Hossein Ahmadi Noubari

Gauss–Laguerre wavelet textural feature fusion with geometrical information for facial expression identification

Facial expressions are a valuable source of information that accompanies facial biometrics. Early detection of physiological and psycho-emotional data from facial expressions is linked to the situational awareness module of any advanced biometric system for personal state re/identification. In this article, a new method that utilizes both texture and geometric information of facial fiducial points is presented. We investigate Gauss–Laguerre wavelets, which have rich frequency extraction capabilities, to extract texture information of various facial expressions. Rotation invariance and the multiscale approach of these wavelets make the feature extraction robust. Moreover, geometric positions of fiducial points provide valuable information for upper/lower face action units. The combination of these two types of features is used for facial expression classification. The performance of this system has been validated on three public databases: the JAFFE, the Cohn-Kanade, and the MMI image.

A novel approach proposed for fractured zone detection using petrophysical logs

Fracture detection is a key step in wellbore stability and fractured reservoir fluid flow simulation. While different methods have been proposed for fractured zones detection, each of them is associated with certain shortcomings that prevent their full use in different related engineering applications. In this paper, a novel combined method is proposed for fractured zone detection, using processing of petrophysical logs with wavelet, classification and data fusion techniques. Image and petrophysical logs from Asmari reservoir in eight wells of an oilfield in southwestern Iran were used to investigate the accuracy and applicability of the proposed method. Initially, an energy matching strategy was utilized to select the optimum mother wavelets for de-noising and decomposition of petrophysical logs. Parzen and Bayesian classifiers were applied to raw, de-noised and various frequency bands of logs after decomposition in order to detect fractured zones. Results show that the low-frequency bands (approximation 2, a2) of de-noised logs are the best data for fractured zones detection. These classifiers considered one well as test well and the other seven wells as train wells. Majority voting, optimistic OWA (ordered weighted averaging) and pessimistic OWA methods were used to fuse the results obtained from seven train wells. Results confirmed that Parzen and optimistic OWA are the best combined methods to detect fractured zones. The generalization of method is confirmed with an average accuracy of about 72%.

Manifold Learning Applied on EEG Signal of the Epileptic Patients for Detection of Normal and Pre-Seizure States

In this paper, several manifold learning (ML) techniques for dimension reduction of EEG feature vectors are introduced and applied on set of epileptic EEG signals. These include principal component analysis (PCA), multidimensional scaling (MDS), isometric mapping (ISOMAP) and locally linear embedding (LLE). While EEG signals of epileptic patients contain necessary information with regards to the various brain states of epileptic patients, for extraction of useful information in the EEG signals and for detection, often construction of high-dimensional feature vectors is utilized. Analysis of such high-dimensional feature vectors are complex and time consuming. This paper deals with dimension reduction of the extracted feature vectors and comparative analysis of the performance of several manifold learning techniques as applied on EEG signals of epileptic patients.

Evoked hemodynamic response estimation using ensemble empirical mode decomposition based adaptive algorithm applied to dual channel functional near infrared spectroscopy (fNIRS).

BACKGROUND The quality of the functional near infrared spectroscopy (fNIRS) recordings is highly degraded by the presence of physiological interferences. It is crucial to efficiently separate the evoked hemodynamic responses (EHRs) from other background hemodynamic activities prior to any further processing. NEW METHOD This paper presents a novel algorithm for physiological interferences reduction from the dual channel fNIRS measurements using ensemble empirical mode decomposition (EEMD) technique. The proposed algorithm is comprised of two main steps: (1) decomposing reference signal into its constituents called intrinsic mode functions (IMFs) and (2) adaptively defining appropriate weights of the corresponding IMFs to estimate the proportion of physiological interference in standard channel measurement. RESULTS Performance of the proposed algorithm was evaluated using both synthetic and semi-real brain hemodynamic data based on four parameters of relative mean squared error (rMSE), Pearsons correlation coefficient (R(2)), percentage estimation error of peak amplitude (EPA) and peak latency (EL). COMPARISON WITH EXISTING METHODS Results obtained from synthetic data revealed that both the EEMD based normalized least mean squares (EEMD-NLMS) and EEMD based recursive least squares (EEMD-RLS) methods could reduce the average rMSE by at least 34% and 49%, respectively, when compared with widely used methods: block averaging, band-pass filtering and principal and/or independent component analysis. Furthermore, the two proposed methods outperform the regression method in reducing rMSE by at least 21% and 35% respectively when applied to semi-real data. CONCLUSIONS An effective algorithm for estimating the EHRs from raw fNIRS data was proposed in which no assumption about the amplitude, shape and duration of the responses is considered.

