Volume 21, Number 1 (2013)

Permanent URI for this collection

http://hdl.handle.net/11025/6864

Browse

Showing 1 - 10 out of 10 results

LightCluster - Clustering Lights to Accelerate Shadow Computation
(Václav Skala - Union Agency, 2013) Wiesenhütter, Daniel; Klein, Andreas; Nischwitz, Alfred; Skala, Václav
In this paper, we propose a method to reduce the amount of shadow maps required for rendering shadows in scenes with many lights. Our idea is to use the spatial relationship of lights to find clusters and replace the lights of a cluster with a single area light. We use a soft shadow algorithm for area lights to approximate the shadows for the clusters. By carefully placing the cluster centers, we can minimize the errors in the shadows. While the clustering only adds a small overhead in the worst case, it can efficiently reduce the number of shadow maps. Thus, in many cases the resulting error in shadows is acceptable compared to the increase in rendering performance.
2.5D Clip-Surfaces for Technical Visualization
(Václav Skala - Union Agency, 2013) Trapp, Matthias; Döllner, Jürgen; Skala, Václav
The concept of clipping planes is well known in computer graphics and can be used to create cut-away views. But clipping against just analytical defined planes is not always suitable for communicating every aspect of such visualization. For example, in hand-drawn technical illustrations, artists tend to communicate the difference between a cut and a model feature by using non-regular, sketchy cut lines instead of straight ones. To enable this functionality in computer graphics, this paper presents a technique for applying 2.5D clip-surfaces in real-time. Therefore, the clip plane equation is extended with an additional offset map, which can be represented by a texture map that contains height values. Clipping is then performed by varying the clip plane equation with respect to such an offset map. Further, a capping technique is proposed that enables the rendering of caps onto the clipped area to convey the impression of solid material. It avoids a re-meshing of a solid polygonal mesh after clipping is performed. Our approach is pixel precise, applicable in real-time, and takes fully advantage of graphics accelerators.
Novel Patterns and Methods for Zooming Camera Calibration
(Václav Skala - Union Agency, 2013) Pennisi, Andrea; Bloisi, Dominico; Gaz, Claudio; Iocchi, Luca; Nardi, Daniele; Skala, Václav
Camera calibration is a necessary step in order to develop applications that need to establish a relationship between image pixels and real world points. The goal of camera calibration is to estimate the extrinsic and intrinsic camera parameters. Usually, for non-zooming cameras, the calibration is carried out by using a grid pattern of known dimensions (e.g., a chessboard). However, for cameras with zoom functions, the use of a grid pattern only is not sufficient, because the calibration has to be effective at multiple zoom levels and some features (e.g., corners) could not be detectable. In this paper, a calibration method based on two novel calibration patterns, specifically designed for zooming cameras, is presented. The first pattern, called in-lab pattern, is designed for intrinsic parameter recovery, while the second one, called on-field pattern, is conceived for extrinsic parameter estimation. As an application example, on-line virtual advertising in sport events, where the objective is to insert virtual advertising images into live or pre-recorded television shows, is considered. A quantitative experimental evaluation shows an increase of performance with respect to the use of standard calibration routines considering both re-projection accuracy and calibration time.
Trixel Buffer Logic for I/O Bound Point in N-Polygon Inclusion Tests of Massive Bathymetric Data
(Václav Skala - Union Agency, 2013) Oliveira, João Fradinho; Ziebart, Marek; Iliffe, Jonathan; Turner, James; Robson, Stuart; Skala, Václav
“Trixel Buffers is a new spatial data-structure for fast point in multiple polygon inclusion queries. The algorithm utilizes a pre-processing step in which the inside/outside status of a quadtree´s leaf triangles without polygon geometry is pre-computed automatically; at run-time point queries lying within these triangles simply inherit their inclusion status. If a point query lies in a leaf triangle enclosing polygon vertices or crossing edges, a ray is fired from the point towards the triangle center whose polygon inclusion properties has also been pre-computed: rules are then applied to the intersection count and center-point properties to infer the polygon inclusion status. Our main contribution is that rays need not be followed until the polygon limits, and consequently the algorithm is I/O bound with shallow trees. It took 1h36m rather than days of using a standard ray test to determine the multiple polygon (~270,000 line segments) inclusion of 1.75 billion points on a 2.5GHz DuoCore computer.
Frequency-based Progressive Rendering of Continuous Scatterplots
(Václav Skala - Union Agency, 2013) Molchanov, Vladimir; Fofonov, Alexey; Linsen, Lars; Skala, Václav
Continuous scatterplots are a consistent tool for the visual representation and exploration of continuous multivariate data defined on a continuous domain. Due to the complexity of the construction algorithm, application of continuous scatterplots is limited in terms of data size and screen resolution when interactive frame rates are desired. Progressive rendering is a paradigm of displaying an approximative visual outcome early on, which iteratively and incrementally gets improved until convergence to the final result is reached. This approach maintains the interactivity of the system and allows the user to make decisions immediately, i.e., much earlier than the end of the computation process. We propose a method for progressive rendering of continuous scatterplots based on a Fourier representation. By iteratively advancing from low to high frequencies and inverting the spectrum representation after each iteration, a series of scatterplots converging to the final result is generated and rendered. We demonstrate that this convergence is monotonic and that the proposed approach is more efficient than state-of-the-art methods, i.e., we can faster produce high-quality approximations. We propose to embed this idea into a hybrid approach which allows balancing the trade-off between quality of the image appearing first and its computation time. The proposed algorithms were implemented on the GPU.
