GRAPP 2012 Abstracts


Area 1 - Geometry and Modeling

Full Papers
Paper Nr: 2
Title:

SALT WEATHERING OF BRICK WALLS

Authors:

Salman Shahidi

Abstract: Human made buildings and constructions are very dependent upon various weathering phenomena, resulting in complex appearances and patterns that vary with time and environment. Handling such complexity is important in computer graphics in order to improve the realism of images. This difficult work is usually done by hand by designers and can lead to non-plausible results. Another way to tackle this problem is to provide new aging algorithms. Among a large number of weathering, salt weathering results in important visual changes affecting widely used materials: bricks and mortars. In this paper, we present a method to generate the texture of deformed and eroded masonry face by efflorescence. To avoid complex and unintuitive formulations while keeping plausible results, we chose a physically-inspired model. Moreover, in order to keep an important artistic control, it is possible for the user to create weathered textures in preferential masonry zones, leading to a great variety of different patterns. First, we synthesize a masonry (brick/mortar) solid texture. Next, we add the effects of efflorescence and sub-florescence using a specific algorithm accounting for regular crystallized or eroded surface masonry by displacement mapping. In addition, to model very strong weathering effects on building walls, we apply a simple geometry modification method to modify the surface before its rendering.

Paper Nr: 6
Title:

FAST DEFORMATION FOR MODELLING OF MUSCULOSKELETAL SYSTEM

Authors:

Josef Kohout, Petr Kellnhofer and Saulo Martelli

Abstract: This paper proposes a gradient domain deformation for wrapping surface models of muscles around bones as they move during a simulation of physiological activities. Each muscle is associated with one or more poly-lines that represent the muscle skeleton to which the surface model of the muscle is bound so that transformation of the skeleton (caused by the movement of bones) produces transformation of the vertices of the mesh subject to Laplacian linear constraints to preserve the local shape of the mesh and non-linear volume constraints to preserve the volume of the mesh. All these constraints form a system of equations that is solved using the iterative Gauss-Newton method with Lagrange multipliers. Our C++ implementation can wrap a muscle of medium size in about a couple of ms up to 400 ms on commodity hardware depending on the type of parallelization, whilst it can keep the change in volume below 0.04%. A preliminary biomechanical assessment of the proposed technique suggests that it can produce realistic results and thanks to its rapid processing speed, it might be an attractive alternative to the methods that are used in clinical practise at present.

Paper Nr: 13
Title:

CONNECTIVITY MAPS FOR SUBDIVISION SURFACES

Authors:

Ali Mahdavi Amiri and Faramarz Samavati

Abstract: In this paper, we introduce a hierarchical indexing for adjacency queries specially for applying subdivision schemes on some simple spherical and toroidal polyhedrons as the base model for the content creation process. The indexing method is created from integer coordinates of regular 2D domains (connectivity maps) resulting from unfolding polyhedrons. All connectivities are found using algebraic relationships of the connectivity map’s indices; therefore, no additional data structure is required and operations are extremely fast and efficient. Although connectivity relationships of the polyhedrons are as simple as the first resolution, the models created by our proposed method is not restricted to the subdivided polyhedrons. Using our connectivity based method, complex objects can be created by adding sharp features and holes and applying deformation and remeshing techniques. We demonstrate capacities and the efficiency of the method with several example results and compare its speed with that of the half-edge data structure.

Paper Nr: 24
Title:

A POLYHEDRAL APPROACH TO COMPUTE THE GENUS OF A VOLUME DATASET

Authors:

D. Ayala, E. Vergés and I. Cruz

Abstract: Topological characteristics are fundamental in many areas. In this paper we present a method that computes the Euler characteristic and the genus of a volume dataset. The followed approach, based on the analogy between binary volume datasets and orthogonal pseudo-polyhedra (OPP), computes the mentioned values using two models that are suitable for representing OPP: the Extreme Vertices Model (EVM) and the Ordered Union of Disjoint Boxes (OUDB). We show the results of these methods for phantom models as well as real datasets, and compare the efficiency of the presented methods with those based on the classic voxel model.

Paper Nr: 38
Title:

DEBLOCKING FOR DYNAMIC TRIANGLE MESHES

Authors:

Jan Rus and Libor Váša

Abstract: Mesh segmentation (clustering) is a useful tool, which improves compression performance. On the other hand, per-partes processing of meshes often leads to new types of artifacts - cracks and shifts on the borders between clusters. These artifacts are detected by both, Human Visual System (HVS) and perceptually-motivated distortion metrics. In this paper, we present a post processing algorithm, which aims at reducing such artifacts without needing any additional data - using only information about the cluster distribution that is already present at the decoder. A rigid transformation, which minimises the border artifacts, is iteratively computed and applied per cluster. Our experiments show that this approach leads to a reduction of distortion, as measured by the STED metric, by up to 18% for low bitrates. We also present visual results confirming that the improvement is well visible.

Paper Nr: 53
Title:

MUSCLE FIBRES MODELLING

Authors:

Josef Kohout, Gordon J. Clapworthy, Saulo Martelli and Marco Viceconti

Abstract: This paper describes a method that represents a muscle by a realistic chaff of muscle fibres that are automatically generated in the volume defined by the surface mesh of the muscle which itself automatically wraps around bones as they move. Our C++ implementation can decompose the volume into muscle fibres, which is done by a slice-by-slice morphing of predefined fibres template into the muscle volume, and visualise the result in only about 1000 ms on commodity hardware. Hence, the method is fast enough to be suitable for interactive educational medical software. Although a biomechanical assessment has yet to be done, we believe that the method could be used also in clinical biomechanical applications to extract information on the current muscle lever arm and fibre path and, thanks to its rapid processing speed, it might be an attractive alternative to current methods.

