Posters

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/238

Browse

Now showing 1 - 20 of 27

Curve Synthesis from Learned Refinement Models
(Eurographics Association, 2003) Simhon, Saul; Dudek, Gregory
We present a method for generating refined 2D illustrations from hand drawn outlines consisting of only curve strokes. The system controllably synthesizes novel illustrations by augmenting the hand drawn curves’ shape, thickness, color and surrounding texture. These refinements are learned from a training ensemble. Users can select several examples that depict the desired idealized look and train the system for that type of refinement. Further, given several types of refinements, our system automatically recognizes the curve and applies the appropriate refinement. Recognition is accomplished by evaluating the likelihood the curve belongs to a particular class based on both its shape and context in the illustration.
Multi-resolution modelling of terrains by using non restricted quadtree triangulation
(Eurographics Association, 2003) Aguilera, A.; Torres, J. C.; Feito, F.
The interactive visualization of terrains requires the processing of a high amount of data in real time, usually it is not possible to work with all of them on main in memory. In this work we present, a method which solves the above problem by means of using spatial indices (quadtree). For that, the terrains are divided into square cells, each of them having a quadtree, which will only be refined up to a the detail level depending on the observer position. With this, allaus to reduce the time used for terrains visualization.
Fast Marching farthest point sampling
(Eurographics Association, 2003) Moenning, Carsten; Dodgson, Neil A.
We introduce the Fast Marching farthest point sampling (FastFPS) approach for the progressive sampling of planar domains and curved manifolds in triangulated, point cloud or implicit form. By using Fast Marching methods2, 3, 6 for the incremental computation of distance maps across the sampling domain, we obtain a farthest point sampling technique superior to earlier point sampling principles in two important respects. Firstly, our method performs equally well in both the uniform and the adaptive case. Secondly, the algorithm is applicable to both images and higher dimensional surfaces in triangulated, point cloud or implicit form. This paper presents the methods underlying the algorithm and gives examples for the processing of images and triangulated surfaces. A companion report4 provides details regarding the application of the FastFPS algorithm to point clouds and implicit surfaces.
Parametric Foveation for Progressive Texture and Model Transmission
(Eurographics Association, 2003) Cheng, Irene; Basu, Anup; Pan, Yixin
Spatially varying sensing (foveation) has been used in many different areas of Computer Vision, such as image compression and video teleconferencing and in perceptually driven Level of Detail (LOD) representations in graphics. In this work, we show that foveation is advantageous for interactive mesh and texture transmission in online 3D applications. Unlike traditional mesh representations where all 3D vertices need to be transmitted, we only need to transmit a collection of points-of-interest (foveae) and information on only one (rather than three) axis. Thereby, we can achieve a threefold reduction in the amount of data that needs to be transmitted to represent a new 3D model. Our research differs from level of detail (LOD) based approaches using perceptually driven simplification in that (i) the mesh and texture resolutions vary smoothly and continuously in our approach compared to distinct levels of details in adjoining regions in other foveated or multiresolution LOD based methods; and (ii) the approach works for an integrated foveated texture and mesh representation. The current implementation extends our past research in image and video compression [1] and is restricted to regular grid mesh representation produced by 3D scanners.
Virtual Modelling
(Eurographics Association, 2003) Kiss, Szilárd; Nijholt, Anton; Zwiers, Job
We concentrate our efforts on building virtual modelling environments where the content creator uses controls (widgets) as an interactive adjustment modality for the properties of the edited objects. Besides the advantage of being an on-line modelling approach (visualised just like any other on-line VRML content), our approach provides also instant visualisation and an intuitive, graphical editing modality. Although visual modelling environments are more powerful in content creation than a text editor, our aim is to include certain domain knowledge to extend their capabilities. Our most recent system is a 3D character modeller capable of handling H-anim or other types of hierarchies (based also on H-anim components) and the geometry attached to them. The modelling is based on the properties of regular grid geometries, on H-anim hierarchies and where applicable, on 3D character symmetry. The novelty of our approach consists of 1) natural graphics components; 2) integrating the interface elements into the virtual environment itself; 3) rule based modelling of 3D characters; 4) on-line modelling.
