No description available.

No description available.

Skeletonization and segmentation are two important techniques for object representation and analysis. Skeletonization algorithm extracts the centre-lines of an object and uses them to efficiently represent the object. It has many applications in various areas, such as computer-aided design, computer-aided engineering, and virtual reality. Segmentation algorithm locates the target object or Region Of Interest (ROI) from images. It has been widely applied to medical image analysis and many other areas. This thesis presents two studies in skeletonization and two studies in segmentation that advanced the state-of-the-art research. The first skeletonization study suggests an improvement of an existing algorithm for connectivity preservation, which is one of the fundamental requirements for skeletonization algorithms. The second skeletonization study proposes a method to generate curve skeletons with unit-width, which is required by many applications. The first segmentation study presents a new approach named Flexible Vector Flow (FVF) to address a few problems of other active contour models such as insufficient capture range and poor convergence for concavities. This approach was applied to brain tumor segmentation in two dimensional (2D) space. The second segmentation study extends the 2D FVF algorithm to three-dimension (3D) and utilizes it to automatically segment brain tumors in 3D.

A Transformada Distância é uma importante ferramenta para o processamento de imagens digitais. A partir dela podemos calcular a dimensão fractal e obter o esqueleto de um objeto. Estas operações são muito importantes na análise de formas e no reconhecimento de padrões. Porém poucas são as implementações em hardware específico para este processamento. Neste trabalho apresentamos a implementação de arquiteturas paralelas para a determinação da Transformada Distância e para a geração de esqueletos baseados no algoritmo de Dilatação Exata e Propagação de Rótulos. Propomos também uma implementação do algoritmo utilizando a biblioteca MPI para o processamento paralelo. / Distance Transform is an important tool for digital imaging processing. Using the Distance Transform it is possible to evaluate the fractal dimension and skeletons of objects. Fractal dimension and skeletons are very important in shape analysis and pattern recognition operations, few are the hardware implementation for these operations. In this work we present a parallel implementation based on the exact dilation and label propagation for the evaluation of the Distance Transform and skeletons generation. We also propose an algorithm implementation using the MPI library for parallel processing.

Honeybees have a rich genetic diversity in Anatolia. This is reflected in the presence of numerous subspecies of honeybee in Turkey. In METU, Department of Biology, honeybee populations of different regions in Turkey are investigated in order to characterize population variation in these regions. A total of 23 length and angle features belonging to the honeybee hindlegs and forewings are measured in these studies using a microscope and a monitor. These measurements are carried out by placing rulers on the monitor that shows the honeybee image and getting the length and angle features. However, performing measurements in this way is a time consuming process and is open to human-dependent errors. In this thesis, a &ldquo / semi-automated honeybee feature extraction system&rdquo / is presented. The aim is to increase the efficiency by decreasing the time spent on handling these measurements and by increasing the accuracy of measured hindleg and forewing features. The problem is studied from the acquisition of the microscope images, to the feature extraction of the honeybee features. In this scope, suitable methods are developed for segmentation of honeybee hindleg and forewing images. Within intermediate steps, blob analysis is utilized, and edges of the forewing and hindlegs are thinned using skeletonization. Templates that represent the forewing and hindleg edges are formed by either Bezier Curves or Polynomial Interpolation. In the feature extraction phase, Active Contour (Snake) algorithm is applied to the images in order to find the critical points using these templates.

A Transformada Distância é uma importante ferramenta para o processamento de imagens digitais. A partir dela podemos calcular a dimensão fractal e obter o esqueleto de um objeto. Estas operações são muito importantes na análise de formas e no reconhecimento de padrões. Porém poucas são as implementações em hardware específico para este processamento. Neste trabalho apresentamos a implementação de arquiteturas paralelas para a determinação da Transformada Distância e para a geração de esqueletos baseados no algoritmo de Dilatação Exata e Propagação de Rótulos. Propomos também uma implementação do algoritmo utilizando a biblioteca MPI para o processamento paralelo. / Distance Transform is an important tool for digital imaging processing. Using the Distance Transform it is possible to evaluate the fractal dimension and skeletons of objects. Fractal dimension and skeletons are very important in shape analysis and pattern recognition operations, few are the hardware implementation for these operations. In this work we present a parallel implementation based on the exact dilation and label propagation for the evaluation of the Distance Transform and skeletons generation. We also propose an algorithm implementation using the MPI library for parallel processing.

In the past years the so-called body-centered cubic grid (bcc) has been examined and proved to be superior over Cartesian lattices for certain applications. Our work deals with parallel thinning on these bcc grids. We introduce conditions which are sufficient for retaining topology and suggest additional conditions to influence the shape of the resulting skeleton. We further developed an algorithm to extract curve skeletons out of 3d objects in parallel which we also present here. We show in our results that the developed thinning approach on bcc grids is extremely efficient.

Master’s thesis proposes a new algorithm which enables efficient conversion of graphical representation of warehouse into graph theory representation and consequently accelerates estimation for route costs. The proposed algorithm computes route distances between any places in warehouse based on Breadth first search, image processing „skeletonization“ and Dijkstra algorithm. Using the proposed algorithm it is possible to search routes in a warehouse effectively and fast using precomputed routing table. Searching time is less then milisecond using routing table and even size of warehouse doesn’t affect it significantly instead of using Dijkstra algorithm.

Corrosion-casted capillary systems of the kidney glomerulus were imaged with confocal microscopy because of the fluorescence properties of the casting plastic. Acquisition of a z-series through the glomerular capillaries provided three-dimensional data sets from which surface-rendered models were generated. These models could be rotated and viewed from any angle and also contained quantitative information allowing cast surface area and volume measurements to be calculated. The computer-generated models were also skeletonized to form a one-voxel-thick skeleton of the original model. The skeleton exhibited the three-dimensional topology and network of the capillary bed, and interior capillary relations could also be viewed. Quantitative information such as the total capillary length and number of capillary intersects was calculated from the skeletonized model. Extending this method to noncorroded kidney specimens revealed not only the casted vessels but also cellular features of the adjacent tissues surrounding the capillaries.

Ce travail a pour but de concevoir des outils d'analyse d'image 3D de matériaux poreux, obtenues par microtomographie à rayons X, afin de caractériser géométriquement la structure micronique des pores et de mettre en évidence le lien entre microgéométrie et propriétés de transport macroscopiques. Partant d'une image segmentée, une séquence complète de traitements (filtrage d'artefacts, squelettisation, LPE, etc.) est proposée pour positionner et délimiter les pores. Une comparaison aux techniques existantes est faite, et une méthodologie qualifiant la robustesse des procédures est présentée. Cette décomposition est utilisée, premièrement pour extraire des descripteurs géométriques de la microstructure porale qui sont examinés en rapport avec la perméabilité intrinsèque ; deuxièmement pour aider à la construction d’un réseau de pores permettant d’effectuer des simulations numériques. / The objective of this work is to develop 3D image analysis tools to study the micronic pore structure of porous materials, obtained by X-ray microtomography, and study the relation between microgeometry and macroscopic transport properties. From a binarised image of the pore space, a complete sequence of processing (artefact filtration, skeletonisation, watershed, etc. ) is proposed for positioning and delimiting the pores. A comparison with available methods is performed, and a methodology to qualify the robustness of these processes is presented. The decomposition is used, firstly for extracting geometric parameters of the porous microstructure and studying the relation with intrinsic permeability; secondly to produce a simplified pore network on which to perform numerical simulations.