Search results for: Laplacian of Gaussian (LoG) method

Authors: Raghad Jawad Ahmed

Abstract:

Segmentation of a color image composed of different kinds of regions can be a hard problem, namely to compute for an exact texture fields. The decision of the optimum number of segmentation areas in an image when it contains similar and/or un stationary texture fields. A novel neighborhood-based segmentation approach is proposed. A genetic algorithm is used in the proposed segment-pass optimization process. In this pass, an energy function, which is defined based on Markov Random Fields, is minimized. In this paper we use an adaptive threshold estimation method for image thresholding in the wavelet domain based on the generalized Gaussian distribution (GGD) modeling of sub band coefficients. This method called Normal Shrink is computationally more efficient and adaptive because the parameters required for estimating the threshold depend on sub band data energy that used in the pre-stage of segmentation. A quad tree is employed to implement the multi resolution framework, which enables the use of different strategies at different resolution levels, and hence, the computation can be accelerated. The experimental results using the proposed segmentation approach are very encouraging.

Keywords: Color image segmentation, Genetic algorithm, Markov random field, Scale space filter.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1548

Authors: Mohammad H. Fattahi

Abstract:

Chaotic analysis has been performed on the river flow time series before and after applying the wavelet based de-noising techniques in order to investigate the noise content effects on chaotic nature of flow series. In this study, 38 years of monthly runoff data of three gauging stations were used. Gauging stations were located in Ghar-e-Aghaj river basin, Fars province, Iran. Noise level of time series was estimated with the aid of Gaussian kernel algorithm. This step was found to be crucial in preventing removal of the vital data such as memory, correlation and trend from the time series in addition to the noise during de-noising process.

Keywords: Chaotic behavior, wavelet, noise reduction, river flow.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2057

Authors: A. Harras, A. Tsuji, K. Terada

Abstract:

Tracking of moving people has gained a matter of great importance due to rapid technological advancements in the field of computer vision. The objective of this study is to design a motion based detection and tracking multiple walking pedestrians randomly in different directions. In our proposed method, Gaussian mixture model (GMM) is used to determine moving persons in image sequences. It reacts to changes that take place in the scene like different illumination; moving objects start and stop often, etc. Background noise in the scene is eliminated through applying morphological operations and the motions of tracked people which is determined by using the Kalman filter. The Kalman filter is applied to predict the tracked location in each frame and to determine the likelihood of each detection. We used a benchmark data set for the evaluation based on a side wall stationary camera. The actual scenes from the data set are taken on a street including up to eight people in front of the camera in different two scenes, the duration is 53 and 35 seconds, respectively. In the case of walking pedestrians in close proximity, the proposed method has achieved the detection ratio of 87%, and the tracking ratio is 77 % successfully. When they are deferred from each other, the detection ratio is increased to 90% and the tracking ratio is also increased to 79%.

Keywords: Automatic detection, tracking, pedestrians.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 801

Authors: John N. Ellinas

Abstract:

In this paper, a robust watermarking algorithm using the wavelet transform and edge detection is presented. The efficiency of an image watermarking technique depends on the preservation of visually significant information. This is attained by embedding the watermark transparently with the maximum possible strength. The watermark embedding process is carried over the subband coefficients that lie on edges, where distortions are less noticeable, with a subband level dependent strength. Also, the watermark is embedded to selected coefficients around edges, using a different scale factor for watermark strength, that are captured by a morphological dilation operation. The experimental evaluation of the proposed method shows very good results in terms of robustness and transparency to various attacks such as median filtering, Gaussian noise, JPEG compression and geometrical transformations.

Keywords: Watermarking, wavelet transform, edge detection.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2308

Authors: Haifeng Wang, Haili Zhang

Abstract:

Most movie recommendation systems have been developed for customers to find items of interest. This work introduces a predictive model usable by small and medium-sized enterprises (SMEs) who are in need of a data-based and analytical approach to stock proper movies for local audiences and retain more customers. We used classification models to extract features from thousands of customers’ demographic, behavioral and social information to predict their movie genre preference. In the implementation, a Gaussian kernel support vector machine (SVM) classification model and a logistic regression model were established to extract features from sample data and their test error-in-sample were compared. Comparison of error-out-sample was also made under different Vapnik–Chervonenkis (VC) dimensions in the machine learning algorithm to find and prevent overfitting. Gaussian kernel SVM prediction model can correctly predict movie genre preferences in 85% of positive cases. The accuracy of the algorithm increased to 93% with a smaller VC dimension and less overfitting. These findings advance our understanding of how to use machine learning approach to predict customers’ preferences with a small data set and design prediction tools for these enterprises.

