Refereed conference proceedings

1. An Evaluation of Scale and Noise Sensitivity of Fibre Orientation Estimation in Volume Images
   Author: Maria Axelsson
   Conference: 15th International Conference on Image Analysis and Processing, Salerno, Italy (ICIAP'09), pp. 975-984
   Abstract: Fibre orientation influences many important properties of fibre-based materials, for example, strength and stiffness. Fibre orientation and the orientation anisotropy in paper and other wood fibre-based materials have previously been estimated using two-dimensional images. Recently, we presented a method for estimating the three-dimensional fibre orientation in volume images based on local orientation estimates. Here, we present an evaluation of the method with respect to scale and noise sensitivity. The evaluation is performed for both tubular and solid fibres. We also present a new method for automatic scale selection for solid fibres. The method is based on a segmentation of the fibres that also provides an estimate of the fibre dimension distribution in an image. The results show that the fibre orientation estimation performs well both in noisy images and at different scales. The presented results can be used as a guide to select appropriate parameters for the method when it is applied to real data. The applicability of the fibre orientation estimation to fibre-based materials with solid fibres is demonstrated for a volume image of a press felt acquired with X-ray microtomography.

2. On Maximal Balls in Three Volume Grids
   Authors: Gunilla Borgefors and Robin Strand
   Conference: 10th International Conference on Pattern Recognition and Information Processing, Minsk, Belarus (PRIP'09), pp, 31-36
   Abstract: A volume image can be digitized in different grids, not only the cubic one. The fcc and bcc grids have many advantages, as they are more dense than the cubic one. The set of maximal balls in a shape in a volume image is a compact but complete description of the shape. The original set, identified by rules dependent on the metric used, can be further reduced, by observing that some balls are covered by groups of other balls. The set of maximal balls can, for example, be used for compression, manipulation and as anchor points for topologically correct medial representations.

3. A Heterogeneous Cluster Framework for Computationally Heavy Visualization
   Authors: Martin Ericsson, Anders Hast and Stefan Seipel
   Journal: International Conference Applied Computing, Rome, Italy (IADIS'09), vol. 2, pp. 337-339
   Abstract: It is desirable to use different kinds of hardware for the same software project but it is in a lot of cases not feasible due to the programming effort that it takes to get it running. We have developed a framework to gain more experience with dealing with several different computer architectures in the same project and to have a test suit for different applications. The hardware we used contains both homogeneous x86 multi-core processors as well as heterogeneous Cell Broadband Engine processors. We developed an example application in form of a distributed rendering system which demands a lot of computational power to test it. Also incorporated in this project is a stereoscopic display that put limits on resolution and frame rate of the system. From the lessons learned in this ongoing work we conclude that it is feasible to build and use heterogeneous resources for this kind of application even though it will take a bit of effort.

4. Suppression of Autofluorescence based on Fuzzy Classification by Spectral Angles
   Authors: Milan Gavrilovic and Carolina Wählby
   Conference: 12th International Conference on Medical Image Computing and Computer Assisted Intervention, London, UK (MICCAI'09)
   The workshop Optical Tissue Image analysis in Microscopy, Histopathology and Endoscopy (OPTIMHisE), pp. 135-146
   Abstract: Background fluorescence, also known as autofluorescence, and cross-talk are two problems in fluorescence microscopy that stem from similar phenomena. When biological specimens are imaged, the detected signal often contains contributions from fluorescence originating from sources other than the imaged fluorophore. This fluorescence could either come from the specimen itself (autofluorescence), or from fluorophores with partly overlapping emission spectra (cross-talk). In order to resolve spectral components at least two distinct wavelength intervals have to be imaged. This paper shows how autofluorescence can be presented statistically using a spectral angle histogram. Pixel classification by spectral angles was previously developed for detection and quantification of colocalization. Here we show how the spectral angle histogram can be employed to suppress autofluorescence. First, classical background subtraction (also referred to as linear unmixing) is presented in the form of a fuzzy classification by spectral angles. A modification of the fuzzy classification rules is also presented and we show that sigmoid membership functions lead to better suppression of background and amplification of true signals.

