No-reference image quality assessment using modified extreme learning machine classifier

In this paper, we present a machine learning approach to measure the visual quality of JPEG-coded images. The features for predicting the perceived image quality are extracted by considering key human visual sensitivity (HVS) factors such as edge amplitude, edge length, background activity and background luminance. Image quality assessment involves estimating the functional relationship between HVS features and subjective test scores. The quality of the compressed images are obtained without referring to their original images ('No Reference' metric). Here, the problem of quality estimation is transformed to a classification problem and solved using extreme learning machine (ELM) algorithm. In ELM, the input weights and the bias values are randomly chosen and the output weights are analytically calculated. The generalization performance of the ELM algorithm for classification problems with imbalance in the number of samples per quality class depends critically on the input weights and the bias values. Hence, we propose two schemes, namely the k-fold selection scheme (KS-ELM) and the real-coded genetic algorithm (RCGA-ELM) to select the input weights and the bias values such that the generalization performance of the classifier is a maximum. Results indicate that the proposed schemes significantly improve the performance of ELM classifier under imbalance condition for image quality assessment. The experimental results prove that the estimated visual quality of the proposed RCGA-ELM emulates the mean opinion score very well. The experimental results are compared with the existing JPEG no-reference image quality metric and full-reference structural similarity image quality metric.

Speeding up AdaBoost Classifier with Random Projection

The development of techniques for scaling up classifiers so that they can be applied to problems with large datasets of training examples is one of the objectives of data mining. Recently, AdaBoost has become popular among machine learning community thanks to its promising results across a variety of applications. However, training AdaBoost on large datasets is a major problem, especially when the dimensionality of the data is very high. This paper discusses the effect of high dimensionality on the training process of AdaBoost. Two preprocessing options to reduce dimensionality, namely the principal component analysis and random projection are briefly examined. Random projection subject to a probabilistic length preserving transformation is explored further as a computationally light preprocessing step. The experimental results obtained demonstrate the effectiveness of the proposed training process for handling high dimensional large datasets.

Use of S1 nuclease and formamide in combination for the reassociation studies on GC-rich DNA

The optimal parameters in the use of nuclease S1 in DNA reassociation kinetics in the presence of formamide have been determined. The conditions are especially suitable for the study of DNA rich in mole percent GC. A 10-fold dilution of the reassociation samples leading to a decrease in both NaCl and formamide concentrations, consequently resulting in a lowering of Tm by only 1.5°C, and the S1 digestion at temperatures identical to the reassociation assay in order to retain the stability of the duplex, are two important aspects of this system. Under these conditions, the kinetics of reassociation followed the theoretically predicted pattern, while the earlier reported methods have shown lower values.

A Combination of Hexagonal and 12-Sided Polygonal Voltage Space Vector PWM Control for IM Drives Using Cascaded Two-Level Inverters

This paper proposes a multilevel inverter configuration which produces a hexagonal voltage space vector structure in the lower modulation region and a 12-sided polygonal space vector structure in the overmodulation region. A conventional multilevel inverter produces 6n plusmn 1 (n = odd) harmonics in the phase voltage during overmodulation and in the extreme square-wave mode of operation. However, this inverter produces a 12-sided polygonal space vector location, leading to the elimination of 6n plusmn 1 (n = odd) harmonics in the overmodulation region extending to a final 12-step mode of operation with a smooth transition. The benefits of this arrangement are lower losses and reduced torque pulsation in an induction motor drive fed from this converter at higher modulation indexes. The inverter is fabricated by using three conventional cascaded two-level inverters with asymmetric dc-bus voltages. A comparative simulation study of the harmonic distortion in the phase voltage and associated losses in conventional multilevel inverters and that of the proposed inverter is presented in this paper. Experimental validation on a prototype shows that the proposed converter is suitable for high-power applications because of low harmonic distortion and low losses.

Validation-Based Sparse Gaussian Process Classifier Design

Gaussian processes (GPs) are promising Bayesian methods for classification and regression problems. Design of a GP classifier and making predictions using it is, however, computationally demanding, especially when the training set size is large. Sparse GP classifiers are known to overcome this limitation. In this letter, we propose and study a validation-based method for sparse GP classifier design. The proposed method uses a negative log predictive (NLP) loss measure, which is easy to compute for GP models. We use this measure for both basis vector selection and hyperparameter adaptation. The experimental results on several real-world benchmark data sets show better orcomparable generalization performance over existing methods.