Keywords: Computational social science, movie preference, machine learning, SVM.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1609

Authors: J. P. Dubois, Omar M. Abdul-Latif

Abstract:

Support Vector Machine (SVM) is a statistical learning tool developed to a more complex concept of structural risk minimization (SRM). In this paper, SVM is applied to signal detection in communication systems in the presence of channel noise in various environments in the form of Rayleigh fading, additive white Gaussian background noise (AWGN), and interference noise generalized as additive color Gaussian noise (ACGN). The structure and performance of SVM in terms of the bit error rate (BER) metric is derived and simulated for these advanced stochastic noise models and the computational complexity of the implementation, in terms of average computational time per bit, is also presented. The performance of SVM is then compared to conventional binary signaling optimal model-based detector driven by binary phase shift keying (BPSK) modulation. We show that the SVM performance is superior to that of conventional matched filter-, innovation filter-, and Wiener filter-driven detectors, even in the presence of random Doppler carrier deviation, especially for low SNR (signal-to-noise ratio) ranges. For large SNR, the performance of the SVM was similar to that of the classical detectors. However, the convergence between SVM and maximum likelihood detection occurred at a higher SNR as the noise environment became more hostile.

Keywords: Colour noise, Doppler shift, innovation filter, least square-support vector machine, matched filter, Rayleigh fading, Wiener filter.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1775

Authors: Cheima Ben Soltane, Ittansa Yonas Kelbesa

Abstract:

Speaker Identification (SI) is the task of establishing identity of an individual based on his/her voice characteristics. The SI task is typically achieved by two-stage signal processing: training and testing. The training process calculates speaker specific feature parameters from the speech and generates speaker models accordingly. In the testing phase, speech samples from unknown speakers are compared with the models and classified. Even though performance of speaker identification systems has improved due to recent advances in speech processing techniques, there is still need of improvement. In this paper, a Closed-Set Tex-Independent Speaker Identification System (CISI) based on a Multiple Classifier System (MCS) is proposed, using Mel Frequency Cepstrum Coefficient (MFCC) as feature extraction and suitable combination of vector quantization (VQ) and Gaussian Mixture Model (GMM) together with Expectation Maximization algorithm (EM) for speaker modeling. The use of Voice Activity Detector (VAD) with a hybrid approach based on Short Time Energy (STE) and Statistical Modeling of Background Noise in the pre-processing step of the feature extraction yields a better and more robust automatic speaker identification system. Also investigation of Linde-Buzo-Gray (LBG) clustering algorithm for initialization of GMM, for estimating the underlying parameters, in the EM step improved the convergence rate and systems performance. It also uses relative index as confidence measures in case of contradiction in identification process by GMM and VQ as well. Simulation results carried out on voxforge.org speech database using MATLAB highlight the efficacy of the proposed method compared to earlier work.

Keywords: Feature Extraction, Speaker Modeling, Feature Matching, Mel Frequency Cepstrum Coefficient (MFCC), Gaussian mixture model (GMM), Vector Quantization (VQ), Linde-Buzo-Gray (LBG), Expectation Maximization (EM), pre-processing, Voice Activity Detection (VAD), Short Time Energy (STE), Background Noise Statistical Modeling, Closed-Set Tex-Independent Speaker Identification System (CISI).

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1834

Authors: Xia Liu, Marek E. Bialkowski

Abstract:

This paper reports on investigations into capacity of a Multiple Input Multiple Output (MIMO) wireless communication system employing a uniform linear array (ULA) at the transmitter and either a uniform linear array (ULA) or a uniform circular array (UCA) antenna at the receiver. The transmitter is assumed to be surrounded by scattering objects while the receiver is postulated to be free from scattering objects. The Laplacian distribution of angle of arrival (AOA) of a signal reaching the receiver is postulated. Calculations of the MIMO system capacity are performed for two cases without and with the channel estimation errors. For estimating the MIMO channel, the scaled least square (SLS) and minimum mean square error (MMSE) methods are considered.