5. Improved Textures for 3D Virtual Reconstruction and Visualization by a Modified Multiscale Texture Synthesis
   Authors: Anders Hast, Martin Ericsson and T. Reiner (1)
   (1) University of Stuttgart, Germany
   Conference: 3rd International Workshop 3D Virtual Reconstruction and Visualization of Complex Architectures, Trento, Italy (3D-ARCH'09), electronic publication
   Abstract: When photos of walls are used in urban 3D visualizations they are often of limited quality due to the fact that it can be very hard, or even impossible, to take close up photos of the whole part of walls, especially for buildings with several floors. Thus walls will appear either pixelized or blurry when the viewer comes close to them. The latter if some kind of interpolation technique is being used to reduce the pixelization. In any case it his has a big impact on how the viewer perceives the 3D environment as it will look far from real. We present how a modified multiscale texture synthesis approach can be used to create highly detailed textures from photos with different levels of detail and scale. The novel idea is to switch colour space in order to improve both quality and speed. By using the HSV space it is possible to maintain colours, especially when the examplar image does not contain all colours present in the target image.

6. Towards Automated TEM for Virus Diagnostics: Segmentation of Grid Squares and Detection of Regions of Interest
   Authors: Gustaf Kylberg, Ida-Maria Sintorn and Gunilla Borgefors
   Conference: 16th Scandinavian Conference Image Analysis, Oslo, Norway (SCIA'09), pp. 169-178
   Abstract: When searching for viruses in an electron microscope the sample grid constitutes an enormous search area. Here, we present methodsfor automating the image acquisition process for an automatic virusdiagnostic application. The methods constitute a multi resolution approach where we first identify the grid squares and rate individual grid squares based on content in a grid overview image and then detect regions of interest in higher resolution images of good grid squares. Our methods are designed to mimic the actions of a virus TEM expert manually navigating the microscope and they are also compared to the expert's performance. Integrating the proposed methods with the microscope would reduce the search area by more than 99.99% and it would also remove the need for an expert to perform the virus search by the microscope.
7. Local Intensity and PCA Based Detection of Virus Particle Candidates in Transmission Electron Microscopy Images
   Authors: Gustaf Kylberg, Ida-Maria Sintorn, Mats Uppström (1) and Martin Ryner (2)
   (1) Vironova AB, Stockholm
   (2) Dept. of Medicine, Centre for Molecular Medicine, Karolinska Institutet, Stockholm; Dept. of Mathematics and NADA, Royal Institute of Technology, Stockholm
   Conference: 6th International Symposium on Image and Signal Processing and Analysis, Salzburg, Austria (ISPA'09), pp. 426-431
   Abstract: We present a general method using local intensity information and PCA to detect objects characterized only by that they differ from their surroundings. We apply ourmethod to the problem of automatically detecting virus particlecandidates in transmission electron microscopy images. Viruses have very different shapes and sizes, many species are spherical whereas others are highly pleomorphic. To detect any kind of virus particles in electron microscopy images it is therefore necessary to use a method not restricted to detection of a specific shape. The method proposed here uses only one input parameter, the approximate virus thickness, which is a conserved feature withina virus species. It is capable to detect virus particles of very varying shapes. Results on images with highly textured background of several different virus species are presented.

8. Improved Quantification of Bone Remodelling by Utilizing Fuzzy Based Segmentation
   Authors: Joakim Lindblad, Natasa Sladoje (1), Vladimir Curic, Hamid Sarve, Carina B Johansson (2) and Gunilla Borgefors
   (1) Faculty of Engineering, University of Novi Sad, Serbia
   (2) Örebro University
   Conference: 16th Scandinavian Conference Image Analysis, Oslo, Norway (SCIA'09), pp. 750-759
   Abstract: We present a novel fuzzy theory based method for the segmentation of images required in histomorphometrical investigations of bone implant integration. The suggested method combines discriminant analysis classification controlled by an introduced uncertainty measure, and fuzzy connectedness segmentation method, so that the former is used for automatic seeding of the later. A thorough evaluation of the proposed segmentation method is performed. Comparison with previously published automatically obtained measurements, as well as with manually obtained ones, is presented. The proposed method improves the segmentation and, consequently, the accuracy of the automatic measurements, while keeping advantages with respect to the manual ones, by being fast, repeatable, and objective.

