WSCG '2007: Full Papers Proceedings

Permanent URI for this collection

http://hdl.handle.net/11025/10981

Browse

Showing 1 - 20 out of 42 results

SIFT Implementation and Optimization for General-Purpose GPU
(Václav Skala - UNION Agency, 2007) Heymann, S.; Müller, K.; Smolic, A.; Fröhlich, B.; Wiegand, T.; Rossignac, Jarek; Skala, Václav
With the addition of free programmable components to modern graphics hardware, graphics processing units (GPUs) become increasingly interesting for general purpose computations, especially due to utilizing parallel buffer processing. In this paper we present methods and techniques that take advantage of modern graphics hardware for real-time tracking and recognition of feature-points. The focus lies on the generation of feature vectors from input images in the various stages. For the generation of feature-vectors the Scale Invariant Feature Transform (SIFT) method [Low04a] is used due to its high stability against rotation, scale and lighting condition changes of the processed images. We present results of the various stages for feature vector generation of our GPU implementation and compare it to the CPU version of the SIFT algorithm. The approach works well on Geforce6 series graphics board and above and takes advantage of new hardware features, e.g. dynamic branching and multiple render targets (MRT) in the fragment processor [KF05]. With the presented methods feature-tracking with real time frame rates can be achieved on the GPU and meanwhile the CPU can be used for other tasks.
Fast Colorization Using Edge and Gradient Constrains
(Václav Skala - UNION Agency, 2007) Li, Yao; Lizhuang, Ma; Di, Wu; Rossignac, Jarek; Skala, Václav
This paper proposes a fast but effective user-guided colorization algorithm. The main difficulty with colorization is its intensive computational cost. And the color sometimes diffuses from one region to others. We make full use of the information given by the gray-scale images, including the edge, gradient and gradient direction, to propagate color over regions from the user’s scribbles. We introduce a novel local distance definition which reduces the color confusion between two regions obviously. Two-Dimensional Programming is used to get the minimum distance from the scribbles to every pixel, and every pixel is blended by the chrominance with top three minimum distances. This speeds up the colorization process. Our algorithm can also be extended to recolorization and movie colorization. Experimental results show that our algorithm outperforms many state-ofthe- art algorithms from small amount of user scribbles. Besides, the implementation require less than 1 second on the image with 320*240 pixels.
Visual Attention for Significantly Influencing the Perception of Virtual Environments
(Václav Skala - UNION Agency, 2007) Hasic, Jasminka; Chalmers, Alan; Rossignac, Jarek; Skala, Václav
The Human Visual System (HVS) is a key part of the rendering pipeline. The human eye is only capable of sensing image detail in a 2 foveal region, relying on rapid eye movements, or saccades, to jump between points of interest. These points of interest are prioritised based on the saliency of the objects in the scene or the task the user is performing. These ”glimpses” of a scene are then assembled by the HVS into a coherent, but inevitably imperfect, visual perception of the environment. In this process, much detail, which the HVS deems unimportant, may literally go unnoticed. In this paper we use knowledge of the HVS to influence what our attention is attracted to in computer graphics imagery, and thus what we actually perceive in those images. We influence the affinity of subjects towards an object based on the complexity of the context that object is put into. The images are rendered using the Radiance lighting simulation system. In this way, we are able to significantly influence users’ preferences in an e-commerce application. Detailed psychophysical studies are used to validate our approach.
Generation of Shadows in Scene Graph based VR
(Václav Skala - UNION Agency, 2007) Kuehl, Bjoern; Blom, Kristopher J.; Beckhaus, Steffi; Rossignac, Jarek; Skala, Václav
In this paper, we present our experience implementing shadows in Scene Graph based Virtual Reality systems. Shadows are an important part of the human perception of both shape and depth and, yet, are a largely missing component in Virtual Environments. In this work, we investigate extending standard Scene Graphs to automatically produce shadowed scenes. This paper presents our experience embedding two popular real-time shadow methods, Shadow Mapping and Stencil Shadow Volumes, in a popular Scene Graph system, OpenGL Performer. Our experience has shown both ways in which they can be used and also a number of weakness in the current state of both Scene Graph systems and shadow methods. Based on our experiences, we present suggestions for the user desiring to include shadows in their Virtual Environment and highlight areas, where further development would benefit users significantly.