Keywords: MIMO, channel capacity, channel estimation, ULA, UCA, spatial correlation

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1332

Authors: Farah Haroon, Haroon Rasheed, Kazi M Ahmed

Abstract:

IEEE 802.15.4a impulse radio-time hopping ultra wide band (IR-TH UWB) physical layer, due to small duty cycle and very short pulse widths is robust against multipath propagation. However, scattering and reflections with the large number of obstacles in indoor channel environments, give rise to dense multipath fading. It imposes serious problem to optimum Rake receiver architectures, for which very large number of fingers are needed. Presence of strong noise also affects the reception of fine pulses having extremely low power spectral density. A robust SRake receiver for IEEE 802.15.4a IRTH UWB in dense multipath and additive white Gaussian noise (AWGN) is proposed to efficiently recover the weak signals with much reduced complexity. It adaptively increases the signal to noise (SNR) by decreasing noise through a recursive least square (RLS) algorithm. For simulation, dense multipath environment of IEEE 802.15.4a industrial non line of sight (NLOS) is employed. The power delay profile (PDF) and the cumulative distribution function (CDF) for the respective channel environment are found. Moreover, the error performance of the proposed architecture is evaluated in comparison with conventional SRake and AWGN correlation receivers. The simulation results indicate a substantial performance improvement with very less number of Rake fingers.

Keywords: Adaptive noise cancellation, dense multipath propoagation, IEEE 802.15.4a, IR-TH UWB, industrial NLOS environment, SRake receiver

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1786

Authors: Liu Dan, Wang Xue, Wang Guiqin, Qian Zhihong

Abstract:

The burst noise is a kind of noises that are destructive and frequently found in semiconductor devices and ICs, yet detecting and removing the noise has proved challenging for IC designers or users. According to the properties of burst noise, a methodological approach is presented (proposed) in the paper, by which the burst noise can be analysed and detected in time domain. In this paper, principles and properties of burst noise are expounded first, Afterwards, feasibility (viable) of burst noise detection by means of wavelet transform in the time domain is corroborated in the paper, and the multi-resolution characters of Gaussian noise, burst noise and blurred burst noise are discussed in details by computer emulation. Furthermore, the practical method to decide parameters of wavelet transform is acquired through a great deal of experiment and data statistics. The methodology may yield an expectation in a wide variety of applications.

Keywords: Burst noise, detection, wavelet transform

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1867

Authors: Nazif Çalış, Murat Erişoğlu, Hamza Erol, Tayfun Servi

Abstract:

In the recent works related with mixture discriminant analysis (MDA), expectation and maximization (EM) algorithm is used to estimate parameters of Gaussian mixtures. But, initial values of EM algorithm affect the final parameters- estimates. Also, when EM algorithm is applied two times, for the same data set, it can be give different results for the estimate of parameters and this affect the classification accuracy of MDA. Forthcoming this problem, we use Self Organizing Mixture Network (SOMN) algorithm to estimate parameters of Gaussians mixtures in MDA that SOMN is more robust when random the initial values of the parameters are used [5]. We show effectiveness of this method on popular simulated waveform datasets and real glass data set.

Keywords: Self Organizing Mixture Network, MixtureDiscriminant Analysis, Waveform Datasets, Glass Identification, Mixture of Multivariate Normal Distributions

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1484

Authors: A. Javed, K. Djidjeli, J. T. Xing

Abstract:

The concept of adaptive shape parameters (ASP) has been presented for solution of incompressible Navier Strokes equations using mesh-free local Radial Basis Functions (RBF). The aim is to avoid ill-conditioning of coefficient matrices of RBF weights and inaccuracies in RBF interpolation resulting from non-optimized shape of basis functions for the cases where data points (or nodes) are not distributed uniformly throughout the domain. Unlike conventional approaches which assume globally similar values of RBF shape parameters, the presented ASP technique suggests that shape parameter be calculated exclusively for each data point (or node) based on the distribution of data points within its own influence domain. This will ensure interpolation accuracy while still maintaining well conditioned system of equations for RBF weights. Performance and accuracy of ASP technique has been tested by evaluating derivatives and laplacian of a known function using RBF in Finite difference mode (RBFFD), with and without the use of adaptivity in shape parameters. Application of adaptive shape parameters (ASP) for solution of incompressible Navier Strokes equations has been presented by solving lid driven cavity flow problem on mesh-free domain using RBF-FD. The results have been compared for fixed and adaptive shape parameters. Improved accuracy has been achieved with the use of ASP in RBF-FD especially at regions where larger gradients of field variables exist.