Paper Nr: 58
Title:

PATCHWORK TERRAINS

Authors:

Luigi Rocca, Daniele Panozzo and Enrico Puppo

Abstract: We present a radically new method for the management, multi-resolution representation and rendering of large terrain databases. Our method has two main benefits: it provides a Ck representation of terrain, with k depending on the type of base patches; and it supports efficient updates of the database as new data come in. We assume terrain data to come as a collection of regularly sampled overlapping grids, with arbitrary spacing and orientation. A multi-resolution model is built and updated dynamically off-line from such grids, which can be queried on-line to obtain a suitable collection of patches to cover a given domain with a given, possibly view-dependent, level of detail. Patches are combined to obtain a Ck surface. The whole framework can is designed to take advantage of the parallel computing power of modern GPUs.

Paper Nr: 68
Title:

GPU OPTIMIZATION AND PERFORMANCE ANALYSIS OF A 3D CURVE-SKELETON GENERATION ALGORITHM

Authors:

J. Jiménez and J. Ruiz de Miras

Abstract: The CUDA programming model allows the programmer to code algorithms for executing in a parallel way on NVIDIA GPU devices. But achieving acceptable acceleration rates writing programs that scale to thousands of independent threads is not always easy, especially when working with algorithms that have high data-sharing or data-dependence requirements. This type of algorithms is very common in fields like volume modelling or image analysis. In this paper we expose a comprehensive collection of optimizations to be considered in any CUDA implementation, and show how we have applied them in practice in a complex and not trivially parallelizable case study: a 3D curve-skeleton calculation algorithm. Two different GPU architectures have been used to test the implications of each optimization, the NVIDIA GT200 architecture and the new Fermi GF100. As a result, although the first direct CUDA implementation of our algorithm ran even slower than its CPU version, overall speedups of 19x (GT200) and 68x (Fermi GF100) were finally achieved.

Paper Nr: 71
Title:

EFFICIENT MANAGEMENT OF FORESTS FOR REAL TIME RENDERING

Authors:

I. Remolar, J. Gumbau, M. Chover and C. Rebollo

Abstract: Real time rendering of outdoor scenes are currently a problem in spite of the advantages of graphics hardware. Multiresolution modeling has been presented as a solution to this problem, but some management of the geometry is required in order to obtain an interactive visualization. This paper presents an efficient administration of all the geometry involved in a forest. Each tree is represented by two multiresolution models: one for the foliage and one for the trunk. The multiresolution model that represents the foliage varies the level of detail of this part of the plant depending on the importance in the scene. As the observer moves away from the tree, the resolution of the foliage is reduced, collapsing two leaves into one. Regarding the trunk, the multiresolution model that represents this part of the plant is based on multiresolution strips. It uniformly reduces the geometry by performing vertex collapses. The management of such amount of geometry is performed by a Forest Manager. This manager exploits graphics hardware and, for the same species of trees, reuses some level of details previously calculated and allocated in this hardware. This paper demonstrates the validity of our solution in an interactive application.

Paper Nr: 96
Title:

A METHOD FOR SPECIFYING SEMANTICS OF LARGE SETS OF 3D MODELS

Authors:

Xin Zhang, Tim Tutenel, Rong Mo, Rafael Bidarra and Willem F. Bronsvoort

Abstract: Semantics of 3D models is playing a crucial role in games and simulations. In this paper we propose a framework to specify semantics of large sets of 3D models with minimal human involvement. The framework consists of three modules: classification, segmentation and annotation. We associate a few models with tags representing their classes and classify the other models automatically. Once all models have been classified in different groups, we take a certain number of models as template models in each group, and segment these template models interactively. We then use the segmentation method (and parameters) of the template models to segment the rest of the models of the same group automatically. We annotate the interactively segmented parts and use an attributed adjacency graph to represent them. Automatic annotation of the rest of the models is then performed by subgraph matching. Experiments show that the proposed framework can effectively specify semantics of large sets of 3D models.

Short Papers
Paper Nr: 7
Title:

3D MODELING OF STREET BUILDINGS FROM PANORAMIC VIDEO SEQUENCES AND GOOGLE MAP IMAGE

Authors:

Fay Huang, Yi-Ju Wu, Jing-Siang Hsu and Augustine Tsai

Abstract: A semi-automatic image-based framework for modeling of street buildings is proposed in this paper. Two types of image sources are used, one is a sequence of ground-level spherical panoramic images captured by panoramic video recorder, and the other is an aerial image of the desired area obtained from Google Map. The advantages of our approach are first that the camera trajectory recovery result is more accurate and stable due to that the spherical panoramic images are used if compared to multiview planar images. Second, since each face texture of a building is extracted from a single panoramic image, there is no need to deal with color blending problem while textures overlapped.