9. On Set Distances and Their Application to Image Registration
   Authors: Joakim Lindblad, Vladimir Curic and Natasa Sladoje
   Conference: 6th International Symposium on Image and Signal Processing and Analysis, Salzburg, Austria (ISPA'09), pp. 449-454
   Abstract: In this paper we study set distances that are used in image processing. We propose a generalization of Sum of minimal distances and show that its special cases include a metric by Symmetric difference. The Hausdorff metric and the Chamfer matching distances are also closely related with the presented framework. In addition, we define the Complement set distance of a given distance. We evaluate the observed distance with respect to applicability to image object registration. We perform comparative evaluations with respect to noise sensitivity, as well as with respect to rigid body transformations. We conclude that the family of Generalized sum of minimal distances has many desirable properties for this application.

10. Closing Curves with Riemannian Dilation: Application to Segmentation in Automated Cervical Cancer Screening
    Authors: Patrik Malm and Anders Brun
    Conference: 5th International Symposium on Visual Computing, Las Vegas, NV (ISVC'09), part II, pp. 337-346
    Abstract: In this paper, we describe a nuclei segmentation algorithm for Pap smears that uses anisotropic dilation for curve closing. Edge detection methods often return broken edges that need to be closed to achieve a proper segmentation. Our method performs dilation using Riemannian distance maps that are derived from the local structure tensor field in the image. We show that our curve closing improve the segmentation along weak edges and significantly increases the overall performance of segmentation. This is validated in a thorough study on realistic synthetic cell images from our Pap smear simulator. The algorithm is also demonstrated on bright-field microscope images of real Pap smears from cervical cancer screening.

11. Binarization of Phase Contrast Volume Images of Fibrous Materials: A Case Study
    Authors: Filip Malmberg, Catherine Östlund and Gunilla Borgefors
    Conference: 4th International Conference on Computer Vision Theory and Applications, Lisboa, Portugal (VISAPP'09), pp. 148-153
    Abstract: In this paper, we present a method for segmenting phase contrast volume images of fibrous materials into fibre and background. The method is based on graph cut segmentation, and is tested on high resolution X-ray microtomography volume images of wood fibres in paper an composites. The new method produces better results than a standard method based on edge-preserving smoothing and hysteresis thresholding. The most important improvement is that the proposed method handles thick and collapsed fibres more accurately than previous methods.

12. Sub-pixel Segmentation with the Image Foresting Transform
    Authors: Filip Malmberg, Joakim Lindblad and Ingela Nyström
    Conference:  13th International Workshop on Combinatorial Image Analysis, Playa del Carmen, Mexico (IWCIA'09), pp. 201-211
    Abstract: The Image Foresting Transform (IFT) is a framework forimage partitioning, commonly used for interactive segmentation. Givenan image where a subset of the image elements (seed-points) have beenassigned user-defined labels, the IFT completes the labeling by computingminimal cost paths from all image elements to the seed-points. Eachimage element is then given the same label as the closest seed-point. Inits original form, the IFT produces crisp segmentations, i.e., each imageelement is assigned the label of exactly one seed-point. Here, we proposea modified version of the IFT that computes region boundaries withsub-pixel precision by allowing mixed labels at region boundaries. Wedemonstrate that the proposed sub-pixel IFT allows properties of thesegmented object to be measured with higher precision.
13. Neighborhood Sequences in the Diamond Grid - Algorithms with Four Neighbors
    Authors: Benedek Nagy (1) and Robin Strand
    (1) Faculty of Informatics, University of Debrecen, Debrecen, Hungary
    Conference:  13th International Workshop on Combinatorial Image Analysis, Playa del Carmen, Mexico (IWCIA'09), theoretical track, pp. 109-121
    Abstract: In digital image processing digital distances are useful; distances based on neighborhood sequences are widely used. In this paper the diamond grid is considered, that is the three-dimensional grid of Carbon atoms in the diamond crystal. This grid can be described by four coordinate values using axes of the directions of atomic bonds. In this way the sum of the coordinate values can be either zero or one. An algorithm to compute a shortest path defined by a neighborhood sequence between any two points in the diamond grid is presented. The metric and non-metric properties of some distances based on neighborhood sequences are also discussed. The constrained distance transformation and digital balls obtained by some distance functions are presented.