Keywords: CFD, Meshless Particle Method, Radial Basis Functions, Shape Parameters

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2790

Authors: Ratchakit Sakuldee, Somkait Udomhunsakul

Abstract:

Measurement of the quality of image compression is important for image processing application. In this paper, we propose an objective image quality assessment to measure the quality of gray scale compressed image, which is correlation well with subjective quality measurement (MOS) and least time taken. The new objective image quality measurement is developed from a few fundamental of objective measurements to evaluate the compressed image quality based on JPEG and JPEG2000. The reliability between each fundamental objective measurement and subjective measurement (MOS) is found. From the experimental results, we found that the Maximum Difference measurement (MD) and a new proposed measurement, Structural Content Laplacian Mean Square Error (SCLMSE), are the suitable measurements that can be used to evaluate the quality of JPEG200 and JPEG compressed image, respectively. In addition, MD and SCLMSE measurements are scaled to make them equivalent to MOS, given the rate of compressed image quality from 1 to 5 (unacceptable to excellent quality).

Keywords: JPEG, JPEG2000, objective image quality measurement, subjective image quality measurement, correlation coefficients.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2137

Authors: Iman Elyasi, Sadegh Zarmehi

Abstract:

Due to some reasons, observed images are degraded which are mainly caused by noise. Recently image denoising using the wavelet transform has been attracting much attention. Waveletbased approach provides a particularly useful method for image denoising when the preservation of edges in the scene is of importance because the local adaptivity is based explicitly on the values of the wavelet detail coefficients. In this paper, we propose several methods of noise removal from degraded images with Gaussian noise by using adaptive wavelet threshold (Bayes Shrink, Modified Bayes Shrink and Normal Shrink). The proposed thresholds are simple and adaptive to each subband because the parameters required for estimating the threshold depend on subband data. Experimental results show that the proposed thresholds remove noise significantly and preserve the edges in the scene.

Keywords: Image denoising, Bayes Shrink, Modified Bayes Shrink, Normal Shrink.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2427

Authors: M. T. Sichani, B. J. Pedersen, S. R. K. Nielsen

Abstract:

Turbulence of the incoming wind field is of paramount importance to the dynamic response of civil engineering structures. Hence reliable stochastic models of the turbulence should be available from which time series can be generated for dynamic response and structural safety analysis. In the paper an empirical cross spectral density function for the along-wind turbulence component over the wind field area is taken as the starting point. The spectrum is spatially discretized in terms of a Hermitian cross-spectral density matrix for the turbulence state vector which turns out not to be positive definite. Since the succeeding state space and ARMA modelling of the turbulence rely on the positive definiteness of the cross-spectral density matrix, the problem with the non-positive definiteness of such matrices is at first addressed and suitable treatments regarding it are proposed. From the adjusted positive definite cross-spectral density matrix a frequency response matrix is constructed which determines the turbulence vector as a linear filtration of Gaussian white noise. Finally, an accurate state space modelling method is proposed which allows selection of an appropriate model order, and estimation of a state space model for the vector turbulence process incorporating its phase spectrum in one stage, and its results are compared with a conventional ARMA modelling method.

Keywords: Turbulence, wind turbine, complex coherence, state space modelling, ARMA modelling.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1607

Authors: Walid S. Alfuhaid, Saud A. Alghamdi, John M. Watkins, M. Edwin Sawan

Abstract:

Designing a controller for stochastic decentralized interconnected large scale systems usually involves a high degree of complexity and computation ability. Noise, observability, and controllability of all system states, connectivity, and channel bandwidth are other constraints to design procedures for distributed large scale systems. The quasi-steady state model investigated in this paper is a reduced order model of the original system using singular perturbation techniques. This paper results in an optimal control synthesis to design an observer based feedback controller by standard stochastic control theory techniques using Linear Quadratic Gaussian (LQG) approach and Kalman filter design with less complexity and computation requirements. Numerical example is given at the end to demonstrate the efficiency of the proposed method.