Paper Nr: 18
Title:

A GENETIC ALGORITHM FOR FACE FITTING

Authors:

David Hunter, Bernard P. Tiddeman and David I. Perrett

Abstract: Accurate estimation of the shape of human faces has many applications from the movie industry to psychological research. One well known method is to fit a Three Dimensional Morphable Model to a target image. This method is attractive as the faces it constructs are already projected onto an orthogonal basis making further manipulation and analysis easier. To date use of Morphable Models have been limited by the inaccuracy and inconvenience of current face-fitting methods. We present a method based on Genetic Algorithms that avoid the local minima and gradient image errors that current methods suffer from. It has the added advantage of requiring no manual interaction to initialise or guide the fitting process.

Paper Nr: 30
Title:

ON THE REAL-TIME PHYSICS SIMULATION OF A SPEED-BOAT MOTION

Authors:

Sergio Casas, Silvia Rueda, José V. Riera and Marcos Fernández

Abstract: Training necessities on watercraft have increased during the last few years and real-time simulators offer a suitable and safe alternative. However, the design of a real-time watercraft simulator implies that, water simulation and water-vehicle interaction have to be addressed efficiently. This paper presents a simplified physics model of the water-vehicle interaction for real-time speed-boat simulators that run over 6-DOF motion platforms. The proposed model is highly parametrizable and can be adapted to any speed-boat by changing the values of the parameters. We propose the evaluation of the designed model in a quantitative and a qualitative way. Evaluations results show that the proposed model behaves like a real one in terms of both objective trajectories and subjective perceived experience.

Paper Nr: 55
Title:

A PROCEDURAL GEOMETRY MODELING API

Authors:

Pedro Brandão Silva, António Coelho, Rui Rodrigues and A. Augusto Sousa

Abstract: This paper presents a solution for geometric manipulation in procedural modeling as an Application Programming Interface (API). This approach intends to enable a more powerful control over the geometric entities by performing selections based on their attributes, similar to picking features in graphical interfaces. This is achieved through the definition of a topological structure, which features a set of properties, such as scope, spatial localization and semantic information. The applicable modeling operations allow a more customized control, as well as successive tracking, which induce a greater, faster and more intuitive approach for geometry generation. This approach still constitutes ongoing work, but has already been successfully applied for the generation of large virtual urban environments.

Paper Nr: 63
Title:

DESIGNING PERSIAN FLORAL PATTERNS USING CIRCLE PACKING

Authors:

Nader Hamekasi and Faramarz Samavati

Abstract: In this paper, we present a novel approach toward generating floral patterns. We extract the essence of a pattern aside from its appearance and geometry into combinatorial elements. As a result, existing patterns can be reshaped while preserving their essence. Furthermore, we can create new patterns that adhere to high level concepts such as imperfect symmetry and visual balance. By decomposing floral patterns into a configuration of circles and angles, we can reconstruct this patterns on different surfaces given a conformal mapping.

Paper Nr: 64
Title:

REVERSE SUBDIVISION FOR OPTIMIZING VISIBILITY TESTS

Authors:

Troy F. Alderson and Faramarz F. Samavati

Abstract: Certain applications require knowledge of whether two entities are visible to each other over a terrain, determined using a line-of-sight computation. Several fast algorithms exist for terrain line-of-sight computations. However, performing numerous line-of-sight computations, particularly over a large terrain data set, can be highly resource-intensive (in run time and/or memory). Methods from the field of terrain simplification can be used to reduce the resource impact of the visibility algorithms. To take advantage of the especially fast algorithms that exist for regular terrain models, we introduce regularity-preserving terrain simplification methods based on reverse subdivision, including a novel reverse subdivision algorithm designed to maximize visibility test accuracy, and compared the resulting visibility test output to several terrain simplification methods. Additionally, the positions of the entities after simplification can have a significant impact on the visibility test results. Hence, we have experimented with different functions that change the positions of the test points in an attempt to maximize visibility test accuracy after simplification.

Paper Nr: 65
Title:

FROM TRIANGULATION TO SIMPLEX MESH AND VICE-VERSA - A Simple and Efficient Conversion

Authors:

Francisco J. Galdames and Fabrice Jaillet

Abstract: We propose an accurate method to convert from a triangular mesh model to a simplex mesh and vice-versa. For this, we are taking advantage of the fact that they are topologically duals, turning it into a natural swap between these two models. Unfortunately, they are not geometrically equivalents, leading to loss of information and to geometry deterioration when performing the conversion. Therefore, optimal positions of the vertices in the dual mesh have to be found while avoiding shape degradation. An accurate and effective transformation technique is described in this paper, where we present a direct method to perform an appropriate interpolation of a simplex mesh to obtain its dual, and/or vice-versa. Our method is based on the distance minimization between the local tangent planes of the mesh and vertices of each face.

Paper Nr: 75
Title:

SURFACE-BASED SEGMENTATION OF 3D POINT CLOUD DATA IN THE VECTOR FIELD REPRESENTATION

Authors:

Van Tung Nguyen and Denis Laurendeau

Abstract: In this paper, we propose an approach for 3D segmentation of point cloud data based on the vector field representation. A volumetric surface representation named the vector field is first computed for the input point cloud. The data description in a voxel of the vector field is then classified into eight surface types by computing the sign of curvatures for the closest point of the voxel on the surface. A region-growing sheme based on bivariate fuctions fitting is finally applied to the closest points of these voxels for refining the segmented regions. This surface-based approach is entirely designed in the vector field surface representation and the advantages of using the vector field are discussed. Experiments demonstrate the peformance of the method.

