Volume 19, number 1-3 (2011)

Permanent URI for this collection

http://hdl.handle.net/11025/1089

Browse

Showing 1 - 15 out of 15 results

Generalized heat kernel signatures
(Václav Skala - UNION Agency, 2011) Zobel, Valentin; Reininghaus, Jan; Hotz, Ingrid; Skala, Václav
In this work we propose a generalization of the Heat Kernel Signature (HKS). The HKS is a point signature derived from the heat kernel of the Laplace-Beltrami operator of a surface. In the theory of exterior calculus on a Riemannian manifold, the Laplace-Beltrami operator of a surface is a special case of the Hodge Laplacian which acts on r-forms, i. e. the Hodge Laplacian on 0-forms (functions) is the Laplace-Beltrami operator. We investigate the usefulness of the heat kernel of the Hodge Laplacian on 1-forms (which can be seen as the vector Laplacian) to derive new point signatures which are invariant under isometric mappings. A similar approach used to obtain the HKS yields a symmetric tensor field of second order; for easier comparability we consider several scalar tensor invariants. Computed examples show that these new point signatures are especially interesting for surfaces with boundary.
Pipeline reconstruction from fisheye images
(Václav Skala - UNION Agency, 2011) Zhang, Yuhang; Hartley, Richard; Mashford, John; Wang, Lei; Burn, Stewart; Skala, Václav
Automatic inspection of pipelines has great potential to increase the efficiency and objectivity of pipeline condition assessment. 3-D pipeline reconstruction aims to reveal the deformation of the pipe surface caused by internal or external influences. We present a system which can reconstruct the inner surface of buried pipelines from multiple fisheye images captured inside the pipes. Whereas the pipelines are huge, a fatal defect can be as tiny as a fine crack. Therefore a reliable system demands both efficiency and accuracy. The repetitive patterns on the pipe surface and the poor illumination condition during photographing further increase the difficulty of the reconstruction. We combine several successful methods found in the literature as well as new methods proposed by ourselves. The proposed system can reconstruct pipe surface not only accurately but also quickly. Experiments have been carried out on real pipe images and show promising performance.
High-performance terrain rendering using hardware tessellation
(Václav Skala - UNION Agency, 2011) Yusov, Edgar; Shevtsov, Maxim; Skala, Václav
In this paper, we present a new terrain rendering approach, with adaptive triangulation performed entirely on the GPU via tessellation unit available on the DX11-class graphics hardware. The proposed approach avoids encoding of the triangulation topology thus reducing the CPU burden significantly. It also minimizes the data transfer overhead between host and GPU memory, which also improves rendering performance. During the preprocessing, we construct a multiresolution terrain height map representation that is encoded by the robust compression technique enabling direct error control. The technique is efficiently accelerated by the GPU and allows the trade-off between speed and compression performance. At run time, an adaptive triangulation is constructed in two stages: a coarse and a fine-grain one. At the first stage, rendering algorithm selects the coarsest level patches that satisfy the given error threshold. At the second stage, each patch is subdivided into smaller blocks which are then tessellated on the GPU in the way that guarantees seamless triangulation.
High-quality cartographic roads on high-resolution DEMs
(Václav Skala - UNION Agency, 2011) Vaaraniemi, Michael; Treib, Marc; Westermann, Rüdiger; Skala, Václav
The efficient and high quality rendering of complex road networks—given as vector data—and high-resolution digital elevation models (DEMs) poses a significant problem in 3D geographic information systems. As in paper maps, a cartographic representation of roads with rounded caps and accentuated clearly distinguishable colors is desirable. On the other hand, advances in the technology of remote sensing have led to an explosion of the size and resolution of DEMs, making the integration of cartographic roads very challenging. In this work we investigate techniques for integrating such roads into a terrain renderer capable of handling high-resolution data sets. We evaluate the suitability of existing methods for draping vector data onto DEMs, and we adapt two methods for the rendering of cartographic roads by adding analytically computed rounded caps at the ends of road segments. We compare both approaches with respect to performance and quality, and we outline application areas in which either approach is preferable.
Fast and memory efficient feature detection using multiresolution probabilistic boosting trees
(Václav Skala - UNION Agency, 2011) Schulze, Florian; Major, David; Bühler, Katja; Skala, Václav
This paper presents a highly optimized algorithm for fast feature detection in 3D volumes. Rapid detection of structures and landmarks in medical 3D image data is a key component for many medical applications. To obtain a fast and memory efficient classifier, we introduce probabilistic boosting trees (PBT) with partial cascading and classifier sorting. The extended PBT is integrated into a multiresolution scheme, in order to improve performance and works on block cache data structure which optimizes the memory footprint. We tested our framework on real world clinical datasets and showed that classical PBT can be significantly speeded up even in an environment with limited memory resources using the proposed optimizations.
Robust feature point matching in general multi-image
(Václav Skala - UNION Agency, 2011) Sellent, Anita; Eisemann, Martin; Magnor, Marcus; Skala, Václav