Keywords: Decentralized, optimal control, output, singular perturb.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1537

Authors: S. Safi, M. Frikel, M. M'Saad, A. Zeroual

Abstract:

This paper describes a blind algorithm for estimating a time varying and frequency selective fading channel. In order to identify blindly the impulse response of these channels, we have used Higher Order Statistics (HOS) to build our algorithm. In this paper, we have selected two theoretical frequency selective channels as the Proakis-s 'B' channel and the Macchi-s channel, and one practical frequency selective fading channel called Broadband Radio Access Network (BRAN A). The simulation results in noisy environment and for different data input channel, demonstrate that the proposed method could estimate the phase and magnitude of these channels blindly and without any information about the input, except that the input excitation is i.i.d (Identically and Independent Distributed) and non-Gaussian.

Keywords: Frequency response, system identification, higher order statistics, communication channels, phase estimation.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1794

Authors: Yilun Shang

Abstract:

We consider a heterogeneously mixing SIR stochastic epidemic process in populations described by a general graph. Likelihood theory is developed to facilitate statistic inference for the parameters of the model under complete observation. We show that these estimators are asymptotically Gaussian unbiased estimates by using a martingale central limit theorem.

Keywords: statistic inference, maximum likelihood, epidemicmodel, heterogeneous mixing.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1375

Authors: Mourad Moussa, Ali Douik, Hassani Messaoud

Abstract:

In this paper, we present a comparative study between two computer vision systems for objects recognition and tracking, these algorithms describe two different approach based on regions constituted by a set of pixels which parameterized objects in shot sequences. For the image segmentation and objects detection, the FCM technique is used, the overlapping between cluster's distribution is minimized by the use of suitable color space (other that the RGB one). The first technique takes into account a priori probabilities governing the computation of various clusters to track objects. A Parzen kernel method is described and allows identifying the players in each frame, we also show the importance of standard deviation value research of the Gaussian probability density function. Region matching is carried out by an algorithm that operates on the Mahalanobis distance between region descriptors in two subsequent frames and uses singular value decomposition to compute a set of correspondences satisfying both the principle of proximity and the principle of exclusion.

Keywords: Image segmentation, objects tracking, Parzen window, singular value decomposition, target recognition.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1938

Authors: Zhichao Jiao, Craig Farnham, Jihui Yuan, Kazuo Emura

Abstract:

It is common to use the outdoor design weather data to select the air-conditioning capacity in the building design stage. The meteorological elements of outdoor design weather data are usually selected based on their excess frequency separately while the dependency between the elements is not well considered. It means that the simultaneous occurrence probability of these elements is smaller than the original excess frequency which may cause an overestimation of selecting air-conditioning capacity. Therefore, the copula approach which can capture the dependency between multivariate data was used to model the joint distributions of the meteorological elements, like air temperature and global solar radiation. We suggest a method based on the specific simultaneous occurrence probability of these two elements of selecting more credible outdoor design conditions. The hourly weather data at 12 noon from 2001 to 2010 in Tokyo, Japan are used to analyze the dependency structure and joint distribution, the Gaussian copula represents the dependence of data best. According to calculating the air temperature and global solar radiation in specific simultaneous occurrence probability and the common exceeding, the results show that both the air temperature and global solar radiation based on simultaneous occurrence probability are lower than these based on the conventional method in the same probability.

Keywords: Copula approach, Design weather database, energy conservation, HVAC.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 291

Authors: Zhengsheng Wang, Jing Qi, Chuntao Liu, Yuanjun Li

Abstract:

The harmonic Arnoldi method can be used to find interior eigenpairs of large matrices. However, it has been shown that this method may converge erratically and even may fail to do so. In this paper, we present a new method for computing interior eigenpairs of large nonsymmetric matrices, which is called weighted harmonic Arnoldi method. The implementation of the method has been tested by numerical examples, the results show that the method converges fast and works with high accuracy.

Keywords: Harmonic Arnoldi method, weighted harmonic Arnoldi method, eigenpair, interior eigenproblem, non symmetric matrix.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1513

Authors: S. Sendilkumar, B. L. Mathur, Joseph Henry

Abstract:

A new approach for protection of power transformer is presented using a time-frequency transform known as Wavelet transform. Different operating conditions such as inrush, Normal, load, External fault and internal fault current are sampled and processed to obtain wavelet coefficients. Different Operating conditions provide variation in wavelet coefficients. Features like energy and Standard deviation are calculated using Parsevals theorem. These features are used as inputs to PNN (Probabilistic neural network) for fault classification. The proposed algorithm provides more accurate results even in the presence of noise inputs and accurately identifies inrush and fault currents. Overall classification accuracy of the proposed method is found to be 96.45%. Simulation of the fault (with and without noise) was done using MATLAB AND SIMULINK software taking 2 cycles of data window (40 m sec) containing 800 samples. The algorithm was evaluated by using 10 % Gaussian white noise.

