Tanuja Srivastava

Fe(III)/Fe(II) — ligand systems for use as negative half-cells in redox-flow cells

Abstract Cyclic voltammetric and controlled-potential electrolysis studies of Fe(III)/Fe(II) complexes with diethylenetriaminepentaacetic, nitrilotriacetic and ethylenediaminetetraacetic acids show that the formal potential of the iron redox couple shifts markedly to negative values. The stability and heterogeneous rate constants appear to favour the application of such systems as the negative half-cell on an all-iron flow cell.

Application of fractal parameters for unsupervised classification of SAR images: a simulation based study

Classification of any satellite image with unsupervised or supervised technique is still very challenging task. A lot of researchers are working for this but still uncertainties exist to label the different classes. Land cover classification with satellite images is very much dependent on the type of satellite images. Nowadays, Synthetic Aperture Radar (SAR) images are giving very promising results in comparison to optical images. Therefore, in this paper a contextual classification has been performed in an unsupervised way for SAR image. For this purpose, fractal parameters viz. fractal dimension and lacunarity is used. In order to apply the methodology, first of all a set of simulated SAR images has been generated and tested for classification and then the proposed methodology is applied on satellite SAR images, i.e., ERS-2 SAR. The classification accuracy for simulated images comes up to 85% whereas for satellite SAR it reaches to 76%.

The potential application of fractal approach for surface roughness retrieval: A study for simulated surfaces

Natural surfaces can be modelled with fractals because fractals properly account for scale invariance and self similarity of these surfaces. The well-known measure, i.e. fractal dimension (D), is the property used to describe the roughness of fractal surfaces. Retrieving the Earths surface roughness with satellite images, particularly synthetic aperture radar (SAR) images, is an interesting and challenging task. Consequently, many researchers are using electromagnetic models, various inversion techniques, semi-empirical models to retrieve the roughness parameters, i.e. RMS surface height (s) and autocorrelation length (l). Most of the models require some a priori information or some given values to solve the equations and retrieve l and s. Uncertainty still exists to retrieve these parameters with minimum or no a priori information. Therefore, in this paper, the fractal dimension approach has been applied to correlate l and s with fractal properties for development of a surface parameter retrieval algorithm. For this purpose, 1500 synthetic surfaces for known l and s have been generated, and their fractal dimension has been computed. D has also been computed after introducing Gaussian and speckle noise to the generated surfaces. The analysis among D, l and s shows the potentiality of relationship among these parameters and is helpful in developing a relationship among them by which l and s can be retrieved. The values of l and s are retrieved with the help of a look-up table for the synthetic surfaces which can be extended for retrieval of roughness parameters from the SAR images.

Multifractal analysis of SAR images for unsupervised classification

In present paper an attempt has been made for unsupervised classification of SAR images based on the surface roughness using multifractal technique. Surface roughness is measured with the help of fractal dimension (D), which lies in the range 2.0 and 3.0. Based on roughness values, i.e., D, various land classes are grouped in different classes. The D values are estimated for a number of local window sizes and thus the window size is very important for classification. The window size is optimized for best classification and in present case it is 9times9. The K-means classifier has been used for this procedure which clusters various land classes according to D values. Although fractal dimension is able to provide the roughness values for various land classes, it can not uniquely identify all classes. In order to remove this discrepancy, the multifractal analysis has been performed. The multifractal dimension has been estimated as 5 generalized dimensions providing 5 multifractal images and then these images are classified. The overall classification accuracy using fractal dimension alone comes to be nearly 60% while it increases to 67% with multifractal images.

Design of Optimal Statistical Filter for the Discrete Convolution Backprojection Method

SYNOPTIC ABSTRACTA statistically optimal convolving function (filter) is derived for discrete implementation of the convolution backprojection method. The method of derivation simultaneously takes into account (i) “the process” generating the object images, (ii) the data collection geometry, (iii) the discretization scheme used, (iv) various interpolations used, and (v) the data noise. The proposed filter minimizes the expected weighted sum of squared errors in pixel reconstructions. The performance of this optimal filter, in reconstructing discretized images by convolution backprojection method, is compared with that of some of the commonly used filters.

Reconstructing h-convex binary images from its horizontal and vertical projections by simulated annealing

The field of Discrete Tomography (DT) deals with the reconstruction of 2D discrete images from a few number of their projections. The ideal problem of DT is to reconstruct a binary image from its horizontal and vertical projections. It turns out that this problem of DT is highly underdetermined and therefore it is inevitable to impose additional constraints to this problem. This paper uses the convexity property of binary images and the problem of reconstruction of h-convex binary images from its horizontal and vertical projections is considered here. This problem is transformed into two different optimization problems by defining two appropriate objective functions. Then two simulated annealing (SA) algorithms to solve the two optimization problems are developed. The SA algorithms are tested on various randomly generated test images. The algorithms are also tested on noisy images. Finally numerical results have been reported showing good reconstruction fidelity.

Image Reconstruction from Projection under Periodicity Constraints Using Genetic Algorithm

In this paper we study the problem of image reconstruction from a small number of projections. This type of problem arises in material science during developing the program for the reconstruction of crystalline structure from their projection image, obtained by high-resolution transmission electron microscopy. The problem has large number of solutions due to few projections. To reduce the number of solutions we can use some priori information about the object. This priori information is called constraints. One of these constraints is periodicity constraint. We use genetic algorithm to optimize the solution, which is an evolutionary technique to solve the problem.

Pitman Estimator to Pareto Scale Parameters

ABSTRACT The Pitman estimators for ordered scale parameters of two Pareto distributions are obtained here. In case of known ordering that have been modified using more informative priors, and found to be better in terms of risks. Some new estimators are also proposed and show that they improve upon Pitman estimators. The results were also verified by simulation study.

Statistical Method of Estimation of ESE in CBP

SYNOPTIC ABSTRACT A linear relationship between the expected squared error (ESE) in reconstruction of images by CBP method, and the covariance kernel of the spatial stochastic process of object function (images) has been established, where the covariance kernel is estimated by sample observations. And a method to estimate the error in reconstruction in practice using the linear relationship is given. When the object covariance kernel is not known an estimate of the covariance kernel is obtained from reconstructed images, and it is used to predict the error. The test cases have been simulated as random images from certain classes of spatial stochastic processes. It is empirically observed that the predicted error in both the cases of known and unknown (estimated) covariance kernel is within reasonable limits of the actual error.

Convolution backprojection algorithm for tomographic image reconstruction with contourlet transform

In the present script, the problem of recovering tomographic images from projection data using the multiresolution approximation of contourlet transform, is considered. This algorithm has similarity with main convolution backprojection but it defines local information in different directions. In this algorithm, error arises due to the approximation of the density function has also been estimated.