Classifier combination for in vivo magnetic resonance spectra of brain tumours

In this paper we present a multi-stage classifier for magnetic resonance spectra of human brain tumours which is being developed as part of a decision support system for radiologists. The basic idea is to decompose a complex classification scheme into a sequence of classifiers, each specialising in different classes of tumours and trying to reproducepart of the WHO classification hierarchy. Each stage uses a particular set of classification features, which are selected using a combination of classical statistical analysis, splitting performance and previous knowledge.Classifiers with different behaviour are combined using a simple voting scheme in order to extract different error patterns: LDA, decision trees and the k-NN classifier. A special label named "unknown¿ is used when the outcomes of the different classifiers disagree. Cascading is alsoused to incorporate class distances computed using LDA into decision trees. Both cascading and voting are effective tools to improve classification accuracy. Experiments also show that it is possible to extract useful information from the classification process itself in order to helpusers (clinicians and radiologists) to make more accurate predictions and reduce the number of possible classification mistakes.

Nuclear magnetic resonance investigation of structure and dynamics of some electrolyte solutions

Electrolyte solutions are of importance in a wide range of scientific contexts and as such have attracted considerable theoretical and experimental effort over many years. Nuclear Magnetic resonance provides a precise and versatile tool for investigation of electrolyte solutions, both in water and in organic solvents. Many structural and dynamic properties can be obtained through NMR experiments. The solution of aluminum chloride in water was studied. Different concentrations were taken for investigation. Independence of maximum line shift from concentration and acidity was shown. Six-coordinated structure of solvation shell was confirmed by experiments on 'H and 27A1 nuclei. Diffusion coefficients were studied. The solution of nickel chloride in methanol was studied. Lines, corresponding to coordinated and bulk methanol were found. Four-, five- and six-coordinated structures were found in different temperatures. The line for coordinated -OD group of deuterated methanol was observed on 2H spectrum for the first time. Partial deuteration of CH3 group was detected. Inability to observe coordinated -OH group was explained.

Developing of AC magnetometer and investigation of magnetic properties of LSMO

In this work AC magnetometer was developed and primary test measurements were performed for temperature range from 77 K up to 350 K in frequency range from 1 kHz up to 20 kHz. In the course of the present work dependencies of magnetization on temperature for Lao7Sr03Mni _yFeyO3 with y = 0.15, 0.20, 0.25 were obtained in DC magnetic field using SQUID magnetometer and in AC magnetic field using the developed AC magnetometer. Lai.XSrXMnO3 (LSMO) compounds belong to the class of Mn perovskites, which demonstrate very high degree of spin polarization. These materials are of great importance for nowadays applications in spintronics, where spin polarized electron transport is used. Spin glass like behavior was found as a characteristic feature of these solid solutions with the freezing temperature in the range 65 — 210 K.

Origin of magnetic and magnetoelastic tweedlike precursor modulations in ferroic materials

Based on experimental observations of modulated magnetic patterns in a Co0.5Ni0.205Ga0.295 alloy, we propose a model to describe a (purely) magnetic tweed and a magnetoelastic tweed. The former arises above the Curie (or Nel) temperature due to magnetic disorder. The latter results from compositional fluctuations coupling to strain and then to magnetism through the magnetoelastic interaction above the structural transition temperature. We discuss the origin of purely magnetic and magnetoelastic precursor modulations and their experimental thermodynamic signatures.

Influence of elastic anisotropy on structural nanoscale textures

We study the influence of elastic anisotropy on nanoscale precursor textures that exist in some shape-memory alloys and show that tweed occurs in the limit of high elastic anisotropy while a nanocluster phase-separated state occurs for values of anisotropy inhibiting the formation of martensite. These results are consistent with specific heat data, elastic constant measurements, and zero-field cooling or field cooling experiments in nonstoichiometric NiTi alloys.

CFD simulation of complex phenomena containing suspensions and flow throughporous media

There is an increasing reliance on computers to solve complex engineering problems. This is because computers, in addition to supporting the development and implementation of adequate and clear models, can especially minimize the financial support required. The ability of computers to perform complex calculations at high speed has enabled the creation of highly complex systems to model real-world phenomena. The complexity of the fluid dynamics problem makes it difficult or impossible to solve equations of an object in a flow exactly. Approximate solutions can be obtained by construction and measurement of prototypes placed in a flow, or by use of a numerical simulation. Since usage of prototypes can be prohibitively time-consuming and expensive, many have turned to simulations to provide insight during the engineering process. In this case the simulation setup and parameters can be altered much more easily than one could with a real-world experiment. The objective of this research work is to develop numerical models for different suspensions (fiber suspensions, blood flow through microvessels and branching geometries, and magnetic fluids), and also fluid flow through porous media. The models will have merit as a scientific tool and will also have practical application in industries. Most of the numerical simulations were done by the commercial software, Fluent, and user defined functions were added to apply a multiscale method and magnetic field. The results from simulation of fiber suspension can elucidate the physics behind the break up of a fiber floc, opening the possibility for developing a meaningful numerical model of the fiber flow. The simulation of blood movement from an arteriole through a venule via a capillary showed that the model based on VOF can successfully predict the deformation and flow of RBCs in an arteriole. Furthermore, the result corresponds to the experimental observation illustrates that the RBC is deformed during the movement. The concluding remarks presented, provide a correct methodology and a mathematical and numerical framework for the simulation of blood flows in branching. Analysis of ferrofluids simulations indicate that the magnetic Soret effect can be even higher than the conventional one and its strength depends on the strength of magnetic field, confirmed experimentally by Völker and Odenbach. It was also shown that when a magnetic field is perpendicular to the temperature gradient, there will be additional increase in the heat transfer compared to the cases where the magnetic field is parallel to the temperature gradient. In addition, the statistical evaluation (Taguchi technique) on magnetic fluids showed that the temperature and initial concentration of the magnetic phase exert the maximum and minimum contribution to the thermodiffusion, respectively. In the simulation of flow through porous media, dimensionless pressure drop was studied at different Reynolds numbers, based on pore permeability and interstitial fluid velocity. The obtained results agreed well with the correlation of Macdonald et al. (1979) for the range of actual flow Reynolds studied. Furthermore, calculated results for the dispersion coefficients in the cylinder geometry were found to be in agreement with those of Seymour and Callaghan.