Volume 10, number 1-2 (2002)
Permanent URI for this collection
Browse
Recent Submissions
Item Interative 3D geometric modelers with 2D UI(UNION Agency, 2002) Wu, Shin-Ting; Malheiros, Marcelo de Gomensoro; Skala, VáclavThis paper presents an object-oriented framework which enhances the collaboration of three categories of experts that play fundamental role in the development of a software for interactive 3D geometric modelers. It aims at three purposes. First, it supports application developers to build a graphics interface for manipulating with 2D devices their own 3D data representations, without intimate knowledge of its internal structure. Second, it provides facilities for interface researchers to create and experiment 3D widgets from reusable draggers and 2D-3D mapping strategies. Finally, it permits graphics experts to implement sophisticated draggers and complex 2D-3D mapping strategies by overriding operations of the predefined abstract classes.
Item Fuzzy linking models for pyramidal edge detection(UNION Agency, 2002) Wang, Zhi-Gang; Wang, Wei; Xu, Xiao-Ming; Skala, VáclavThis paper presents a novel approach to multiresolution edge detection, which combines the grayscale morphological filtering, pyramid data structure and fuzzy technique. It mainly addresses the linking of edge nodes at adjacent levels in image pyramid. In previous pyramidal approaches, linking is based on linear relationship and intensity proximity only. The approach proposed here contains multiple linking mechanisms and introduces fuzzy technique. It considers the parent-child linking relationship of edge nodes between the two adjacent levels as fuzzy model, which is trained offline using real image data. Through this fuzzy linking model, the coarse, low-resolution edge map is propagated and refined to the fine, high-resolution edge map in the pyramid. The validation experiment is carried out on one synthetic image and two real images, and the results show that our approach has better performance on the localization and detection of continuous large-scale object boundaries than Canny’s edge detector and other previous multiresolution approaches. In addition, the proposed approach has high computational efficiency.Item Synthesis of normal and abnormal human gait animation(UNION Agency, 2002) Jing, Cui; Prakash, Cyril; Skala, VáclavThe mechanics of gait animation for human normal walking and abnormal walking has been investigated using a novel hybrid approach with kinematic determinants and biomedical constraints. The kinematic determinants help to achieve efficient control of gait synthesis whereas the masked constraints help in the specification of biomedical correct postures. In the proposed new masked constraints approach, a mask is added to each constraint cone to indicate particular subsets of abnormal postures. In summary, our gait system is an attempt to adapt biomedical analysis of human walking for synthesis of gait in human computer animation.Item Shape invariantsand principal directions from 3D points and normals(UNION Agency, 2002) Kamberov, George; Kamberova, Gerda; Skala, VáclavA new technique for computing the differential invariants of a surface from 3D sample points and normals is presented. It is based on a new conformal geometric approach to computing shape invariants directly from the Gauss map. In the current implementation we compute the mean curvature, the Gauss curvature, and the principal curvature axes at 3D points reconstructed by area-based stereo. The differential invariants are computed directly from the points and the normals without prior recovery of a 3D surface model and an approximate surface parameterization. The technique is stable computationally.Item A study about the form factors kernel functions(UNION Agency, 2002) Zéghers, Eric; Renaud, Christophe; Skala, VáclavThis paper presents a study of the form factors (FF) function kernel. The accuracy of FF estimate is known as a difficult problem when simulating radiative energy exchanges between objects inside an enclosure. By carefully studying the FF function between two polygons we are able to propose a very interesting characterization of its behaviour according to the relative distance between those polygons (general form of the function, location and height of its unique maximum, effect of polygons orientation and distance, ...). According to the results of this study we estimate the FFs between any two polygons by distinguishing the areas where the kernel has smooth variations from the those where it changes quickly. A fine integration is thus performed for the more varying parts of the kernel whereas the other parts are computed more easily. We show that even a very simple implementation of our approach provides accurate estimates of the FF close to the results provided by the Schroeder formula in a time 8 up to 10 times faster. Moreover our approach does not suffer from lack of accuracy when surfaces are very closed from each other thus outperforming classical methods.Item Arbitrary viewpoint video synthesis from uncalibrated multiple cameras(UNION Agency, 2002) Yaguchi, Satoshi; Saito, Hideo; Skala, VáclavIn this paper, we propose a method for arbitrary view synthesis from uncalibrated multiple camera system, targeting large space such as soccer stadium. In Projective Grid Space (PGS), that is the three-dimensional space defined by epipolar geometry between the two basis cameras in the camera system, we reconstruct three dimensional shape models from the silhouette images. By using the three dimensional shape model reconstructed in PGS, we can obtain the dense point correspondence map between reference images. The obtained correspondence can synthesize the image of arbitrary view between the reference images. We also propose the method for merging the synthesized images with the virtual background scene in PGS. We apply the proposed method to image sequences taken by the multiple camera system, which is developed in a large space on a concert hall. The synthesized image sequences of virtual camera have enough quality to demonstrate effectiveness of the proposed method.Item Image rotation without scaling on spiral architecture(UNION Agency, 2002) Wu, Qiang; Xianghjian, He; Hintz, Tom; Skala, VáclavSpiral Architecture is a relatively new and powerful approach to general purpose