Paper Nr: 77
Title:

MULTI-RESOLUTION DUAL CONTOURING FROM VOLUMETRIC DATA

Authors:

Ricardo Uribe Lobello, Florent Dupont and Florence Denis

Abstract: We present a Multi-Resolution Dual method based on an incremental octree-based refinement strategy. Our solution is able to generate multi-resolution surfaces from segmented volumetric data. It extends the Dual Marching Cubes algorithm over a generalized octree and guarantees that the produced surfaces are always manifold by introducing a new cell-based criterion for dual vertices generation. Moreover, we propose a top-bottom refinement algorithm that is able to locally adapt the mesh resolution according to a curvature parameter. Our algorithm is suitable to process volumetric data sets and we show on different volumes that the produced surfaces are manifold and approximate well the original object.

Paper Nr: 103
Title:

SEMANTIC AND TOPOLOGICAL REPRESENTATION OF BUILDING INTERIORS - An Overview

Authors:

B. Domínguez, Á. L. García and F. R. Feito

Abstract: Software related to urban environments has experimented a considerable growth over the last years. This trend makes necessary the creation and management of urban models, both for cities and building indoors. There exists a huge variety of proposals dealing with the geometry, topology and semantics of building indoors. In this paper we give an overview of a number of recent works from several research areas such as Geographic Information Systems (GIS), Building Information Models (BIM), Building Product Models (BPM) and Computer Aided Design (CAD) tools and propose a reference architecture which gathers the analyzed features.

Paper Nr: 104
Title:

SURFACE RECONSTRUCTION FROM 3D MEDICAL IMAGES BASED ON TRI-TREE CONTOURING - Seeking Geometrically Valid Surfaces

Authors:

Rubén Pulido, Juan J. Jiménez and Félix Paulano

Abstract: Surface reconstruction from 3D medical images is fundamental for bio-medical applications. Nevertheless, the generation of valid geometric surfaces is a difficult task due to the complexity of the structures necessary to represent the human body. In this paper, we describe the main strategies concerning surface reconstruction and planar contour extraction algorithms. In addition, we propose a new approach to surface reconstruction from medical images. Our approach is divided into two main parts. In the first part, contour extraction is performed using a hierarchical spatial decomposition. In the second part, these contours can be triangulated using a table of patterns in order to obtain geometrically valid surfaces.

Posters
Paper Nr: 10
Title:

FAST WAY TO CREATE SEAM BOUNDARY FOR SQUARE PARAMETERIZATION WITH LOW-DISTORTION

Authors:

Anuwat Dechvijankit, Hiroshi Nagahashi and Kota Aoki

Abstract: In order to parameterize any three-dimensional surface like a closed boundary one, we need to convert its polygonal mesh into a disk topology surface. For the quality of texturing or re-meshing that uses a parameterization technique, it is more effective if the distortion of two-dimensional manifold planar domain map is as small as possible. We introduce a fast way to create seams along the three-dimensional surface for square boundary of a two-dimensional planar domain map. This paper describes a faster and less error technique than the existing seam cutting methods by simply observing the location of high distortion area in the planar domain. The distortion rate of our proposed method is low as the original one but with less time consuming.

Paper Nr: 46
Title:

TREE MODELING WITH DSM DATA

Authors:

Keonsoo Park, Jehoon Park, Choi Ji-Hoon, Sun-Jeong Kim and Chang-Geun Song

Abstract: This study aims to resolve the problem of not being able to directly examine forests or each individual tree of a forest. In order to get the specific information on actual trees. Such as their locations, heights and the number of trees. We used an aerial photograph that is 4096x4096 pixels. And process the DSM/DEM data with a raw 16 bit-‘unsigned short’ data value. Through the collected information, we might model trees and a forest.

Paper Nr: 46
Title:

TREE MODELING WITH DSM DATA

Authors:

Keonsoo Park, Jehoon Park, Choi Ji-Hoon, Sun-Jeong Kim and Chang-Geun Song

Abstract: This study aims to resolve the problem of not being able to directly examine forests or each individual tree of a forest. In order to get the specific information on actual trees. Such as their locations, heights and the number of trees. We used an aerial photograph that is 4096x4096 pixels. And process the DSM/DEM data with a raw 16 bit-‘unsigned short’ data value. Through the collected information, we might model trees and a forest.

Paper Nr: 79
Title:

ORTHOGONAL SIMPLIFICATION OF OBJECTS REPRESENTED BY THE EXTREME VERTEX MODEL

Authors:

Irving Cruz-Matías and Dolors Ayala

Abstract: This work presents a new approach to simplify objects. It is restricted to orthogonal pseudo-polyhedra (OPP) and binary volumes. The method produces a level-of-detail (LOD) sequence of OPP and it is incremental. Any object in this sequence is a bounding OPP of the previous objects. The sequence finishes with the minimum axis-aligned bounding box. OPP are represented by the Extreme Vertices Model. Simplification is achieved using a new approach called merging faces that relies on the application of 2D Boolean operations.

Paper Nr: 82
Title:

A 3D PHYSICS-BASED MODEL TO SIMULATE NORMAL AND PATHOLOGICAL GAIT PATTERNS

Authors:

Christian Cifuentes, Fabio Martínez and Eduardo Romero

Abstract: This article presents a novel 3D physics-based human gait model that allows to quantify and simulate the dynamic patterns of normal and pathological movements in the sagittal and coronal views, using an enhanced inverted pendulum approach. The method outperforms the classic planar representations that do not consider important gait phases like the double stance phase and the heelstrike, crucial in proper gait descriptions on clinical routine. The model was assessed by simulating gait cycles and comparing the obtained trajectories with actual normal and pathological gait data. Results showed that the normal and pathological kinematic patterns generated by our model are highly similar to the actual data, obtaining an accuracy of about 87%.