We present a robust feature matching approach that considers features from more than two images during matching. Traditionally, corners or feature points are matched between pairs of images. Starting from one image, corresponding features are searched in the other image. Yet, often this two-image matching is only a subproblem and actually robust matches over multiple views and/ or images acquired at several instants in time are required. In our feature matching approach we consider the multi-view video data modality and find matches that are consistent in three images. Requiring neither calibrated nor synchronized cameras, we are able to reduce the percentage of wrongly matched features considerably. We evaluate the approach for different feature detectors and their natural descriptors and show an application of our improved matching approach for optical flow calculation on unsynchronized stereo sequences.
Multiscale visualization of 3D geovirtual environments using view-dependent multi-perspective views
(Václav Skala - UNION Agency, 2011) Pasewaldt, Sebastian; Trapp, Matthias; Döllner, Jürgen; Skala, Václav
3D geovirtual environments (GeoVEs), such as virtual 3D city models or landscape models, are essential visualization tools for effectively communicating complex spatial information. In this paper, we discuss how these environments can be visualized using multi-perspective projections [10, 13] based on view-dependent global deformations. Multi-perspective projections enable 3D visualization similar to panoramic maps, increasing overview and information density in depictions of 3D GeoVEs. To make multi-perspective views an effective medium, they must adjust to the orientation of the virtual camera controlled by the user and constrained by the environment. Thus, changing multi-perspective camera configurations typically require the user to manually adapt the global deformation — an error prone, non-intuitive, and often time-consuming task. Our main contribution comprises a concept for the automatic and view-dependent interpolation of different global deformation preset configurations (Fig. 1). Applications and systems that implement such view-dependent global deformations, allow users to smoothly and steadily interact with and navigate through multi-perspective 3D GeoVEs.
Analysis and design of the dynamical stability of collective behavior in crowds
(Václav Skala - UNION Agency, 2011) Mukovskiy, Albert; Slotine, Jean-Jacques E.; Giese, Martin A.; Skala, Václav
The modeling of the dynamics of the collective behavior of multiple characters is a key problem in crowd animation. Collective behavior can be described by the solutions of large-scale nonlinear dynamical systems that describe the dynamical interaction of locomoting characters with highly nonlinear articulation dynamics. The design of the stability properties of such complex multi-component systems has been rarely studied in computer animation. We present an approach for the solution of this problem that is based on Contraction Theory, a novel framework for the analysis of the stability complex nonlinear dynamical systems. Using a learning-based realtime-capable architecture for the animation of crowds, we demonstrate the application of this novel approach for the stability design for the groups of characters that interact in various ways. The underlying dynamics specifies control rules for propagation speed and direction, and for the synchronization of the gait phases. Contraction theory is not only suitable for the derivation of conditions that guarantee global asymptotic stability, but also of minimal convergence rates. Such bounds permit to guarantee the temporal constraints for the order formation in self-organizing interactive crowds.
Deducing explicit from implicit visibility for global illumination with antiradiance
(Václav Skala - UNION Agency, 2011) Mückl, Gregor; Dachsbacher, Carsten; Skala, Václav
The antiradiance method, a variant of radiosity, allows the computation of global illumination solutions without determining visibility between surface patches explicitly, unlike the original radiosity method. However, this method creates excessively many links between patches, since virtually all elements exchange positive and negative energy whose interplay replaces the visibility tests. In this paper we study how and if explicit visibility information can be recovered only by analyzing the link mesh to identify chains of links whose light transport cancels out. We describe heuristics to reduce the number of links by extracting visibility information, still without resorting to explicit visibility tests, e.g. using ray casting, and use that in combination with the remaining implicit visibility information for rendering. Further, to prevent the link mesh from growing excessively in large scenes in the beginning, we also propose a simple means to let graphic artists define blocking planes as a way to support our algorithm with coarse explicit visibility information. Lastly, we propose a simple yet efficient image-space approach for displaying radiosity solutions without any meshing for interpolation.
Interactive volume rendering aurora on the GPU
(Václav Skala - UNION Agency, 2011) Lawlor, Orion Sky; Genetti, Joe; Skala, Václav
We present a combination of techniques to render the aurora borealis in real time on a modern graphics processing unit (GPU). Unlike the general 3D volume rendering problem, an auroral display is emissive and can be factored into a height-dependent energy deposition function, and a 2D electron flux map. We also present a GPU-friendly atmosphere model, which includes an integrable analytic approximation of the atmosphere’s density along a ray. Together, these techniques enable a modern consumer graphics card to realistically render the aurora at 20–80fps, from any point of view either inside or outside the atmosphere.