Ashli – Advanced Shading Language Interface
(Eurographics Association, 2003) Preetham, Arcot J.; Bleiweiss, Avi
The support for IEEE floating point computation and the exposure of shading functionality in a standardize API form have made graphics hardware a viable workflow solution to an artist involved in digital content creation (DCC). Nevertheless, there still remains a significant interface void between the abstract shading description an artist is used to and the low level shading constructs the hardware expec ts. We have developed Ashli, an advanced shading language interface tool, with the primary motivation to bridge the interface gap fore mentioned. Ashli is developed in the form of a case study. It takes in high level shading languages and descriptions and at the end emits standard graphics hardware shading API (e.g. Microsoft DirectX and OpenGL). Ashli‘s main contribution is in its seamless cooperation with the DCC graphics application. The RenderMan®‡ Shading Language, Maya® Shading Network and 3ds Max® Standard Materials are the subset of input abstraction we have experimented with to validate our approach. Shading path computation complexity and conformance to hardware resource constraints are owned by Ashli and for the most part made transparent to the artist. A tool like Ashli essentially retains rendered image appearance quality comparable to the level produced by a software renderer, but at a significant higher rate of interaction efficiency. We present Ashli in the context of a stand -alone viewing application framework, depicting shading path computation bound to scene rendering. We demonstrate the application design and results for using Ashli as a shading coprocessor assist.
GreenArt: A Tool for Non-Photorealistic Rendering of Plants and Trees
(Eurographics Association, 2003) Campos, C.; Quirós, R.; Huerta, J.; Camahort, E.; Lluch, J.; Vivo, R.
Plant and tree rendering is critical to many computer graphics applications, like interactive walkthroughs, architectural presentations, entertaintment and computer games. Current research focuses on realistic modeling and rendering of plants and trees. In this paper, however, we focus on non-photorealistic rendering of these models. We introduce GreenArt, a tool for a modeling and rendering non-photorealistic plants and trees. We present GreenArt’s functionality and capabilities and we show some examples of scenes generated using the system.
Interactive demo: OpenGLfly, version Euskadi
(Eurographics Association, 2003) Gómez de Segura, José Daniel; Fernández, Borja; Peral, Rosa; López, Susana; Ibáñez, Eduardo
This paper describes the demo associated to one of our projects, which is currently under development. The version presented can be considered as a large terrain simulation over the Basque Country area (North of Spain), with real imagery, georeferenced information and a user friendly and interactive interface. The main goal is to demonstrate the applications of virtual reality techniques to fields such as virtual tourism, low cost simulators or edutainment. The requirements of the demo are low enough for it to be run in a home PC with a middle range graphics accelerator card. Currently, there are some OpenGL optimisations specific to NVIDIA chipsets and consequently this platform has a slight advantage in terms of performance.
Persepolis: Recovering history with a handheld camera
(Eurographics Association, 2003) Sainz, M.; Susin, A.; Cervantes, A.; Bagherzadeh, N.
In this paper we present new improvements to our novel pipeline for image based modeling of objects using a camcorder. Our system takes an uncalibrated sequence of images recorded around a scene, it automatically recovers the underlying 3D structure and camera path and then a volumetric scene reconstruction is performed using a hardware accelerated voxel carving approach. Finally a triangular mesh is obtained and the available information from the images is combined to generate a full 3D photo-realistic reconstruction. As an application, we use this system to reconstruct parts of the archeological site of Persepolis (Iran).
Combining 3D Scans and Motion Capture for Realistic Facial Animation
(Eurographics Association, 2003) Breidt, Martin; Wallraven, Christian; Cunningham, Douglas W.; Buelthoff, Heinrich H.
We present ongoing work on the development of new methods for highly realistic facial animation. One of the main contributions is the use of real-world, high-precision data for both the timing of the animation and the deformation of the face geometry. For animation, a set of morph shapes acquired through a 3D scanner is linearly morphed according to timing extracted from point tracking data recorded with an optical motion capture system.