14. Neighborhood Sequences on nD Hexagonal/Face-Centered-Cubic Grids
    Authors: Benedek Nagy (1) and Robin Strand
    (1) Faculty of Informatics, University of Debrecen, Debrecen, Hungary
    Conference:  13th International Workshop on Combinatorial Image Analysis, Playa del Carmen, Mexico (IWCIA'09), theoretical track, pp. 96-108
    Abstract: The two-dimensional hexagonal grid and the three-dimensional face-centered cubic grid can be described by intersecting and with a (hyper)plane. Corresponding grids in higher dimensions (nD) are examined. In this paper, we define distance functions based on neighborhood sequences on these, higher dimensional generalizations of the hexagonal grid. An algorithm to produce a shortest path based on neighborhood sequences between any two gridpoints is presented. A formula to compute distance and condition of metricity are presented for neighborhood sequences using two types of neighbors. Distance transform as an application of these distances is also shown.

15. An Automatic Method for Counting Annual Rings in Noisy Sawmill Images
    Author: Kristin Norell
    Conference: 15th International Conference on Image Analysis and Processing, Salerno, Italy (ICIAP'09), pp. 307-316
    Abstract: The annual ring pattern of a log end face is related to the quality of the wood. We propose a method for computing the number of annual rings on a log end face depicted in sawmill production. The method is based on the grey-weighted polar distance transform and registration of detected rings from two different directions. The method is developed and evaluated on noisy images captured in on-line sawmill production at a Swedish sawmill during 2008, using an industrial colour camera. We have also evaluated the method using synthetic data with different ring widths, ring eccentricity, and noise levels.

16. Creating Synthetic Log End Face Images
    Author: Kristin Norell
    Conference: 6th International Symposium on Image and Signal Processing and Analysis, Salzburg, Austria (ISPA'09), pp. 353-358
    Abstract: In this paper we present the design and creation of synthetic images of log end faces. The images are constructed to resemble images of Scots pine taken in on-line sawmill production. Wood features such as knots, heartwood, and annual rings, as well as the sawing procedure, storage, and imaging, including camera position, are simulated. A dataset of 100 images is provided, together with code for generating new synthetic data.

17. Segmentation and Visualization of 3D Medical Images through Haptic Rendering
    Authors: Ingela Nyström, Filip Malmberg, Erik Vidholm, Ewert Bengtsson
    Conference: 10th International Conference on Pattern Recognition and Information Processing, Minsk, Belarus (PRIP'09), pp. 43-48
    Abstract: High-dimensional and high-resolution image data is increasingly produced by modern medical imaging equipment. As a consequence, the need for efficient interactive tools for segmentation and visualization of these medical images is also increasing. Existing software include state-of-the-art algorithms, but in most cases the interaction part is limited to 2D mouse/keyboard, despite the tasks being highly 3D oriented. This project involves interactive medical image visualization and segmentation, where true 3D interaction is obtained with stereo graphics and haptic feedback. Well-known image segmentation algorithms, e.g., fast marching, fuzzy connectedness, deformable models, and live-wire, have been implemented in a framework allowing the user to interact with the algorithms and the volumetric data in an efficient manner. The data is visualized via multi-planar reformatting, surface rendering, and hardware-accelerated volume rendering. We present a case study where liver segmentation is performed in CT images with high accuracy and precision.

