to be changed repeatedly. The validity of the proposed algorithm is assessed by three different simulation scenarios and the results verify its proper operation  

This article introduces a new time-varying model to generate synthetic non-stationary acceleration records using the modified Kanaii-Tajimi model with time-variant parameters. The proposed method can capture two different dominant frequencies per time which makes it suitable for synthesizing the near-field no-pulse earthquake records. A number of closed-form relationships are developed to describe the frequency dependent time-domain level crossings of the simulated records under white noise excitation. The model parameters are optimized using the crossing and Arias intensity properties of the synthetic and target records. The efficiency of the proposed model is demonstrated using a data base... 

This study presents a new approach to sensor placement strategy in emitter localisation problems based on time difference of arrival measurements. The method addresses a flexible procedure which is capable of positioning the sensors in constrained environments or non-stationary situations where the positions of the sensors are restricted to certain parts of the space and/or need to be changed repeatedly. This method is sequential and has lower computation burden compared to other methods. The validity of the proposed algorithm is assessed by many different numerical scenarios and the results verify its proper operation. © The Institution of Engineering and Technology 2017  

We describe a general method for analyzing a nonstationary stochastic process X (t) which, unlike many of the previous analysis methods, does not require X (t) to have any scaling feature. The method is used to study the fluctuations in the daily price of oil. It is shown that the returns time series, y (t) =ln [X (t+1) X (t)], is a stationary and Markov process, characterized by a Markov time scale tM. The coefficients of the Kramers-Moyal expansion for the probability density function P (y,t y0, t0) are computed. P (y,t, y0, t0) satisfies a Fokker-Planck equation, which is equivalent to a Langevin equation for y (t) that provides quantitative predictions for the oil price over times that... 

This paper presents a novel HMM-based speech enhancement framework based on Laplace and Gaussian distributions in DCT domain. We propose analytical procedures for training clean speech and noise models with the aim of Baum's auxiliary function and present two MMSE estimators based on Gaussian-Gaussian (for clean speech and noise respectively) and Laplace-Gaussian combinations in the HMM framework. The performance evaluation is done using SNR and PESQ measures and the results of the proposed techniques are compared with AR-HMM approach. Higher SNR improvement is achieved for the proposed method in the Gaussian-Gaussian case in comparison with AR-HMM and Laplace-Gaussian techniques for both... 

In this letter, a new adaptive Gaussian process (GP) frame work for signal extrapolation is proposed. Signal extrapolation is an essential task in many applications such as concealment of corrupted data in image and video communications. While possessing many interesting properties, Gaussian process priors with inappropriate stationary kernels may create extremely blurred edges in concealed areas of the image. To address this problem, we propose adaptive non-stationary kernels in a Gaussian process framework. The proposed adaptive kernel functions are defined based on the hypothesized edges of the missing areas. Experimental results verify the effectiveness of the proposed method compared to... 

This paper introduces an efficient stochastic method to produce fully nonstationary records having multiple peaks in power spectrum. The zero-crossing characteristics of the acceleration, velocity, and displacement time histories of the output signal are used to estimate the model parameters. This procedure is utilized to resimulate 252 non-pulse-like, horizontal near-field records with rupture distance of less than 10 km and strike–slip mechanism. The model parameters are regressed against moment magnitude, rupture distance, hypo-central depth, and shear wave velocity, enabling the scenario-based simulation of the records. The response spectra of the simulated records are compared with... 

Time-frequency and time-scale transforms are one of the powerful mathematical methods for feature extraction of non-stationary signals such as partial discharge (PD) signals. Modified Fourier-based transforms and Wavelet transforms are well known among them. PD signals can be analyzed by these influential methods but normally with some serious limitations and inadequacies. This paper proposes a new approach based on Wavelet-Hilbert transform for the analysis of electrical PD signals. The mathematical model of a Wavelet-Hilbert transform is described. Contemplating of natural behavior of PD, that is, stochastic and non-stationary an artificial multi-source of PD is simulated in a Detail model... 

This study introduces a novel identification method for recognition of nuclear power plants (NPPs) transients by combining the autoregressive integrated moving-average (ARIMA) model and the neural network with error back-propagation (EBP) learning algorithm. The proposed method consists of three steps. First, an EBP based identifier is adopted to distinguish the plant normal states from the faulty ones. In the second step, ARIMA models use integrated (I) process to convert non-stationary data of the selected variables into stationary ones. Subsequently, ARIMA processes, including autoregressive (AR), moving-average (MA), or autoregressive moving-average (ARMA) are used to forecast time... 

Data streams are endless flow of data produced in high speed, large size and usually non-stationary environments. The main property of these streams is the occurrence of concept drifts. Using decision trees is shown to be a powerful approach for accurate and fast learning of data streams. In this paper, we present an incremental regression tree that can predict the target variable of newly incoming instances. The tree is updated in the case of occurring concept drifts either by altering its structure or updating its embedded models. Experimental results show the effectiveness of our algorithm in speed and accuracy aspects in comparison to the best state-of-the-art methods  

This paper studies the automatic noise recognition problem based on RBF and MLP neural networks classifiers using linear predictive and Mel-frequency cepstral coefficients (LPC and MFCC). We first briefly review the architecture of each network as automatic noise recognition (ANR) approach, then, compare them to each other and investigate factors and criteria that influence final recognition performance. The proposed networks are evaluated on 15 stationary and non-stationary types of noises with frame length of 20 ms in term of correct classification rate. The results demonstrate that the MLP network using LPCs is a precise ANR with accuracy rate of 99.9%, while the RBF network with MFCCs... 

This paper presents a method for ECG characteristic points detection based on Wavelet Transform (WT). Wavelet Transform leads to more accurate results in analyzing nonstationary signals such as ECG. The selected wavelet is quadratic spline wavelet. Using quadratic spline mother wavelet, a method for detection of QRS complex and T and P waves is presented and evaluated with the help of MIT-BIH Arrhythmia database. Experimental results show excellent performance of the proposed method. © 2009 IEEE  

The problem of finding the optimal node density for reconstructing a stochastic signal from its noisy samples in sensor networks is considered. The signal could be nonstationary and nonbandlimited. A weight is assigned to each location that indicates the relative importance of the signal at that location. It is shown that when the number of samples is very large, the optimal density of the samples at each location is proportional to the square root of the weight associated to that location  

This study aims to show the efficiency of a proposed spectral matching technique for the reduction of required ground motions in the dynamic time history analysis. In this non-stationary spectral matching approach, unconstrained optimization is employed to adjust the signal to match a target spectrum. Adjustment factors of discrete wavelet transform (DWT) coefficients associated with the signals are then considered as decision variables and the Levenberg-Marquardt algorithm is employed to find the optimum values of DWT coefficients. This matching algorithm turns out to be quite effective in the spectral matching objective, where matching at multiple damping ratios can be readily achieved....