Tematem pracy jest wizualizacja danych wielowymiarowych reprezentowanych przy pomocy wektorów cech. Celem pracy jest stworzenie interaktywnej przegladarki graficznej do analizy wyników skalowania wielowymiarowego realizowanego przy pomocy kilku metod. Zarówno silnik jak i przegladarka maja działac w środowisku karty graficznej CUDA. Szczegółowe specyfikacje oprogramowania zostaną sformułowane w trakcie realizacji pracy.
Celem pracy jest stworzenie wydajnej symulacji dynamiki płynów oraz przewodnictwa
cieplnego działajacej w srodowisku przegladarki internetowej. Aplikacja ma mieć
charakter edukacyjny, a jej głównym zadaniem jest wspieranie uzytkownika w dogłebnym
zrozumieniu procesu transferu energii, w tym zjawisk takich jak przewodnictwo cieplne,
konwekcja czy różnorodne przepływy gazów i cieczy.
Growing cities around the world are seeking efficient methods of maintenance of their extensive
infrastructure. Every large city needs a department which role is to collect information
about failures in the city’s infrastructure, manage and schedule repairs. It is desirable to collect
feedback from the city’s residents, to discover and react quickly to irritating malfunctions such
as potholes or graffiti. In the age of smartphones, the common use of mobile devices opens
up new perspectives to build public utility systems. The system should enable straight forward
interactions between the society and the department.
The aim of this master thesis is to create an image categorization software component to
be integrated into a city’s infrastructure maintenance system. The component’s task is to match
an input image to a single, predefined category. Input images are to present failures in a city’s
infrastructure. The scope of the thesis includes thorough algorithmic research, reference implementation,
Multidimensional scaling (MDS) is a very popular and reliable procedure used in visualization
of multidimensional data. The role of MDS is to reconstruct the topology of an original N-
dimensional feature space consisting of M feature vectors in target 2-D (3-D) Euclidean space.
It can be achieved by minimization of the error - \stress" function - F(jjR
The mathematical modeling of natural phenomena is understood as an approximation of a real-world process extracting the underlying physics and leaving unimportant details out of the scope of interest. Following Einstein’s dictum everything should be made as simple as possible, but not simpler. The need for such models emerge from the strive to understand the world. Realistic modeling of natural phenomena are vital in the computer graphics applications for the movies industry. Also a range of medical and engineering areas draw from the numerical models and their solutions. Level Set Method since its emergence as an approach to Hamilton Jacobi problems  have been widely used in areas such as contour capturing for moving objects tracing, image processing, optimizations problems and computer graphics. Despite successive advances in performance Level Set methods are still computationally demanding and their use for large domain problems involves time-consuming calculations. Recent emergence of the general purpose graphical processing units (GP-GPU) enabled optimization possibilities in many computationally demanding areas. This thesis presents an approach to realistic three-dimensional free surface fluid volume simulation with an implementation using the CUDA language for GPU computations and level sets for interface tracking.
This thesis presents new approach for solving coarse-raining problem for Cellular Automata.
The coarse-graining is a process of replacement a fine-rained description with a lower-resolution coarse-grained model. We use coarse-graining procedure in many physical systems, to hide some fine-grained details. The idea of applying this process to Cellular-Automata was introduced by Israeli et al.
in 2006, although solution presented in that paper was not very efficient. We will focus on application of coarse-graining process into cryptographic field and demonstrate new public-key crypto-system based on that concept.
The new cryptographic method will be analyzed in terms of safeness, complexity and probability of finding valid keys. We will try to compare this method with other commonly used public-key cryptography models and other methods based on cellular automata.
At last We would try to present some output from novel fast coarse-graining method which is computationally unreachable by origin approach by Israeli et al. These results might have some impact on exploring complex characteristics of cellular automata.