18. Quantification of Bone Remodeling in SR µCT Images of Implants
    Authors: Hamid Sarve, Joakim Lindblad and Carina B. Johansson (1)
    (1) Örebro University
    Conference: 16th Scandinavian Conference Image Analysis, Oslo, Norway (SCIA'09), pp. 770-779
    Abstract: For quantification of bone remodeling around implants, wecombine information obtained by two modalities: 2D histological sectionsimaged in light microscope and 3D synchrotron radiation-based computedmicrotomography, SRµCT. In this paper, we present a methodfor segmenting SµCT volumes. The impact of shading artifact at theimplant interface is reduced by modeling the artifact. The segmentationis followed by quantitative analysis. To facilitate comparison with existingresults, the quantification is performed on a registered 2D slicefrom the volume, which corresponds to a histological section from thesame sample. The quantification involves measurements of bone areaand bone-implant contact percentages.We compare the results obtained by the proposed method on the SµCTdata with manual measurements on the histological sections and discussthe advantages of including SRµCT data in the analysis.

19. Designing Efficient Visualizations for Applications in the Paper and Pulp Industry
    Authors: Stefan Seipel and Ann-Kristin Forsberg (1)
    (1) Dept. of Mathematics, Natural- and Computer Science, Gävle University
    Conference: International Conference Applied Computing, Rome, Italy (IADIS'09), vol. 1, pp. 403-410
    Abstract: 2D and 3D visualization has become a rapidly growing area of research during the past years. For a long time advanced graphical techniques where mainly used within the scientific community, however, their potential benefits are now increasingly recognized also for industrial applications. The visualization community is meanwhile following a theoryanchored and evaluation based approach paving the way for the design of perceptually efficient visualizations. In this paper we describe the development, evaluation and deployment of efficient visualizations to support process operators in the paper and pulp process industry. We first identify and describe the data analysis task of the real working situation at hand. Starting from existing theory in the field of visualization and vision esearch, we then describe the process of designing perceptually motivated new visualizations for the specific task of the operators. The result is a new deviation color scale that we apply to 2D color map and 3D height-field representations. We then describe an experiment to formally evaluate the efficiency of these visualizations for the visual detection of thermal overheating of rotary kilns. The results of this study showed that our new differential color scale lead to significantly reduced detection times when compared with traditional color coding schemes. Also, when conventional color scales are used for the visualization of absolute temperature levels, the inclusion of the 3D cues in the visualization for the visual encoding of the rate of change contributes to faster detection of temperature increases.

20. On the Applicability of Direct and Indirect Input in Table-Top and Vertical Displays
    Authors: Stefan Seipel and Ann-Kristin Forsberg (1)
    (1) Dept. of Mathematics, Natural- and Computer Science, Gävle University
    Conference: International Conference Applied Computing, Rome, Italy (IADIS'09), vol. 2, pp. 8-11
    Abstract: Recent developments in display and interaction technologies afford new IT based applications e.g. for use cases with several collaborating users. Table-top display environments are among those techniques that are expected to boost the development of an entire range of new computer applications. While technological development has advanced rapidly, little research has been done to evaluate the applicability of such techniques for general interaction. In this paper we present latest results of a study on the efficiency of direct interaction and indirect nteraction with a table-top display and we compare it with indirect input on a conventional vertical display. The results show that users made significantly shorter cursor movements in the table-top interface than in the conventional vertical interface. In regard to time, task performance for direct, pen-based input on a table-top interface was even superior to conventional, mouse based interaction on a vertical display. Users performance was worst for mouse based interaction in the table-top condition despite shorter travel distances. The results of our work give some interesting criteria for the development of applicable systems. That is to say, contrary to earlier research we find that the use of a table-top display yields to more efficient interaction than a comparable vertical display as long as direct input is used.