Accelerating and enhancing rendering of realistic ocean scenes
(Václav Skala - UNION Agency, 2007) Darles, E.; Crespin, B.; Ghazanfarpour, D.; Rossignac, Jarek; Skala, Václav
Modeling and rendering realistic ocean scenes has been investigated by many researchers in the past few years, providing different models of ocean waves, finescale details, and lightwater interactions. This paper describes a practical data structure for accelerating raymarching for a wellknown wave model defined as a sum of trochoids waves. This data structure exhibits nice properties that allow to take into account spatial and temporal coherence as well as reducing aliasing effects. In addition to this, it offers a unified framework in which finescale details, such as scattering phenomena due to spray, foam or particles suspended in the water, can be added easily. Finally, we present some benefits of the model, which can handle physicallybased lightwater interactions and glare effects.
Efficient Region-Based Pencil Drawing
(Václav Skala - UNION Agency, 2007) Sun, Shuo; Huang, Dongwei; Rossignac, Jarek; Skala, Václav
This paper proposes an extension to the existing automatic pencil drawing generation technique based on Line Integral Convolution (LIC). The original LIC pencil filter utilizes image segmentation and texture direction detection techniques for defining outlines and stroke directions, and the quality of a resulting image depends largely on the result of the white noises and the texture directions. It may fail to generate a reasonable result when the white noises and the texture directions are not consistent with the texture structure of the input image. To solve this problem, we propose in this paper to improve the existed LIC-based method. First, a more accurate and rapid graphbased image segmentation method is introduced to divide the image into different regions. Second, we present a new region-based way to produce white noises and texture directions. We also demonstrate the enhanced LIC pencil drawing is closer to the real artistic style.
Triangulation of uniform particle systems: its application to the implicit surface texturing
(Václav Skala - UNION Agency, 2007) Levet, Florian; Granier, Xavier; Schlick, Christophe; Rossignac, Jarek; Skala, Václav
Particle systems, as originally presented by Witkin and Heckbert [32], offer an elegant solution to sample implicit surfaces of arbitrary genus, while providing an extremely regular distribution of samples over the surface. In this paper, we present an ef cient technique that uses particle systems to rapidly generate a triangular mesh over an implicit surface, where each triangle is almost equilateral. The major advantage of such a triangulation is that it minimizes the deformations between the mesh and the underlying implicit surface. We exploit this property by using few triangular texture samples mapped in a non-periodic fashion as presented by Neyret and Cani [16]. The result is a pattern-based texturing method that maps homogeneous non-periodic textures to arbitrary implicit surfaces, with almost no deformation.
A Benchmarking Suite for Static Collision Detection Algorithms
(Václav Skala - UNION Agency, 2007) Trenkel, Sven; Weller, Rene; Zachmann, Gabriel; Rossignac, Jarek; Skala, Václav
In this paper, we present a benchmarking suite that allows a systematic comparison of pairwise static collision detection algorithms for rigid objects. The benchmark generates a number of positions and orientations for a predefined distance. We implemented the benchmarking procedure and compared a wide number of freely available collision detection algorithms.