Paper Nr: 101
Title:

HARMONIC OSCILLATIONS MODELLING FOR THE PURPOSE OF CAMERA SYNCHRONIZATION

Authors:

Tomislav Pribanic, Adam Hedi and Vedran Gracanin

Abstract: The goal of this paper to present a general method to synchronize cameras from the image modelled harmonic oscillations, which are in this work produced by a simple pendulum. The method essence is to recover from camera images a sine trajectory of a small ball, attached to 45cm string and suspended from a pivot so that it can swing freely. From this trajectory and given an equilibrium position of ball in space, for each camera a time is estimated needed to reach this equilibrium position from a neighboring frame during a ball swing. The difference in those times for two cameras yields a subframe time difference between cameras. The subframe time differences between cameras were computed using the proposed method and they were compared against the ground truth measurement values given by a camera manufacturer. Linear correlation coefficient is 0.996, while the mean absolute difference between two methods measurements is 3.5ms.

Paper Nr: 102
Title:

MEPP - 3D MESH PROCESSING PLATFORM

Authors:

Guillaume Lavoué, Martial Tola and Florent Dupont

Abstract: This paper presents MEPP, an open source platform for 3D mesh processing. This platform already contains a large set of processing tools from classical ones (simplification, subdivision, segmentation) to more technical algorithms (compression, watermarking, Boolean operation, perceptual metrics, etc.). Its main objective is to allow a quick start for both users and developers by providing highly detailed tutorials and simple integration mechanisms, through a modular architecture where components are implemented as dynamic plugins.

Area 2 - Rendering

Full Papers
Paper Nr: 51
Title:

IMPROVING RAY TRAVERSAL BY USING SEVERAL SPECIALIZED KD-TREES

Authors:

Roberto Torres, Pedro J. Martín, Antonio Gavilanes and Luis F. Ayuso

Abstract: In this paper, we present several variants of the Surface Area Heuristics (SAH) to build kd-trees for specific sets of rays’ directions. In order to cover the whole space of directions, several sets of directions are considered and each of them leads to a different specialized kd-tree. We call Multi-kd-tree to the set of these kd-trees. During rendering, each ray will traverse the kd-tree associated with the set containing its direction. In order to evaluate the efficiency of our proposal, we have implemented a Path Tracing and an Ambient Occlusion renderer on GPU with CUDA. A SAH-based kd-tree has been compared to a Multi-kd-tree and we show that all the new heuristics exhibit a better performance than SAH over usual scenes.

Paper Nr: 60
Title:

VISUALIZATION OF LONG SCENES FROM DENSE IMAGE SEQUENCES USING PERSPECTIVE COMPOSITION

Authors:

Siyuan Fang and Neill Campbell

Abstract: This paper presents a system for generating multi-perspective panoramas for long scenes from dense image sequences. Panoramas are created by combining different perspectives, including both original and novel perspectives. The latter are rendered using our perspective synthesis algorithm, which employs geometrical information to eliminate the sampling error distortion caused by depth parallax of on-planar scenes. Our approach for creating multi-perspective panoramas is different from existing methods in that a perspective composition framework is presented to combine various perspectives to form a panorama without undesired visual artifacts, through suppressing both colour inconsistencies and structural misalignments among input perspectives. We show that this perspective composition can facilitate the generation of panoramas from user specified multi-perspective configurations.

Short Papers
Paper Nr: 14
Title:

REVERSE PAINTERLY RENDERING

Authors:

Ying Wang and Masahiro Takatsuka

Abstract: Traditional painterly rendering algorithms transform photographic images to their artistic representations such as paintings or drawings. However, the reverse process of painterly rendering has never been discussed, which could be useful for studies in heritage and forensic science. This paper presents a novel reverse painterly rendering algorithm which restores “oil painting-like” images to photo-like images. Results show that our algorithm can improve the photorealistic appearance of “oil painting-like” images.

Paper Nr: 16
Title:

USER-CENTERED DESIGN OF GPU-BASED SHADER PROGRAMS

Authors:

Martin Kraus

Abstract: In the context of game engines with graphical user interfaces, shader programs for GPUs (graphics processing units) are an asset for game development that is often used by artists and game developers without knowledge of shader programming. Thus, it is important that non-programmers are enabled to explore and exploit the full potential of shader programs. To this end, we develop principles and guidelines for the design of usercentered graphical interfaces for shaders. With the help of several examples, we show how the requirements of a user-centered interface design influence the choice of widgets as well as the choice of the underlying shader parameters.

Paper Nr: 29
Title:

GPU-BASED VISUALIZATION OF HYBRID TERRAIN MODELS

Authors:

E. G. Paredes, M. Bóo, M. Amor, J. Döllner and J. D. Bruguera

Abstract: Hybrid terrain models formed by a large regular mesh refined with detailed local TIN meshes represent an interesting and efficient approach for the representation of complex terrains. However, direct rendering of the component meshes would lead to overlapping geometries and discontinuities around their boundaries. The Hybrid Meshing algorithm solves this problem by generating an adaptive tessellation between the boundaries of the component meshes in real-time. In this paper, we present a highly parallel implementation of this algorithm using the Geometry Shader on the GPU.