Plausible and realtime rendering of scratched metal by deforming MDF of normal mapped anisotropic surface
(Václav Skala - UNION Agency, 2011) Kang, Young-Min; Cho, Hwan-Gue; Kim, Sung-Soo; Skala, Václav
An effective method to render realistic metallic surface in realtime application is proposed. The proposed method perturbs the normal vectors on the metallic surface to represent small scratches. General approach to the normal vector perturbation is to use bump map or normal map. However, the bumps generated with those methods do not show plausible reflectance when the surface is modeled with a microfacet-based anisotropic BRDF. Because the microfacet-based anisotropic BRDFs are generally employed in order to express metallic surface, the limitation of the simple normal mapping or other normal vector perturbation techniques make it difficult to render realistic metallic object with various scratches. The proposed method employs not only normal perturbation but also deformation of the microfacet distribution function (MDF) that determines the reflectance properties on the surface. The MDF deformation enables more realistic rendering of metallic surface. The proposed method can be easily implemented with GPU programs, and works well in realtime environments.
Evaluation of the linear box-spline filter from trilinear texture samples: a feasibility study
(Václav Skala - UNION Agency, 2011) Domonkos, Balázs; Csébfalvi, Balázs; Skala, Václav
The major preference for applying B-spline filtering rather than non-separable box spline filtering on the BCC lattice is the fact that separable filtering can be performed more efficiently on current GPUs due to the utilization of the hardware-accelerated trilinear texture fetching. In order to make a fair comparison, a similar, efficient evaluation scheme is required that uses trilinear texture fetches instead of nearest-neighbor ones also for the box splines. Thus, in this paper, we propose an evaluation scheme for the linear BCC box spline built upon a trilinear B-spline basis. We compare our trilinearly evaluated linear box spline scheme to the latest method, that uses twice as many nearest neighbor fetches. Then we give a comparison to the major competitive methods: the BCC B-spline filtering and the BCC DC-spline filtering in terms of their performance.
Accurate real-time multi-camera stereo-matching on the GPU for 3D reconstruction
(Václav Skala - UNION Agency, 2011) Denker, Klaus; Umlauf, Georg; Skala, Václav
Using multi-camera matching techniques for 3d reconstruction there is usually the trade-off between the quality of the computed depth map and the speed of the computations. Whereas high quality matching methods take several seconds to several minutes to compute a depth map for one set of images, real-time methods achieve only low quality results. In this paper we present a multi-camera matching method that runs in real-time and yields high resolution depth maps. Our method is based on a novel multi-level combination of normalized cross correlation, deformed matching windows based on the multi-level depth map information, and sub-pixel precise disparity maps. The whole process is implemented completely on the GPU. With this approach we can process four 0.7 megapixel images in 129 milliseconds to a full resolution 3d depth map. Our technique is tailored for the recognition of non-technical shapes, because our target application is face recognition.
Ambient occlusion opacity mapping for visualization of internal molecular structure
(Václav Skala - UNION Agency, 2011) Borland, David; Skala, Václav
Molecular surfaces often exhibit a complicated interior structure that is not fully visible from exterior viewpoints due to occlusion. In many cases this interior cavity is the most important feature of the surface. Applying standard blended transparency can reveal some of the hidden structure, but often results in confusion due to impaired surface-shape perception. We present ambient occlusion opacity mapping (AOOM), a novel visualization technique developed to improve understanding of the interior of molecular structures. Ambient occlusion is a shading method used in computer graphics that approximates complex shadows from an ambient light source by rendering objects darker when surrounded by other objects. Although ambient occlusion has previously been applied in molecular visualization to better understand surface shape, we instead use ambient occlusion information to determine a variable opacity at each point on the molecular surface. In this manner, AOOM enables rendering interior structures more opaque than outer structures, displaying the inner surface of interest more effectively than with constant-opacity blending. Furthermore, AOOM works for cases not handled by previous cavity-extraction techniques. This work has been driven by collaborators studying enzyme-ligand interactions, in which the active site of the enzyme is typically formed as a cavity in the molecular surface. In this paper we describe the AOOM technique and extensions, using visualization of the active site of enzymes as the driving problem.
A caching approach to real-time procedural generation of cities from GIS data
(Václav Skala - UNION Agency, 2011) Cullen, Brian; O’Sullivan, Carol; Skala, Václav
This paper presents a method for real-time generation of detailed procedural cities. Buildings are generated as needed from real GIS data, using modern techniques that can generate realistic content and without having a huge impact on the rendering system. The system uses a client-server approach allowing multiple clients to generate any part of the city the user wishes without requiring the full data-set, or any pre-generated models. The paper introduces the use of object oriented shape grammars to reduce redundant code and presents a parallel cache to allow real-time generation of detailed cities.

Browse

Recent Submissions