Large XML Document Manager and Visualizer
(Eurographics Association, 2003) Merienne, J. P.; Jacquemin, C.
This work presents an interactive Graphical User Interface used to display and manipulate collection of large XML documents. It relies on a geometrical model that combines hierarchical and sequential representations. In this interface, XML structures are displayed on an inclined plane in a 3D environment. This interface provides users with a visualization that speeds up searches in collection of documents through simple interaction techniques. Document browsing is facilitated through display space optimization and direct manipulation of document structures. Besides, this approach tries to take advantage of 3D by engaging human 3D spatial cognitive capabilities while using 2D interaction to not disrupt user's habits. A user evaluation for document access is reported that shows that our interface tends to be faster than 2D interface in data searching while taking advantage of the user's natural attraction to 3D interface.
Resampling structured grids into regular ones
(Eurographics Association, 2003) Dimitrov, Leonid I.; Kolcun, Aleksej; Sramek, Milos
Structured grids and some of their applications in natural sciences are discussed. The problem of their visualization and quantitative evaluation is considered and possible ways for its solution sketched. Resampling a structured grid onto a regular one is one of the possible solutions offering the additional benefit of opening up more possibilities for qualification and quantification, too. Some methods for achieving this goal are presented, their advantages and disadvantages compared, and preliminary results from the application of our own method shown.
Pedagogic discrete visualization of electromagnetic waves
(Eurographics Association, 2003) Mora, F.; Ruillet, G.; Andres, E.; Vauzelle, R.
A dynamic electromagnetic wave propagation visualization tool is presented in this paper. It can simulate wave reflection and diffraction in 2D. It allows to visualize the electric field magnitude. The original approach is based on discrete analytical geometry allowing a dynamic visualization of the wave propagation contrary to existing models. We use it for the moment as pedagogic tool for the understanding of this complex phenomenon.
Point in Solid Tests for Triangle Meshes. Comparative Study
(Eurographics Association, 2003) Ogayar, Carlos J.; Segura, Rafael J.; Feito, Francisco R.
This document presents a comparative study of several point inclusion tests for triangle meshes. We also discuss some issues about the usefulness of each method depending on the situation, taking into account memory and CPU usage. The goal is to compare the performance of well known spatial classification algorithms with a modified implementation of Feito-Torres, optimized to work only with triangle meshes. A description of studied methods is also presented, with implementation issues and time tables showing the performance of each algorithm. The tests highlight both weak points and virtues of each approach.
VR Game Interfaces for Interaction onto 3D Display
(Eurographics Association, 2003) Choi, Jeong-Dan; Jang, Byung-Tae; Hwang, Chi-Jeong
This paper describes the VR game interfaces for interaction onto 3D display. The challenge we are interested in is how to build and use a low cost 3D display that satisfies semi-immersive requirements and how to coordinate it with interactions into the immersive VR game environments. So, the objective of our display system is to produce a viewer-oriented, stereoscopic display; each of the user’s eyes should see an image correctly rendered from that eye’s position through the window of the displays. To accomplish this project, we concrete a projection-based multiple window rendering system on passive 3D stereo display, synchronization multichannel system for wide field of view on different PCs, and game user interactions using real-time virtual character animations. Our preliminary results are well suited for VR Safari as game contents with surround displays running on a PC clusters.
3D Model Search Engine Based on Lightfield Descriptors
(Eurographics Association, 2003) Shen, Yu-Te; Chen, Ding-Yun; Tian, Xiao-Pei; Ouhyoung, Ming