Keywords: Power Transformer, differential Protection, internalfault, inrush current, Wavelet Energy, Db9.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3093

Authors: Jin Sup Kim, Woo Young Jung, Minho Kwon

Abstract:

In general dynamic analyses, lower mode response is of interest, however the higher modes of spatially discretized equations generally do not represent the real behavior and not affects to global response much. Some implicit algorithms, therefore, are introduced to filter out the high-frequency modes using intended numerical error. The objective of this study is to introduce the P-method and PC α-method to compare that with dissipation method and Newmark method through the stability analysis and numerical example. PC α-method gives more accuracy than other methods because it based on the α-method inherits the superior properties of the implicit α-method. In finite element analysis, the PC α-method is more useful than other methods because it is the explicit scheme and it achieves the second order accuracy and numerical damping simultaneously.

Keywords: Dynamic, α-Method, P-Method, PC α-Method, Newmark method.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3030

Authors: Changqiao Wu, Guoqing Ding, Xin Chen

Abstract:

In this paper, ways of modeling dynamic measurement systems are discussed. Specially, for linear system with single-input single-output, it could be modeled with shallow neural network. Then, gradient based optimization algorithms are used for searching the proper coefficients. Besides, method with normal equation and second order gradient descent are proposed to accelerate the modeling process, and ways of better gradient estimation are discussed. It shows that the mathematical essence of the learning objective is maximum likelihood with noises under Gaussian distribution. For conventional gradient descent, the mini-batch learning and gradient with momentum contribute to faster convergence and enhance model ability. Lastly, experimental results proved the effectiveness of second order gradient descent algorithm, and indicated that optimization with normal equation was the most suitable for linear dynamic models.

Keywords: Dynamic system modeling, neural network, normal equation, second order gradient descent.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 733

Authors: Cheng Zeng, George Michailidis, Hitoshi Iyatomi, Leo L Duan

Abstract:

The conditional density characterizes the distribution of a response variable y given other predictor x, and plays a key role in many statistical tasks, including classification and outlier detection. Although there has been abundant work on the problem of Conditional Density Estimation (CDE) for a low-dimensional response in the presence of a high-dimensional predictor, little work has been done for a high-dimensional response such as images. The promising performance of normalizing flow (NF) neural networks in unconditional density estimation acts a motivating starting point. In this work, we extend NF neural networks when external x is present. Specifically, they use the NF to parameterize a one-to-one transform between a high-dimensional y and a latent z that comprises two components [zP , zN]. The zP component is a low-dimensional subvector obtained from the posterior distribution of an elementary predictive model for x, such as logistic/linear regression. The zN component is a high-dimensional independent Gaussian vector, which explains the variations in y not or less related to x. Unlike existing CDE methods, the proposed approach, coined Augmented Posterior CDE (AP-CDE), only requires a simple modification on the common normalizing flow framework, while significantly improving the interpretation of the latent component, since zP represents a supervised dimension reduction. In image analytics applications, AP-CDE shows good separation of x-related variations due to factors such as lighting condition and subject id, from the other random variations. Further, the experiments show that an unconditional NF neural network, based on an unsupervised model of z, such as Gaussian mixture, fails to generate interpretable results.

Keywords: Conditional density estimation, image generation, normalizing flow, supervised dimension reduction.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 43

Authors: D Sudheer Reddy, N Gopal Reddy, P V Radhadevi, J Saibaba, Geeta Varadan

Abstract:

Smoothing or filtering of data is first preprocessing step for noise suppression in many applications involving data analysis. Moving average is the most popular method of smoothing the data, generalization of this led to the development of Savitzky-Golay filter. Many window smoothing methods were developed by convolving the data with different window functions for different applications; most widely used window functions are Gaussian or Kaiser. Function approximation of the data by polynomial regression or Fourier expansion or wavelet expansion also gives a smoothed data. Wavelets also smooth the data to great extent by thresholding the wavelet coefficients. Almost all smoothing methods destroys the peaks and flatten them when the support of the window is increased. In certain applications it is desirable to retain peaks while smoothing the data as much as possible. In this paper we present a methodology called as peak-wise smoothing that will smooth the data to any desired level without losing the major peak features.