machine vision system. On this novel architecture, image rotation is achieved by Spiral Multiplication. However, the general image rotation on Spiral Architecture has two effects. One is scaling segmentation and the other is rotation. This paper presents an algorithm to achieve image rotation without scaling on Spiral Architecture, which improves the Spiral Architecture’s usage in image processing.Item Curvature measures of 3D vector fields and their applications(UNION Agency, 2002) Weinkauf, T.; Theisel, H.; Skala, VáclavTangent curves are a powerful tool for analyzing and visualizing vector fields. In this paper two of their most important properties are examined: their curvature and torsion. Furthermore, the concept of normal surfaces is introduced to the theory of 3D vector fields, and their Gaussian and mean curvature are analyzed. It is shown that those four curvature measures tend to infinity near critical points of a 3D vector field. Applications utilizing this behaviour for the (topological) treatment of critical points are discussed.Item The synthesis of trees in Chinese landscape painting using silhoutte and texture strokes(UNION Agency, 2002) Way, De-Lor; Lin, Yu-Ru; Shih, Zen-Chung; Skala, VáclavPracticed for more than three thousand years, Chinese painting emphasizes "implicit meaning", and involves painters’ using a minimal number of brush strokes to express their deepest feelings. Landscapes are one of the most important themes in Chinese painting. Trees are the essential painting objects. This paper presents a set of novel methods to automatically draw trees in Chinese ink painting from 3D polygonal models. Outline rendering and texture generation uses the information of the silhouette, shade and orientation of three-dimensional model’s surface to draw a particular tree. Four reference maps are established to analyze the information for the bark texture. These methods can draw various styles of bark texture by defining the texture patterns. Finally, this paper demonstrates some results obtained with our method.Item EndoView: a phantom study of a tracked virtual bronchoscopy(UNION Agency, 2002) Wagner, Daniel; Wegenkittl, Rainer; Gröller, Eduard; Skala, VáclavVirtual endoscopy can be used for preoperative planning, for training and intraoperatively. Surface rendering displays the inner lumen very well. Volume rendering has to be used if the external structures are of interest. For certain applications, e.g. endoluminal biopsy, it is of great advantage to be able to use both techniques at once. In this work we describe an approach that allows using these two methods in combination on a low-end standard personal computer. Since image generation is done in a preprocessing step, any high quality volume or polygonal rendering technique can be used and mixed together without any loss in performance at run-time. This work extends a previous image based rendering system for virtual bronchoscopy to include tracking of a rigid or flexible endoscope and finding one's way in the tracheal tree by displaying the endoscope's position in a top-view map of the trachea. Natural landmarks, i.e. bifurcations in the bronchial tree, are used for registration. Properties of the technique are explored on a phantom data set.Item Human behaviour visualization and simulation for automatic video understanding(UNION Agency, 2002) Vu, Van Thinh; Brémond, François; Thonnat, Monique; Skala, VáclavItem Line scratch on digital images: an energy based model(UNION Agency, 2002) Vitulano, D.; Bruni, V.; Ciarlini, P.; Skala, VáclavIn this paper, a model for automatic detection of line scratches on digital images is proposed. It consists in a generalization of the Kokaram's model using width and height features of the line profile along with Weber's law, to test scratches visibility on the image. Experimental results show that the algorithm works better than the existing models in detecting true scratches with a lower computing time.Item Direct 3D pattern matching(UNION Agency, 2002) Vergeest, Yoris S. M.; Spanjaard, Sander; Song, Yu; Skala, VáclavWe have studied the problem of extracting shape parameters from freeform features in full 3D. The freeform features are typically encountered in laser-scanned data from physical parts or in point clouds from any source. Matching involves 1) the search for optimal position and orientation of the template shape and 2) the adjustment of a number (d) of intrinsic shape parameters. The type and total number of parameters (6+d), and hence the family of template shapes is determined by the type of feature. We have analyzed the robustness, accuracy and the efficiency of Hausdorff-like shape distance measures. A number of search strategies have been tested and were evaluated against convergence and computational performance. The practical relevance of the technique is addressed as well.Item Automatic keyframe selection for high-quality image-based wlkthrough animation using viewpoint entropy(UNION Agency, 2002) Vázquez, Pere-Pau; Sbert, Mateu; Skala, VáclavThe computation of high quality animation sequences is expensive. Generation time for each frame can take a few hours. Recently, Image-Based Rendering methods have been proposed to solve this problem. As these techniques obtain new arbitrary views from precomputed ones at low cost, walkthroughs may be computed faster. Consequently, the selection of the precomputed images is a very important step. The initial set of views should fulfill two requirements, it must be small but provide as much information as possible on the scene. In this paper we review several keyframe selection strategies and then we propose a new method based on entropy that achieve similar, and in some cases better, results.Item Surface-based cloud visualization for animation applications(UNION Agency, 2002) Trembilski, Andrzej; Broßler, Andreas; Skala, VáclavFor the local TV presentation of weather forecast data it is important to have high-quality and fast visualisation of clouds. In this paper we present surface-based methods for the high performance visualisation of clouds from data produced by a routine meteorological weather simulation. Isosurfaces, which are originally too coarse because of the data grid resolution are refined and deformed. The refined geometry is used for a light simulation and transparency