With the development of modern 3D modelling and digitizing tools, more and more models have been created recently, which leads to the necessity of the technique of 3D model retrieval system. Our 3D model search engine has been designed with the goal to meet entertaining, industrial, commercial, medical, and educational needs. The system is available on the Web (http://3d.csie.ntu.edu.tw) with a database containing over 10,000 models free downloaded through the Internet. Users can query 3D models by text, drawing 2D shapes using a friendly painting interface, or selecting one 3D model as a query key interactively. The features representing a 3D object, namely, the Lightfield Descriptors, are extracted from 2D images, which are rendered from cameras positioned on the vertices of a regular dodecahedron. The Lightfield Descriptors of each model are used to compare similarities among each other, and the retrieval process takes only 2 and 0.1 seconds with a 3D model and 2D shapes, respectively. During four months, the system has been widely used for querying over three thousand times from 418 IP addresses in at least 27 countries.
Anisotropic Participating Media at Near Interactive Rates
(Eurographics Association, 2003) Jimenez, Juan-Roberto; Pueyo, Xavier
This article presents an algorithm for the realistic simulation, at near interactive rates, of globally illuminated scenes including participating media, isotropic and anisotropic. The proposed algorithm is divided in three phases: particle tracing, reconstruction and rendering. The particle tracing phase is based on the Monte Carlo light tracing. The goal of this phase is to deliver the light energy from the sources, over the scene’s object. In the reconstruction phase the algorithm derives one texture for each object in the scene, 2D textures for surfaces and 3D textures for participating media. Each texel represents the outgoing radiance, in a given direction, from a region of the object. The method uses density estimation techniques to compute the textures. In the final phase, the algorithm renders the textures using hardware graphics capabilities.
Obscurances for ray-tracing
(Eurographics Association, 2003) Mendez, A.; Sbert, M.; Neumann, L.
We present a powerful method to create realistic-looking pictures of scenes with objects that have diffuse and non-diffuse properties. The method recreates the obscurances technique, introduced some years ago, with a new approach based on ray-tracing. The first version of the obscurances technique was used only for diffuse environments. It is already working successfully in some widely known video games because it is able to avoid the high cost of radiosity techniques in the interactive and real-time game environments. Soft shadows, nice colour reflection effects, visually pleasant rendering of corners and other partly occluded surfaces of the scene are reproduced at a small fraction of the cost of radiosity. We extend here obscurances to handle non-diffuse environments via ray-tracing. Instead of computing the obscurance of every patch in the scene, we will compute view-dependent obscurances. The direct illumination is separately handled and diffuse color albedo functions are used for obscurance computation. Ray-traced obscurances can be useful in animation, both in the editing phase and/or in final images, and in ray-traced games, when the use of future graphics cards will decrease dramatically the cost of tracing a ray.
Real time positioning system based on active stereo and facial features identification
(Eurographics Association, 2003) Frosio, Iuri; Lanzarotti, Raffaella; Campadelli, Paola; Ferrigno, Giancarlo; Borghese, N. Alberto
We propose a new system for patient’s head positioning for radiology and radiotherapy. The system is constituted of two low-cost digital cameras and a digital projector. As a first step a bright pattern is projected onto the subject’s face and a 3D model of the face is reconstructed in real-time. The pattern guides the selection of feature regions of the face: the eyes, the lips and the nose. Inside these regions features are recognized using image processing techniques; 3D many-fold analysis can improve this stage. From these features the position and orientation over time of the face can be determined.
3deSoundBox – a Scalable, Platform-Independent 3D Sound System for Virtual and Augmented Reality Applications
(Eurographics Association, 2003) Stampfl, Philipp
Sound is one of the most important components in animations, presentations and especially immersive environments. Many aspects are recognized more intuitively when sound supports vision and it is easier to follow a story and to believe in a virtual world of illusion. The 3deSoundBox was designed to provide this important acoustic component for such kind of 3D applications in an easy and intuitive way. It is scalable, platform-independent, has an easy to use interface and offers a lot of new possibilities in the field of virtual and augmented reality applications. The 3deSoundBox system can drive any number of speakers and works with any application on any platform with the same range of service. Once programmed, an application can easily take advantage of these features, as it will work on any sound-system with any number of speakers without requiring any changes, due to the scalable architecture of the system.

Browse

Browsing Posters by Issue Date

Results Per Page

Sort Options