Dimensional Induced Clustering for Surface Recognition
(Václav Skala - UNION Agency, 2007) Porcu, Massimiliano B.; Scateni, Riccardo; Rossignac, Jarek; Skala, Václav
Understanding when a cloud of points in three-dimensional space can be, semantically, interpreted as a surface, and then being able to describe the surface, is an interesting problem in itself and an important task to tackle in several application elds. Finding a possible solution to the problem implies to answer to many typical questions about surface acquisition and mesh reconstruction: how one can build a metric telling whether a point in space belongs to the surface? Given data from 3D scanning devices, how can we tell apart (and eventually discard) points representing noise from signal? Can the reached insight be used to align point clouds coming from di erent acquisitions? Inside this framework, the present paper investigates the features of a new dimensional clustering algorithm. Unless standard clustering methods, the peculiarity of this algorithm is, using the local fractal dimension, to select subsets of lower dimensionality inside the global of dimension N. When applied to the study of discrete surfaces embedded in three dimensional space, the algorithm results to be robust and able to discriminate the surface as a subset of fractal dimension two, differentiating it from the background, even in the presence of an intense noise. The preliminary tests we performed, on points clouds generated from known surfaces, show that the recognition error is lower than 3 percent and does not a ffect the visual quality of the final result.
Viewpoint Entropy-Driven Simplification
(Václav Skala - UNION Agency, 2007) Castelló, P.; Sbert, M.; Chover, M.; Feixas, M.; Rossignac, Jarek; Skala, Václav
In this paper, a new viewpoint-based simplification approach is proposed for polygonal meshes. This approach is driven by an information-theoretic metric, viewpoint entropy, which measures the amount of information from a scene or object that arrives at a certain viewpoint. Our algorithm applies the best half-edge collapse as a decimation criterion and uses the variation of viewpoint entropy to measure the edge collapse error. Compared to pure geometric-based simplifications, the models produced by our method are closer to the original model according to visual similarity. Our approach also achieves a higher simplification in hidden interiors, by being able to remove them all and to leave the visible areas of the mesh intact. Models generated by CAD applications can benefit from this feature, since these models are usually constructed by assembling smaller objects which can become partially hidden during joining operations. The main application of our method is for video games where models come from CAD applications and are geometrically not very complex, a few thousand polygons at the most, and in which visual similarity is the most important requirement.
Assembly Simulation on Collaborative Haptic Virtual Environments
(Václav Skala - UNION Agency, 2007) Iglesias, R.; Prada, E.; Uribe, A.; Garcia-Alonso, A.; Casado, S.; Gutierrez, T.; Rossignac, Jarek; Skala, Václav
Currently, Virtual Environments (VEs) are used within engineering industry: physical prototypes or mock-ups are replaced by virtual prototypes. Increasingly, these VEs also allow designers and engineers to carry out assembly/disassembly processes and assembly sequences before any physical prototype is built. Moreover, different designers or engineers, which may be situated in the same or in geographically dispersed locations, often collaborate in the design of products. This allows developing more complex products within a shorter time scale and lowered costs. On the other hand, the utilization of haptic feedback has been found to significantly enhance task performance, for instance, in assembly tasks. In this paper we describe an assembly simulation application on a collaborative haptic virtual environment, where several users interact with virtual models to perform assembly operations within the same virtual scene. The paper also summarizes results achieved with experiments which evaluated different collaborative architectures. Furthermore, it reports on the goals that can be achieved and the limitations for haptic collaborative interaction in each case.
Tweakable Shadows for Cartoon Animation
(Václav Skala - UNION Agency, 2007) Nakajima, Hidehito; Sugisaki, Eiji; Morishima, Shigeo; Rossignac, Jarek; Skala, Václav
The role of shadow is important in cartoon animation. Shadows in hand-drawn animation reflect the expression of the animators' style, rather than mere physical phenomena. However shadows in 3DCG cannot express such an animators' touch. In this paper we propose a novel method for editing the shadow with both advantages of hand drawn animation and 3DCG technology. In particular, we discuss two phases that enable animators to transform and deform the shadow tweakably. The first phase is that a shadow projection matrix is applied to deform the shape of the shadow. The second one is that we manipulate vectors to transform the shadow such as scaling and translation. These vectors are used in shadow volume method. The shadows are edited directably by integration of these two phases. Our approach enables animators to edit the shadow by simple mouse operations. As a result, the animators can not only produce shadows automatically but also reflect their style easily. Once the shape and location of shadow are decided by animators' style in our method, 3DCG techniques can produce consistent shadow in object motion interactively.