CSSI-PRO: a method for secondary structure type editing, assignment and estimation in proteins using linear combination of backbone chemical shifts

Estimation of secondary structure in polypeptides is important for studying their structure, folding and dynamics. In NMR spectroscopy, such information is generally obtained after sequence specific resonance assignments are completed. We present here a new methodology for assignment of secondary structure type to spin systems in proteins directly from NMR spectra, without prior knowledge of resonance assignments. The methodology, named Combination of Shifts for Secondary Structure Identification in Proteins (CSSI-PRO), involves detection of specific linear combination of backbone H-1(alpha) and C-13' chemical shifts in a two-dimensional (2D) NMR experiment based on G-matrix Fourier transform (GFT) NMR spectroscopy. Such linear combinations of shifts facilitate editing of residues belonging to alpha-helical/beta-strand regions into distinct spectral regions nearly independent of the amino acid type, thereby allowing the estimation of overall secondary structure content of the protein. Comparison of the predicted secondary structure content with those estimated based on their respective 3D structures and/or the method of Chemical Shift Index for 237 proteins gives a correlation of more than 90% and an overall rmsd of 7.0%, which is comparable to other biophysical techniques used for structural characterization of proteins. Taken together, this methodology has a wide range of applications in NMR spectroscopy such as rapid protein structure determination, monitoring conformational changes in protein-folding/ligand-binding studies and automated resonance assignment.

Experimental investigation of interaction of strong shock heated oxygen gas on the surface of ZrO2 - a novel method to understand re-combination heating

Interaction of shock heated test gas in the free piston driven shock tube with bulk and thin film of cubic zirconium dioxide (ZrO2) prepared by combustion method is investigated. The test samples before and after exposure to the shock wave are analyzed by X-ray diffraction (XRD), X-Ray Photoelectron Spectroscopy (XPS) and Scanning Electron Microscope (SEM). The study shows transformation of metastable cubic ZrO2 to stable monoclinic ZrO2 phase after interacting with shock heated oxygen gas due to the heterogeneous catalytic recombination surface reaction.

Diffusion of Pure CH4 and Its Binary Mixture with CO2 in Faujasite NaY: A Combination of Neutron Scattering Experiments and Molecular Dynamics Simulations

The self-diffusion properties of pure CH4 and its binary mixture with CO2 within MY zeolite have been investigated by combining an experimental quasi-elastic neutron scattering (QENS) technique and classical Molecular dynamics simulations. The QENS measurements carried out at 200 K led to an unexpected self-diffusivity profile for Pure CH4 with the presence of a maximum for a loading of 32 CH4/unit cell, which was never observed before for the diffusion of apolar species in azeolite system With large windows. Molecular dynamics simulations were performed using two distinct microscopic models for representing the CH4/NaY interactions. Depending on the model, we are able to fairly reproduce either the magnitude or the profile of the self-diffusivity.Further analysis allowed LIS to provide some molecular insight into the diffusion mechanism in play. The QENS measurements report only a slight decrease of the self-diffusivity of CH4 in the presence of CO2 when the CO2 loading increases. Molecular dynamics simulations successfully capture this experimental trend and suggest a plausible microscopic diffusion mechanism in the case of this binary mixture.

Automatic text block separation in document images

Separation of printed text blocks from the non-text areas, containing signatures, handwritten text, logos and other such symbols, is a necessary first step for an OCR involving printed text recognition. In the present work, we compare the efficacy of some feature-classifier combinations to carry out this separation task. We have selected length-nomalized horizontal projection profile (HPP) as the starting point of such a separation task. This is with the assumption that the printed text blocks contain lines of text which generate HPP's with some regularity. Such an assumption is demonstrated to be valid. Our features are the HPP and its two transformed versions, namely, eigen and Fisher profiles. Four well known classifiers, namely, Nearest neighbor, Linear discriminant function, SVM's and artificial neural networks have been considered and efficiency of the combination of these classifiers with the above features is compared. A sequential floating feature selection technique has been adopted to enhance the efficiency of this separation task. The results give an average accuracy of about 96.

On the momentum balance in linear-combination models for the transition zone

The momentum balance of the linear-combination integral model for the transition zone is investigated for constant pressure flows. The imbalance is found to be small enough to be negligible for all practical purposes. [S0889-504X(00)00703-0].