Keywords: smoothing, moving average, peakwise smoothing, spatialdensity models, planar shape models, wavelets.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1718

Authors: B. Bagheri Nakhjavanlo, T. J. Ellis, P. Raoofi, J. Dehmeshki

Abstract:

This paper presents a customized deformable model for the segmentation of abdominal and thoracic aortic aneurysms in CTA datasets. An important challenge in reliably detecting aortic aneurysm is the need to overcome problems associated with intensity inhomogeneities and image noise. Level sets are part of an important class of methods that utilize partial differential equations (PDEs) and have been extensively applied in image segmentation. A Gaussian kernel function in the level set formulation, which extracts the local intensity information, aids the suppression of noise in the extracted regions of interest and then guides the motion of the evolving contour for the detection of weak boundaries. The speed of curve evolution has been significantly improved with a resulting decrease in segmentation time compared with previous implementations of level sets. The results indicate the method is more effective than other approaches in coping with intensity inhomogeneities.

Keywords: Abdominal and thoracic aortic aneurysms, intensityinhomogeneity, level sets, local fitting binary.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1775

Authors: C. G. Jinimole, A. Harsha

Abstract:

One of the key problems facing in the analysis of Computed Tomography (CT) images is the poor contrast of the images. Image enhancement can be used to improve the visual clarity and quality of the images or to provide a better transformation representation for further processing. Contrast enhancement of images is one of the acceptable methods used for image enhancement in various applications in the medical field. This will be helpful to visualize and extract details of brain infarctions, tumors, and cancers from the CT image. This paper presents a comparison study of five contrast enhancement techniques suitable for the contrast enhancement of CT images. The types of techniques include Power Law Transformation, Logarithmic Transformation, Histogram Equalization, Contrast Stretching, and Laplacian Transformation. All these techniques are compared with each other to find out which enhancement provides better contrast of CT image. For the comparison of the techniques, the parameters Peak Signal to Noise Ratio (PSNR) and Mean Square Error (MSE) are used. Logarithmic Transformation provided the clearer and best quality image compared to all other techniques studied and has got the highest value of PSNR. Comparison concludes with better approach for its future research especially for mapping abnormalities from CT images resulting from Brain Injuries.

Keywords: Computed tomography, enhancement techniques, increasing contrast, PSNR and MSE.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1342

Authors: Othman Belgassim, Krzysztof Jemielniak

Abstract:

The analysis of Acoustic Emission (AE) signal generated from metal cutting processes has often approached statistically. This is due to the stochastic nature of the emission signal as a result of factors effecting the signal from its generation through transmission and sensing. Different techniques are applied in this manner, each of which is suitable for certain processes. In metal cutting where the emission generated by the deformation process is rather continuous, an appropriate method for analysing the AE signal based on the root mean square (RMS) of the signal is often used and is suitable for use with the conventional signal processing systems. The aim of this paper is to set a strategy in tool failure detection in turning processes via the statistic analysis of the AE generated from the cutting zone. The strategy is based on the investigation of the distribution moments of the AE signal at predetermined sampling. The skews and kurtosis of these distributions are the key elements in the detection. A normal (Gaussian) distribution has first been suggested then this was eliminated due to insufficiency. The so called Beta distribution was then considered, this has been used with an assumed β density function and has given promising results with regard to chipping and tool breakage detection.

Keywords: AE signal, skew, kurtosis, tool failure

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1807

Authors: R. Senthilkumar

Abstract:

Most simple nonlinear thresholding rules for wavelet- based denoising assume that the wavelet coefficients are independent. However, wavelet coefficients of natural images have significant dependencies. This paper attempts to give a recipe for selecting one of the popular image-denoising algorithms based on VisuShrink, SureShrink, OracleShrink, BayesShrink and BiShrink and also this paper compares different Bivariate models used for image denoising applications. The first part of the paper compares different Shrinkage functions used for image-denoising. The second part of the paper compares different bivariate models and the third part of this paper uses the Bivariate model with modified marginal variance which is based on Laplacian assumption. This paper gives an experimental comparison on six 512x512 commonly used images, Lenna, Barbara, Goldhill, Clown, Boat and Stonehenge. The following noise powers 25dB,26dB, 27dB, 28dB and 29dB are added to the six standard images and the corresponding Peak Signal to Noise Ratio (PSNR) values are calculated for each noise level.

Keywords: BiShrink, Image-Denoising, PSNR, Shrinkage function

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1304