Fast Skeleton Estimation from Motion Capture Data using Generalized Delogne-Kåsa method
(Václav Skala - UNION Agency, 2007) Knight, Jonathan Kipling; Semwal, S. K.; Rossignac, Jarek; Skala, Václav
This paper presents a fast closed-form solution estimating the rotation points of joints relative to the motion capture data. The proposed solution estimates the physical location of joints inside the body of the person wearing the trackers. The Generalized Delogne-Kåsa method used for our implementation fits spheres, cylinders, circles and planes to the motion capture data without the need for initial guessing. This non-iterative, closed-form solution is fast as it calculates the rotation point with O(N) averaging along with one inversion of a 3x3 positive semi-definite matrix for each joint. The error in the joint location is on average 3σ/√N which is low. In addition, sample points for every joint can be from different time sequence allowing flexibility in recovering the joint locations. Once the joint location relative to the tracker position is determined, it could be used for the remainder of the data set. Publicly available CMU motion capture data was used for this study. Two animation sequences, showing our method, are included with this paper. These results can be compared to that available at the CMU site for the same animation. Since the pose is found relative to the given data, our pose estimation provide better fit to the given data, revealing subtle, individual nuances of the person used for the motion capture. Because of the closed form solution, our technique is ideally suited for the use of motion captured data to create skeletal motion in 3D games or applications where real time performance is essential.
Inverse Modeling and Animation of Growing Single-stemmed Trees at Interactive Rates
(Václav Skala - UNION Agency, 2007) Rudnick, Steffen; Linsen, Lars; McPherson, E. Gregory; Rossignac, Jarek; Skala, Václav
For city planning purposes, animations of growing trees of several species can be used to deduce which species may best fit a particular environment. The models used for the animation must conform to real measured data. We present an approach for inverse modeling to fit global growth parameters. The model comprises local production rules, which are iteratively and simultaneously applied to build a fractal branching structure, and incorporates the propensity of trees to grow towards light. The parameters of the local production rules are derived from global functions that describe the measured tree growth data over time. The production rules are influenced by the global light distribution, which is represented by the amount of light available at each position within the tree’s crown. Since we want to allow the user to explore the tree’s appearance interactively at any time during the animation, all modeling computations must be within a time frame that allows for interactive rendering rates. To this end, we developed a fast approximate algorithm for computing the light distribution. The rendering itself must also be fast; therefore, we sought a well-balanced compromise between photo-realism and performance. Because shadow computations play a key role for photo-realism, we developed a fast approximate shadow computation algorithm including soft shadows and self-shadowing. We applied our methods in order to model and animate the growth of seven single-stemmed tree species in an interactive setting.
Memory Efficient and Robust Software Implementation of the Raycast Algorithm
(Václav Skala - UNION Agency, 2007) de Pina, Aline A.; Bentes, Cristiana; Farias, Ricardo; Rossignac, Jarek; Skala, Václav
In this paper we propose two novel software implementations of the ray-casting volume rendering algorithm for irregular grids, called ME-Raycast (Memory Efficient Ray-casting) and EME-Raycast (Enhanced Memory Efficient Ray-Casting). Our algorithms improve previous work by Bunyk et al [1] in terms of complete handling of degenerate cases, memory consumption, and type of cell allowed in the grid (tetrahedral and/or hexahedral). The use of a more compact and non-redundant data structure, allowed us to achieve higher memory efficiency. Our results show consistent and significant gains in the memory usage of ME-Raycast and EME-Raycast when compared to Bunyk et al implementation. Furthermore, our results also show that handling of degenerate cases generates accurate images, correctly rendering all the pixels in the image, while Bunyk et al implementation fails in rendering up to 38 pixels in the final image. When we compare our algorithms to other robust rendering algorithm, like ZSweep [2], we have considerable performance gains and competitive memory consumption. We conclude that ME-Raycast and EME-Raycast are efficient methods for ray-casting that allow in-core rendering of large datasets with no image errors.