Detection of cancerous zones in mammograms using fractal modeling and classification by probabilistic neural network

Recent studies on the geometry of fractals indicate that tumors with irregular shapes can be utilized for the study of the morphology and diagnosis of cancerous cases. In this paper, we deal with the fractal modeling of the mammographic images and their background morphology. It is shown that the use of fractal modeling as applied to a given image can clearly discern cancerous zones from noncancerous areas. Our results show that fractal modeling of images can be used as an effective tool for identification of cancerous cells. For fractal modeling, the original image is first segmented into appropriate fractal boxes followed by identifying the fractal dimension of each windowed section. We have used two dimensional box counting algorithm after which based on the order of the computations, they are placed in an appropriate matrix to facilitate the required computations. Finally using eight features identified as characteristic features of tumors extracted from mammogram images, the results obtained from the preliminary analysis stages, were utilized in a neural network for classification of cells into malignant and benign with the accuracy of 89.21 % classification results.

Hardware implementation of lifting based wavelet transform

In this paper, a VLSI implementation of the lifting-based Discrete Wavelet Transform (DWT) is presented. The behavioral description of integer-to-integer CDF (2,2) lifting wavelet, which is used in image compression has been coded in Verilog Hardware Description Language (HDL). The code has been synthesized and then implemented using both Field Programmable Gate Array (FPGA) and Application Specific Integrated Circuit (ASIC) design approaches. Post-synthesis and post-layout simulations verify the appropriate operation of he architecture. The resulting hardware can be used in image compression applications such as JPEG2000.

A novel model predictive controller for uncertain constrained systems

Abstract Model Predictive Control (MPC) is a method of choice for control of many industrial applications. MPC uses optimization based prediction of a model in a predetermined horizon to determine control action. While uncertainty is an inherent part of most dynamical systems, model mismatches in the case of uncertain dynamical system lead to undesirable closed loop response from steady state errors to instability. There are several research studies for robust MPC. However, most of the results are very conservative, where optimality is sacrificed to obtain worst case constraint satisfaction. This paper presents a novel MPC for control of non-linear uncertain dynamical systems with reliability type constraints, where the probability of violation of constraints is limited. We utilize polynomial chaos expansion (PCE) in order to propagate uncertainties. We show that it is possible to transform probabilistic constraints to a deterministic constraint on PCE coefficients. This approach prevents sampling of the uncertainty, thus it is efficient numerically and computationally.

Autonomic-cardiorespiratory regulation: A physiology-based mathematical model

This paper presents a novel physiology-based mathematical model of autonomic-cardiorespiratory regulation described by a set of three nonlinear, coupled differential equations. We improved our previously proposed autonomic-cardiac regulation model by considering neuromechanical and mechanical coupling of cardiovascular and respiration systems including lung stretch-receptor reflex and venous return variation. We also introduced a differential equation describing respiration rate regulation which mainly originates in the medullary respiratory center. The results of simulation experiments suggest that the venous return variation generates a higher perturbation on heart rate and blood pressure than lung stretch-receptor reflex. The proposed model is also powerful in determining and removing direct respiratory impacts on parasympathetic activation tone to accurately extract parasympathetic activity caused by emotional states and environmental conditions.

Application of wavelets and fractal-based methods for detection of microcalcification in mammograms: a comparative analysis using neural network

Recent studies on the wavelet transform and geometry of fractals indicate that microcalcification can be utilized for the study of the morphology and diagnosis of cancerous cases. In this paper we deal with the fractal modeling of the mammographic images and their background morphology. It is shown that the use of fractal modeling as applied to a given image can clearly discern cancerous zones from noncancerous areas. Our results show that fractal modeling of images can be used as an effective tool for identification of cancerous cells. For fractal modeling, the original image is first segmented into appropriate fractal boxes followed by identifying the fractal dimension of each windowed section. We have used two dimensional box counting algorithm after which based on the order of the computations, they are placed in an appropriate matrix to facilitate the required computations.For wavelet transform,the original image is first analysed by db2 to 3 different resolution levels and for detection of microcalcification we just need to nullify wavelet coefficients of the image at first scale and low frequency at the third scale subimages and take reverse wavelet transform of the remaining coefficients to reconstruct mammogram.Finally using eight features identified as characteristic features of microcalcification extracted from mammograms, the results obtained from the preliminary analysis stages, were utilized in a neural network for classification of cells into malignant and benign with the accuracy of 89.21 % classification results in fractal method and accuracy of 88.23 % classification results in wavelet method.

Design of Optimum Wavelet for Noise Suppression and its Application to Ultrasonic Echo Delay Estimation

In this paper an optimum wavelet for denoising applications based on signal to noise ratio (SNR) improvement is designed. The approach is based on searching for a wavelet basis that provides the best non linear approximation of a given signal. It is shown that such a basis will have the best wavelet denoising performance in the sense of signal estimation error. The performance of this method is studied on a few standard test signals and as a practical application, the proposed method is used for denoising and estimating the time difference of arrival (TDOA) of ultrasonic echoes which are widely used in non destructive testing. Simulation results show an improvement of up to 1.2 dB in SNR enhancement for the given test signals comparing with standard wavelets.