Paper Nr: 37
Title:

LOW-RANK RADIOSITY USING SPARSE MATRICES

Authors:

Eduardo Fernández, Pablo Ezzatti, Sergio Nesmachnow and Gonzalo Besuievsky

Abstract: Radiosity methods are part of the global illumination techniques, which deal with the problem of generating photorealistic images in 3D scenes with Lambertian surfaces. Low-rank radiosity is a O(nk) method, where n is the number of polygons and k is the rank of the matrix used as a direct transport operator. This method allows calculating, in real-time and with infinite bounces, the illumination of a scene with static geometry and dynamic lighting. In this paper we present a new methodology for low-rank radiosity calculation based on the use of sparse matrices, which significantly reduces the memory storage required and achieves speedup improvements over the original low-rank method. Experimental analysis was performed in both traditional computers and new graphics processing unit architectures.

Paper Nr: 57
Title:

CURVATURE-APPROXIMATED ESTIMATION OF REAL-TIME AMBIENT OCCLUSION

Authors:

Tomohito Hattori, Hiroyuki Kubo and Shigeo Morishima

Abstract: We present a novel technique for computing ambient occlusion (Akenine-Möller et al., 2008) on real-time graphics hardware. Our method approximates the occlusion for a local illumination model by introducing curvature-dependent function. Using our method, we are able to acquire occlusion at lower computational cost than conventional methods such as SSAO (Bavoil et al., 2008). Our method requires a multi-pass algorithm with the graphics processing unit (GPU). In the first pass curvature is acquired, and in the second pass the occlusion is computed from the curvature. In the calculating occlusion from the curvature, we approximate the geometric shape by a quadric surface function, and then obtain a curvature dependent function which is an approximation of geometric surface. This function depends only on local variables and we are able to calculate the ambient occlusion for the local illumination model. According to our results, both the computational and storage costs are sufficiently low for the technique to be applied to current graphics hardware supported computer games.

Paper Nr: 69
Title:

SYNTHESIS OF MULTIRESOLUTION SCENES WITH GLOBAL ILLUMINATION ON A GPU

Authors:

Raquel Concheiro, Margarita Amor, Montserrat Bóo, Iago Iglesias, Emilio J. Padrón and Ramón Doallo

Abstract: The radiosity computation has the important feature of producing view independent results, but these results are mesh dependent and, in consequence, are attached to a specific level of detail in the input mesh. Therefore, rendering at iterative frame rates would benefit from the utilization of multiresolution models. In this paper we focus on the rendering stage of a solution for hierarchical radiosity for multiresolution systems. This method is based on the application of an enriched hierarchical radiosity algorithm to an input scene with low resolution objects (represented by coarse meshes), and the efficient data management of the resulting values. The proposed encoding makes it possible to apply the color values obtained for the coarse objects to detailed versions of these objects during the rendering phase. These finer meshes are obtained by a standard mesh subdivision strategy, such as the Loop subdivision scheme. Our solution performs the whole rendering stage of this multiresolution approach on the GPU, implementing it in the geometry shader using Microsoft HLSL. Results of our implementation show an important reduction in computational costs.

Paper Nr: 93
Title:

REAL-TIME VIDEO MATTING FOR MIXED REALITY USING DEPTH GENERATED TRIMAPS

Authors:

Nicholas Beato, Remo Pillat and Charles E. Hughes

Abstract: In Mixed Reality scenarios, background replacement is a common way to immerse a user in a synthetic environment. Properly identifying the background pixels in an image or video is a difficult problem known as matting. Proper alpha mattes usually come from human guidance, special hardware setups, or color dependent algorithms. This is a consequence of the under-constrained nature of the per pixel alpha blending equation. While the field of natural image matting has made progress finding a least squares solution for an alpha matte, the generation of trimaps, indicating regions of known foreground and background pixels, normally requires human interaction or offline computation. We overcome these limitations by combining a low fidelity depth image that segments the original video signal with a real-time parallel natural image matting technique that favors objects with similar colors in the foreground and background. This allows us to provide real-time alpha mattes for Mixed Reality scenarios that take place in relatively controlled environments. As a consequence, while monochromatic backdrops (such as green screens or retro-reflective material) aid the algorithm’s accuracy, they are not an explicit requirement.

Posters
Paper Nr: 19
Title:

REAL-TIME WALKTHROUGH RENDERING - Speed Improvement of Photon Mapping Algorithm

Authors:

Florian Graglia, Jean Sequeira and Sébastien Mavromatis

Abstract: The main purpose of this paper is to discuss the global illumination methods in the context of real-time walkthrough. Our work focuses on an accurate illumination of complex scenes; the intensity of the lights can also be interactively modified. This feature is particularly relevant within the context of a production pipeline: After the 3D modeling, photorealistic walkthrough is often used to detect inappropriate reflections or to measure illumination rates. The known methods that usually work in real-time do not simulate all light paths for large scenes. However, some methods provide a full global illumination with a computation time close to a few seconds. This paper shows how this calculation time can be reduced to approach real-time animation thanks to a specific method derived from the photon mapping.

Paper Nr: 22
Title:

REAL-TIME AMBIENT OCCLUSION ON THE PLAYSTATION3

Authors:

Dominic Goulding, Richard Smith, Lee Clark, Gary Ushaw and Graham Morgan

Abstract: This paper describes how to implement ambient occlusion effects on the Playstation3 (PS3) while alleviating processing demands on the GPU. The solutions proposed here are implementations that utilize the parallel processing available on the PS3’s synergistic processing units (SPUs). Two well-known ambient occlusion techniques are evaluated as candidate solutions for PS3 SPU implementations.