Rigid registration of different poses of animated shapes
(Václav Skala - Union Agency, 2013) Livesu, Marco; Scatenni, Ricardo; Skala, Václav
Different poses of 3D models are very often given in different positions and orientations in space. Since most of the computer graphics algorithms do not satisfy geometric invariance, it is very important to bring shapes into a canonical coordinate frame before any processing. In this paper we consider the problem of finding the best alignment between two or more different poses of the same object represented by triangle meshes sharing the same connectivity. Firstly, we developed a method to select a region of interest (ROI) which has a perfect alignment over the two poses (up to a rigid movement). Secondary, we solved a simplified version of the Largest Common Point-set (LCP) problem with a-priori knowledge about point correspondence, in order to align the ROIs. We eventually align the poses performing least square rigid registration. Our method makes no assumption about the starting positions of the objects and can also be used with more than two poses at once. It is fast, non-iterative, easy to reproduce and brings the poses into the best alignment whatever the initial positions are.
Optimal RANSAC - Towards a Repeatable Algorithm for Finding the Optimal Set
(Václav Skala - Union Agency, 2013) Hast, Anders; Nysjö, Johan; Marchetti, Andrea; Skala, Václav
A novel idea on how to make RANSAC repeatable is presented, which will find the optimal set in nearly every run for certain types of applications. The proposed algorithm can be used for such transformations that can be constructed by more than the minimal points required. We give examples on matching of aerial images using the Direct Linear Transformation, which requires at least four points. Moreover, we give examples on how the algorithm can be used for finding a plane in 3D using three points or more. Due to its random nature, standard RANSAC is not always able to find the optimal set even for moderately contaminated sets and it usually performs badly when the number of inliers is less than 50%. However, our algorithm is capable of finding the optimal set for heavily contaminated sets, even for an inlier ratio under 5%. The proposed algorithm is based on several known methods, which we modify in a unique way and together they produce a result that is quite different from what each method can produce on its own.
Using holistic features for scene classification by combining classifiers
(Václav Skala - Union Agency, 2013) de Souza Gazolli, Kelly Assis; Salles, Evandro Ottoni Teattini; Skala, Václav
Scene classification is a useful, yet challenging problem in computer vision. Two important tasks for scene classification are the image representation and the choice of the classifier used for decision making. This paper proposes a new technique for scene classification using combined classifiers method. We run two classifiers based on different features: GistCMCT and spatial MCT and combine the output results to obtain the final class. In this way, we improve accuracy, by taking advantage from the qualities of the two descriptors, without increasing the final size of the feature vector. Experimental results on four used datasets demonstrate that the proposed methods could achieve competitive performance against previous methods.
Constrained Up-Scaling for Direct and Global Image Components
(Václav Skala - Union Agency, 2013) Bader, Julian; Pätzold, Martin; Kolb, Andreas; Skala, Václav
The separation of direct and global illumination components is interesting for many applications in Computer Graphics and Computer Vision, such as BRDF estimation or material classification. However, for full-resolution images, a large number of coded images have to be acquired. For many interactive applications, such as the acquisition of dynamic scenes or video capturing, this is not feasible. In this paper, a new constrained up-scaling technique for separated direct and global illumination images is proposed which requires two to three coded input images, only. Our approach imposes the boundary condition that the sum of the direct and global components equals the fully illuminated image. We work in a predictive-corrective manner where we first use a single-image up-scaling method in order to predict the higher resolution images. Afterwards, the missing higher frequencies are determined using a fully illuminated image. As the distribution of the higher frequencies differs among the various frequency bands, we apply our approach in an iterative way for small up-scaling steps distributing the missing information by minimizing the overall frequencies. We evaluate the up-scaling scheme and demonstrate the improvement compared to single-image approached. As our method aims at minimizing the structured light patterns needed for acquisition, we additionally discuss the performance of existing pattern sets in terms of applicability for dynamic scenes.
On the Parameterization and the Geometry of the Configuration Space of a Single Planar Robot
(Václav Skala - Union Agency, 2013) Atariah, Dror; Ghosh, Sunayana; Rote, Günter; Skala, Václav
Translating and rotating planar polygonal robots are studied in the literature for decades. An integral part of this study is the configuration space which corresponds to the work space. In the context of motion planning problems, the boundary between the free and forbidden parts of the configuration space plays a major role. In this paper we find an explicit parameterization of the boundary of the forbidden space. Using this parameterization we detail several geometrical properties of the various elements which constitute this boundary. In addition, this parameterization enables us to visualize these elements.

Browse

Recent Submissions