Seamless Patches for GPU-Based Terrain Rendering
(Václav Skala - UNION Agency, 2007) Livny, Yotam; Kogan, Zvi; El-Sana, Jihad; Rossignac, Jarek; Skala, Václav
In this paper we present a novel approach for interactive rendering of large terrain datasets which is based on subdividing the terrain into rectangular patches at different resolutions. Each patch is represented by four triangular tiles which can be at different resolutions; and four strips which are used to stitch the four tiles in a seamless manner. As a result, our scheme maintains resolution changes within patches and not across patches. At runtime, the terrain patches are used to construct a level of detail based on view-parameters. The selected level of detail only includes the layout of the patches and the resolutions at boundary edges. Since adjacent patches agree on the resolution of common edges, the resulted mesh does not include any cracks or degenerate triangles. The GPU generates the meshes of the patches by using scaled instances of cached tiles and assigning elevation for each vertex from the cached textures. Our algorithm manages to achieve quality images at high frame rates while providing seamless transition between different levels of detail.
Octree-based view-dependent triangle meshes
(Václav Skala - UNION Agency, 2007) Fairén, Marta; Trueba, Ramón; Rossignac, Jarek; Skala, Václav
In this paper we present a new technique for view-dependent LOD rendering, where the scene is represented through an octree model from which we can obtain a triangle mesh corresponding to a view-dependent LOD. We present the construction of this octree model and the visualization algorithm that generates on-the-fly a closed and valid triangle mesh for each frame of the visualization. This visualization algorithm is a depth-first traversal algorithm which also allows to re-use triangles from one frame to another.
Upper and Lower Bounds on the Quality of the PCA Bounding Boxes
(Václav Skala - UNION Agency, 2007) Dimitrov, Danko; Knauer, Christian; Kriegel, Klaus; Rote, Günter; Rossignac, Jarek; Skala, Václav
Principle component analysis (PCA) is commonly used to compute a bounding box of a point set in Rd. The popularity of this heuristic lies in its speed, easy implementation and in the fact that usually, PCA bounding boxes quite well approximate the minimum-volume bounding boxes. In this paper we give a lower bound on the approximation factor of PCA bounding boxes of convex polytopes in arbitrary dimension, and an upper bound on the approximation factor of PCA bounding boxes of convex polygons in R2.
Hardware-Independent Clipmapping
(Václav Skala - UNION Agency, 2007) Seoane, Antonio; Taibo, Javier; Hernández, Luiz; López, Ruben; Jaspe, Alberto; Rossignac, Jarek; Skala, Václav
We present a technique for efficient management of large textures and its real-time application to geometric models. The proposed technique is inspired by the clipmap [12] idea, that caches in video memory a subset of the texture mipmap pyramid. Based on this concept, we define some structures and a different management allowing its implementation on a personal computer without specific graphics hardware. Finally, we present the results of the application in a terrain visualization system, using several simultaneous textures with a detail up to 0.25 meters per texel, covering a 60,000 km2 area.
Automated 3D scan multi-view registration based on rotation estimation
(Václav Skala - UNION Agency, 2007) Wang, Huaxin; Vergeest, Joris S. M.; Song, Yu; Wiegers, Tjamme; Rossignac, Jarek; Skala, Václav
A new approach is presented for automatic registration of a sequence of 3D scanned range images into a complete model. Provided that the user follows the proposed scanning strategy, the difficulty to register the split views into a complete model is alleviated and success rate of automatic registration is improved by an adaptive forward estimation based on the extracted rotation axes and angles of the object between each scanning steps. The basic idea of this improvement is that when each of the scan differs slightly from its previous one in a small rotation about an approximately constant axis, the rotation axes and angle in the previous scanning step can be used as the first estimation of the rotation in the next step, so as to avoid manual coarse initial alignment. Formulation of the principle is provided and experimental results demonstrated.

Browse

Recent Submissions