Paper Nr: 27
Title:

IMAGE LIGHTING EFFECT MANIPULATION FOR AN EFFICIENT STYLIZATION

Authors:

Catherine Sauvaget and Vincent Boyer

Abstract: We propose a model to stylize images through the manipulation of different lighting effects and their stylization. Given an input image, we propose to generate semi-automatically a map containing each effect identified by the user. Then, we propose different distortions and coloring to obtain artistic stylizations. Our model is flexible and well designed to help users who desire to stylize their images following light effects. This model is intended for an image editing tool dedicated to comics stylizations.

Area 3 - Animation and Simulation

Full Papers
Paper Nr: 8
Title:

GENERATING 3D MORPHABLE MODEL PARAMETERS FOR FACIAL TRACKING - Factorising Identity and Expression

Authors:

Alexander Davies, Carl Henrik Ek, Colin Dalton and Neill Campbell

Abstract: The ability to factorise parameters into identity and expression parameters is highly desirable in facial tracking as it requires only the identity parameters to be set in the initial frame leaving the expression parameters to be adjusted in subsequent frames. In this paper we introduce a strategy for creating parameters for a data-driven 3D Morphable Model (3DMM) which are able to separately model the variance due to identity and expression found in the training data. We present three factorisation schemes and evaluate their appropriateness for tracking by comparing the variances between the identity coefficients and expression coefficients when fitted to data of individuals performing different facial expressions.

Paper Nr: 89
Title:

VISUAL SIMULATION OF MAGNETIC FLUIDS

Authors:

Tomokazu Ishikawa, Yonghao Yue, Kei Iwasaki, Yoshinori Dobashi and Tomoyuki Nishita

Abstract: In this paper, we focus on simulation of magnetic fluids. Magnetic fluids behave as both fluids and as magnetic bodies, and these characteristics allow them to generate ‘spike-like’ shapes along a magnetic field. Magnetic fluids are popular materials for use in works of art. Our goal is to simulate such works of art. In the field of electromagnetic hydrodynamics, many methods have also been proposed for simulating such spike shapes based on numerical fluid analysis. However, those methods are computationally expensive and they typically require tens of hours just to simulate a single spike. We propose a more efficient method by combining a procedural approach and the SPH method (smoothed particle hydrodynamics). Our method simulates overall behaviors of the magnetic fluids using the SPH method and then synthesizes the spike shapes by using the procedural approach. We demonstrate our method can generate visually plausible results within a reasonable computational cost.

Paper Nr: 91
Title:

SKETCHING FLUID FLOWS - Combining Sketch-based Techniques and Gradient Vector Flow for Lattice-Boltzmann Initialization

Authors:

Sicilia Ferreira Judice and Gilson Antonio Giraldi

Abstract: This work proposes an intuitive fluid flow initialization for computer graphics applications. A combination of sketching techniques and Gradient Vector Flow is proposed to obtain a smooth initialization for the simulation using a Lattice Boltzmann Method (LBM). The LBM is based on the fundamental idea of constructing simplified kinetic models, which incorporates the essential physics of microscopic processes so that the macroscopic averaged properties satisfy macroscopic equations. The application of sketching techniques is proposed in order to enable the user to draw freely the initial state of the fluid flow using an intuitive interface. Moreover, it will be possible for the user to define multiply connected domains with suitable boundary conditions.

Short Papers
Paper Nr: 25
Title:

ANATOMICALLY CORRECT ADAPTION OF KINEMATIC SKELETONS TO VIRTUAL HUMANS

Authors:

Christian Rau and Guido Brunnett

Abstract: The adaption of a given kinematic skeleton to a triangular mesh representing the shape of a human is a task still done manually in most animation pipelines. In this work we propose methods for automating this process. By thoroughly analyzing the mesh and utilizing anatomical knowledge, we are able to provide an accurate skeleton alignment, suited for animating the surface of a virtual human based on motion capture data. In contrast to existing work, our methods on the one hand support the anatomically correct adaption of detailed skeletal structures, but on the other hand are robust against the reduction of the mesh’s tessellation complexity.

Paper Nr: 72
Title:

A GENERAL MOTION REPRESENTATION - Exploring the Intrinsic Viewpoint of a Motion

Authors:

Gutemberg Guerra-Filho

Abstract: We propose a novel motion representation, named General Motion Representation (GMR), which explicitly contains the absolute world coordinates of the meaningful joints, while still specifying every other joint with rotational data relative to their respective parent joints. More specifically, our representation supports multiple roots where any joint can be a root due to the use of a novel pre-rotation format for the construction of local transformation matrices. Hence, our general motion representation allows for all intermediate data structures between fully rotational and fully translational data. The use of multiple roots also implies the representation of partial motion considering a subset of joints. We introduce the general motion representation to consider multiple roots with the support of three operations (shift root, split skeleton tree, and join skeleton trees). These operations allow reduced skeletal complexity because of the application of pre-rotation local transformation matrices which eliminates the requirement for dummy joints. We also present procedures to convert from raw marker data or post-rotation formats. We demonstrate the highly efficient computation of per-frame joint positions and orientations. Our experimental results show that GMR outperforms traditional motion formats in both speed and flexibility. At the full translational configuration, GMR is around seven times faster than bvh.