21. Prediction of Ventilation Paths in Urban Environments using Digitized Maps
    Authors: Stefan Seipel and Julia Åhlén (1)
    (1) Dept. of Computer Science, Gävle University
    Conference: International Conference Applied Computing, Rome, Italy (IADIS'09), vol. 2, pp. 217-221
    Abstract:  Urban populations and the density of settlements constantly become higher. This leads to higher energy consumption and generally to deterioration of life comfort. Study of urban heat and cool islands is of great importance for social community planners and building engineers. Ventilation paths are defined by turbulent mass, momentum and energy transport conditions and can thus be modeled. This area is usually studied by measurements of the conditions and air flows in laboratory environments. This paper presents a method for the prediction of free ventilation paths in a small city using digital imagery. A digitized map created from a eographic data base is used as input. Image analysis is performed in order to create an optimal edge image. A modified Hough transform is applied. Points of interest are defined and surroundings are calculated. These points are inputs to a parameter space. As a result a free wind passage is predicted based on the position of the observer. Prediction is done by calculation of possible straight lines in a parameter space. Finally, the method is verified by comparison with position vectors from the same space in the image and the best fitted path is chosen.

22. A collaborative visualization environment for natural interaction with architectural content
    Authors: Stefan Seipel and Ann-Kristin Forsberg (1)
    (1) Dept. of Mathematics, Natural- and Computer Science, Gävle University
    Conference: 3rd International Workshop 3D Virtual Reconstruction and Visualization of Complex Architectures, Trento, Italy (3D-ARCH'09), electronic publication
    Abstract: Interactive exploration and assessment of architectural design studies or reconstructions of historical built environments have for centuries been based on physical models from wood, plaster or cardboard. With the development of powerful 3D graphics functionality on personal computers, digital models of complex architecture (constructed or digitized) can be visualized and be explored interactively by means of advanced 3D computer displays. Virtual Reality based experiences can be used efficiently to provide detail views by means of virtual architectural walkthroughs as well as facilitate contextual views by adopting a birds-eye metaphor upon the data. One of the drawbacks of many 3D architectural presentations is that a correct 3D perspective is presented for a virtual camera with one predefined central perspective projection. In consequence only one if any of several observers benefits from a spatially correct view of the virtual scenery. In addition, for 3D presentations on vertical computer screens, natural interaction between two ore more collaborating users is hampered as direct face-to-face communication is distracted. In this paper we present results of our most recent development and ongoing work towards a more usable tabletop display system for two collaborating users and we present its application in the visualization of public buildings and historic environments. What renders our display environment specific is a combination of several features that make it feasible for everyday use: The technical design of our system allows for a compact form factor allowing the system to be used in everyday office situations. The system is capable of providing a dynamic stereoscopic 3D view for one moving observer, or alternatively monoscopic dynamic 3D views for two independently moving observers. This visualization environment is based on rear-projection and it incorporates an optical film into the screen which allows for high-resolution multi-point interaction at pixel accuracy using several optical digital pens that communicate wireless with the computer. In this paper we present the technical design of the system as well as its use in the visual assessment of building structures and presentation of pre-historical architecture. The main and most novel contributions of this paper are the results of an experimental study that investigated performance differences between natural pen-based direct interaction versus tradition mouse-based interaction in this new visualization environment.

23. Segmentation of Highly Lignified Zones in Wood Fiber Cross-Sections
    Authors: Bettina Selig, Cris L. Luengo Hendriks, Stig Bardage (1) and Gunilla Borgefors
    (1) Dept. of Forest Products, SLU
    Conference: 16th Scandinavian Conference Image Analysis, Oslo, Norway (SCIA'09), pp. 369-378
    Abstract: Lignification of wood fibers has important consequences tothe paper production, but its exact effects are not well understood. Tocorrelate exact levels of lignin in wood fibers to their mechanical proper-ties, lignin autofluorescence is imaged in wood fiber cross-sections. Highlylignified areas can be detected and related to the area of the whole cellwall. Presently these measurements are performed manually, which is te-dious and expensive. In this paper a method is proposed to estimate thedegree of lignification automatically. A multi-stage snake-based segmen-tation is applied on each cell separately. To make a preliminary evaluationwe used an image which contained 17 complete cell cross-sections. Thisimage was segmented both automatically and manually by an expert.There was a highly significant correlation between the two methods, al-though a systematic difference indicates a disagreement in the definitionof the edges between the expert and the algorithm.