computation.Item Real-time view morphing for web applications(UNION Agency, 2002) Terasawa, Mikio; Yamaguchi, Yasushi; Odaka, Kinji; Skala, VáclavIn this paper, real-time view morphing that is an extension of view morphing is proposed for web applications such as on-line shopping and remote instruction. Real-time view morphing is an image-based rendering method that generates an image of an intermediate view from two or more photographs in real-time without 3D models. The method is combined with conventional view morphing and 2D polygon rendering with texture mapping for real-time processing. Techniques to avoid discontinuity of texture and artifacts at contours without generating holes are proposed to keep the quality of original images. The advantages of the method are small data size, high image quality and real-time rendering that are important for web applications. Users can change object-centered viewpoints interactively in a web browser on a local machine that is connected to the narrow band Internet. The real-time view morphing program also runs in banner advertisements and desktop accessories. An editor is developed for preparing the data such as corresponding points, faces, and face order manually or semi-automatically using the epipolar constraint.Item Linear-time CSG rendering of intersected convex objects(UNION Agency, 2002) Stewart, Nigel; Leach, Geoff; Sabu, John; Skala, VáclavThe Sequenced Convex Subtraction(SCS) algorithm is a hardware based multi-pass image-space lgorithm for general purpose Constructive Solid Geometry (CSG)Rendering. Convex objects combined by volumetric intersection, difference and union are rendered in real-time without b-rep re-processing. OpenGL stencil and depth testing is used to determine the visible surface for each pixel on the screen. This paper introduces a specialised algorithm for CSG Rendering of intersected convex objects,we call SCS-Intersect. This new technique requires linear time with respect to the number of intersections. SCS-Intersect is primarily of interest as an optimisation to the SCS algorithm for rendering CSG trees of convex objects. A revised formulation of the SCS CSG Rendering algorithm is presented in this pape.Item Monitoring and correction of geometric distortion in projected displays(UNION Agency, 2002) Steele, R. Matt; Webb, Stephen; Jaynes, Christopher; Skala, VáclavOff-axis placement of light projectors induces significant planar parallax on the display surface. Although commodity solutions exist for removing this distortion, they involve iterative, menu-driven user interaction or physical alignment of the projector, and in either case interrupt the use of the display. User interaction is infeasible in a number of scenarios including mechanically aligned multi-projector displays that are subject to mechanical drift and situations in which projectors are often reconfigured. We present a general technique for continuous rectification of arbitrary off-axis distortions that does not require user interaction. A camera automatically detects when the projector's orientation has changed, without requiring explicit fiducials or targets in the world. The method runs in concert with interactive display applications and has minimal impact on framerate. An initial rectifying transform is recovered automatically by projecting target points and observing them in the camera. The display is then warped and passively monitored for calibration error and motion of the projector. The technique distinguishes between distortions due to miscalibration and intentional framebuffer changes. A consistency score is measured by generating a predicted view based on the current framebuffer contents and correlating this prediction with the camera's captured image. Poor correlation scores indicate that the projector has moved and re-calibration and geometric correction is required. Initial experiments show that the calibration consistency measures are sufficiently robust to distinguish small motion of the projector from continuously changing imagery.Item PGF: a new progressive file format for lossy and lossless image compression(UNION Agency, 2002) Stamm, Christoph; Skala, VáclavWe present a new image file format, called Progressive Graphics File (PGF), which is based on a discrete wavelet transform with progressive coding features. We show all steps of a transform based coder in detail and discuss some important aspects of our careful implementation. PGF can be used for lossless and lossy compression. It performs best for natural images and aerial ortho-photos. For these types of images it shows in its lossy compression mode a better compression efficiency than JPEG. This efficiency gain is almost for free, because the encoding and decoding times are only marginally longer. We also compare PGF with JPEG 2000 and show that JPEG 2000 is about ten times slower than PGF. In its lossless compression mode PGF has a slightly worse compression efficiency than JPEG 2000, but a clearly better compression efficiency than JPEG-LS and PNG. If both, compression efficiency and runtime, is important, then PGF is the best of the tested algorithms for compression of natural images and aerial photos.Item Human pose estimation from silhouettes: a consistent approach using distance level sets(UNION Agency, 2002) Sminchisescu, C.; Telea, A.; Skala, VáclavWe present a novel similarity measure (likelihood) for estimating three-dimensional human pose from image silhouettes in model-based vision applications. One of the challenges in such approaches is the construction of a model-to-image likelihood that truly reflects the good configurations of the problem. This is hard, commonly due to the violation of consistency principle resulting in the introduction of spurious, unrelated peaks/minima that make the search for model localization difficult. We introduce an entirely continuous formulation which enforces model estimation consistency by means of an attraction/explanation silhouette-based term pair. We subsequently show how the proposed method provides significant consolidation and improved attraction zone around the desired likelihood configurations and elimination of some of the spurious ones. Finally, we present a skeleton-based smoothing method for the image silhouettes that stabilizes and accelerates the search process.