Two-dimensional NMR spectroscopy of connected transitions by linear combination of flip angle dependent cosy: An alternative scheme for E.COSY

An alternative pulse scheme which simplifies and improves the recently proposed P.E.COSY experiment is suggested for the retention of connected or unconnected transitions in a coupled spin system. An important feature of the proposed pulse scheme is the improved phase characteristics of the diagonal peaks. A comparison of various experiments designed for this purpose, namely COSY-45, E.COSY, P.E.COSY and the present scheme (A.E.COSY), is also presented. The suppression of unconnected transitions and the measurement of scalar coupling constants and their relative signs are illustrated from A.E.COSY spectra of 2,3-dibromopropionic acid and 2-(2-thienyl)pyridine.

Design of speaker identification schemes for large number of speakers (A)

Design of speaker identification schemes for a small number of speakers (around 10) with a high degree of accuracy in controlled environment is a practical proposition today. When the number of speakers is large (say 50–100), many of these schemes cannot be directly extended, as both recognition error and computation time increase monotonically with population size. The feature selection problem is also complex for such schemes. Though there were earlier attempts to rank order features based on statistical distance measures, it has been observed only recently that the best two independent measurements are not the same as the combination in two's for pattern classification. We propose here a systematic approach to the problem using the decision tree or hierarchical classifier with the following objectives: (1) Design of optimal policy at each node of the tree given the tree structure i.e., the tree skeleton and the features to be used at each node. (2) Determination of the optimal feature measurement and decision policy given only the tree skeleton. Applicability of optimization procedures such as dynamic programming in the design of such trees is studied. The experimental results deal with the design of a 50 speaker identification scheme based on this approach.

On uniqueness, transfer and combination of acoustic sources in one-dimensional systems

The use of electroacoustic analogies suggests that a source of acoustical energy (such as an engine, compressor, blower, turbine, loudspeaker, etc.) can be characterized by an acoustic source pressure ps and internal source impedance Zs, analogous to the open-circuit voltage and internal impedance of an electrical source. The present paper shows analytically that the source characteristics evaluated by means of the indirect methods are independent of the loads selected; that is, the evaluated values of ps and Zs are unique, and that the results of the different methods (including the direct method) are identical. In addition, general relations have been derived here for the transfer of source characteristics from one station to another station across one or more acoustical elements, and also for combining several sources into a single equivalent source. Finally, all the conclusions are extended to the case of a uniformly moving medium, incorporating the convective as well as dissipative effects of the mean flow.

Ribavirin-Induced Anemia in Hepatitis C Virus Patients Undergoing Combination Therapy

The current standard of care for hepatitis C virus (HCV) infection - combination therapy with pegylated interferon and ribavirin - elicits sustained responses in only similar to 50% of the patients treated. No alternatives exist for patients who do not respond to combination therapy. Addition of ribavirin substantially improves response rates to interferon and lowers relapse rates following the cessation of therapy, suggesting that increasing ribavirin exposure may further improve treatment response. A key limitation, however, is the toxic side-effect of ribavirin, hemolytic anemia, which often necessitates a reduction of ribavirin dosage and compromises treatment response. Maximizing treatment response thus requires striking a balance between the antiviral and hemolytic activities of ribavirin. Current models of viral kinetics describe the enhancement of treatment response due to ribavirin. Ribavirin-induced anemia, however, remains poorly understood and precludes rational optimization of combination therapy. Here, we develop a new mathematical model of the population dynamics of erythrocytes that quantitatively describes ribavirin-induced anemia in HCV patients. Based on the assumption that ribavirin accumulation decreases erythrocyte lifespan in a dose-dependent manner, model predictions capture several independent experimental observations of the accumulation of ribavirin in erythrocytes and the resulting decline of hemoglobin in HCV patients undergoing combination therapy, estimate the reduced erythrocyte lifespan during therapy, and describe inter-patient variations in the severity of ribavirin-induced anemia. Further, model predictions estimate the threshold ribavirin exposure beyond which anemia becomes intolerable and suggest guidelines for the usage of growth hormones, such as erythropoietin, that stimulate erythrocyte production and avert the reduction of ribavirin dosage, thereby improving treatment response. Our model thus facilitates, in conjunction with models of viral kinetics, the rational identification of treatment protocols that maximize treatment response while curtailing side effects.