24. Pixel Coverage Segmentation for Improved Feature Estimation
    Authors: Natasa Sladoje (1) and Joakim Lindblad
    (1) Faculty of Engineering, University of Novi Sad, Serbia
    Conference: 15th International Conference on Image Analysis and Processing, Salerno, Italy (ICIAP'09), pp. 923-938
    Abstract: By utilizing intensity information available in images, partial coverage of pixels at object borders can be estimated. Such information can, in turn, provide more precise feature estimates. We present a pixel coverage segmentation method which assigns pixel values corresponding to the area of a pixel that is covered by the imaged object(s). Starting from any suitable crisp segmentation, we extract a one-pixel thin 4-connected boundary between the observed image components where a local linear mixture model is used for estimating fractional pixel coverage values. We evaluate the presented segmentation method, as well as its usefulness for subsequent precise feature estimation, on synthetic test objects with increasing levels of noise added. We conclude that for reasonable noise levels the presented method outperforms the achievable results of a perfect crisp segmentation. Finally, we illustrate the application of the suggested method on a real histological colour image.

25. Recovering Affine Deformations of Fuzzy Shapes
    Authors: Attila Tenacs (1), Csaba Domokos (1), Natasa Sladoje (2), Joakim Lindblad and Zoltan Kato (1)
    (1) Dept. of Image Processing and Computer Graphics, University of Szeged, Hungary
    (2) Faculty of Engineering, University of Novi Sad, Serbia
    Conference: 16th Scandinavian Conference Image Analysis, Oslo, Norway (SCIA'09), pp. 735-744
    Abstract: Fuzzy sets and fuzzy techniques are attracting increasing attention nowadays in the field of image processing and analysis. It has been shown that the information preserved by using fuzzy representation based on area coverage may be successfully utilized to improve precision and accuracy of several shape descriptors; geometric moments of a shape are among them. We propose to extend an existing binary shape matching method to take advantage of fuzzy object representation. The result of a synthetic test show that fuzzy representation yields smaller registration errors in average. A segmentation method is also presented to generate fuzzy segmentations of real images. The applicability of the proposed methods is demonstrated on real X-ray images of hip replacement implants.

26. Generating synthetic µCT images of wood fibre materials
    Authors: Erik Wernersson, Cris L. Luengo Hendriks and Anders Brun
    Conference: 6th International Symposium on Image and Signal Processing and Analysis, Salzburg, Austria (ISPA'09)
    Abstract: X-ray Computerized Tomography at micrometer resolution (µCT) is an important tool for understanding the properties of wood fibre materials such as paper, carton and wood fibre composites. While many image analysis methods have been developed for µCT images in wood science, the evaluation of these methods if often not thorough enough because of the lack of a dataset with ground truth. This paper describes the generation of synthetic µCT volumes of wood fibre materials. Fibres with a high degree of morphological variations are modeled and densely packed into a volume of the material. Using a simulation of the µCT image acquisition process, realistic synthetic images are obtained. This simulation uses noise characterized from a set of µCT images. The synthetic images have a known ground truth, and can therefore be used when evaluating image analysis methods.

27. Segmentation of Wood Fibres in 3D CT Images Using Graph Cuts
    Authors: Erik Wernersson, Anders Brun and Cris L. Luengo Hendriks
    Conference: 15th International Conference on Image Analysis and Processing, Salerno, Italy (ICIAP'09)
    Abstract: To completely segment all individual wood fibres in volume images of fibrous materials presents a challenging problem but is important in understanding the micro mechanical properties of composite materials. This paper presents a filter that identifies and closes pores in wood fibre walls, simplifying the shape of the fibres. After this filter, a novel segmentation method based on graph cuts identifies individual fibres. The methods are validated on a realistic synthetic fibre data set and then applied on µCT images of wood fibre composites.