Posters
Paper Nr: 20
Title:

MODEL OF AGGREGATION - A Topological Approach

Authors:

Masud Rana and Dongsheng Cai

Abstract: The aggregate motion of flocks of birds, a herd of land animals, Mexican wave forming in stadia are beautiful and nice examples of collective behaviour. The aggregation is constructed by the action of each individual, each action solely on basis of its local perception of the world. Scientists from different backgrounds have tried to model collective behaviour. Most of the models are strictly metric (based on Euclidian distance among individuals) but flocks of birds do not act on metric perception. In this paper we proposed a model based on topological perspective to construct a flock of birds with large number of individuals and checked flock’s density independent behaviour.

Paper Nr: 39
Title:

SIERPINSKI JELLY - Iterated Function Systems as Elastic Bodies

Authors:

Pawel Filipczuk, Slawomir Nikiel and Korneliusz Warszawski

Abstract: Relatively simple ideas of fractal geometry result in an infinite number of complex images and objects. Fractals are used in computer graphics to increase visual fidelity of the vector models. Although derived from dynamical systems, fractals are usually presented as static objects. The paper presents a new concept of embedding physical description in the model of IFS (Iterated Function System). Dynamically changing fractal structures offer better sense of ‘material’ than static images or key-framed animations. The model can augment IFS attractors with the illusion of softness, weight and other material-related features. The proposed model is flexible, deterministic and offers high rendering performance.

Paper Nr: 54
Title:

SIMULATION SYSTEM FOR DANCE GROUPS USING A GAMEPAD

Authors:

Asako Soga and Itsuo Yoshida

Abstract: A dance group requires a formation, quite unlike a solo dance, and thus the arrangement on the stage is critical. We have developed a system to simulate the formations of dance groups. The simulation is executed by specifying the number of dancers, the formation pattern, and the dancing motion with a gamepad. Typical formations used for classical ballet pieces were analyzed, and three formation patterns were obtained. Each dancer’s position is calculated by the number of dancers and the selected formations. The system arranges dancers on a virtual stage and simulates dance animations by using motion-capture data obtained from a professional dancer.

Paper Nr: 85
Title:

AN AUTOMATIC APPROACH FOR FACIAL FEATURE POINTS EXTRACTION FROM 3D HEAD

Authors:

Achraf Othman, Oussama El Ghoul and Mohamed Jemni

Abstract: This work presents a novel approach for 3D head segmentation as well as for extraction of facial features vertices. These tasks are the preliminary task for retrieval, recognition, classification and tracking processes. 3D head regions are detected from a cloud of points using the ratio of geodesic to Euclidean distances between pairs of vertices that combine neighbor coordinates and position information of the point. In this paper, we propose a new algorithm to extract automatically feature points inside the facial region based on the topology of 3D face. Furthermore, in order to validate our approach, we have run experimentations on several types of full 3D head that differ in number of points and surfaces, gender and ethnic. Also, to approve its robustness, we apply our approach on a benchmark of scanned faces containing 67 models.

Area 4 - Interactive Environments

Full Papers
Paper Nr: 49
Title:

AN OCCLUSION-AWARE AR AUTHORING TOOL FOR ASSEMBLY AND REPAIR TASKS

Authors:

Jesús Gimeno, Pedro Morillo, Juan Manuel Orduña and Marcos Fernández

Abstract: The use of authoring tools has become a valuable trend for the fast development of Augmented Reality (AR) applications in industrial organizations. However, most of current AR authoring tools are actually programming interfaces that are exclusively suitable for programmers. Moreover, they do not provide advanced visual effects, such as occlusion or object collision detection, to the resulting AR applications. In this paper, we propose an easy-to-use AR authoring tool oriented to the development of AR applications for the execution of industrial sequential procedures. Unlike other recent easy-to-use AR authoring tools, this software framework allows non-programming users to develop low-cost AR applications, including occlusion capabilities, by means of the use of a Kinect sensor. These applications could be used in on-site assembly and maintenance/repair tasks where a certain level of depth perception is needed. In order to validate our AR authoring tool, we have developed four AR applications belonging to different industrial areas. The evaluation results show that overlaying 3D instructions on the actual work pieces reduces the error rate for an assembly task by more than a 75%, particularly diminishing cumulative errors common in sequential procedures. Also, the results show that the time required by non-programming users to develop the AR prototypes using our tool was more than 90% lower than the time required for developing the same prototypes with computer graphics programmers. These results show the potential of our approach and validate the tool as a general-purpose AR authoring tool for industrial AR applications.

Paper Nr: 76
Title:

DISTANCE-ADAPTED 2D MANIPULATION TECHNIQUES FOR LARGE HIGH-RESOLUTION DISPLAY ENVIRONMENTS

Authors:

Anke Lehmann and Oliver Staadt

Abstract: Physical navigation presents an intuitive overview+detail technique in large high-resolution display environments where users interact at different distances. Thus, the user interaction should be adapted to the user’s current position. This ongoing work presents bimanual interaction techniques that use distance-adapted mapping of physical hands motion to virtual relative objects motion (named interaction scaling). Interaction scaling allows the user to manipulate high-resolution content with the appropriate accuracy at diverse distances. In our initial experiment, we considered two relative mapping functions: continuous mapping and zone mapping. The mapping functions manipulate the 2D position of the virtual cursor. The parameters for the 2D manipulation tasks (i.e., scaling, rotation, translation) are calculated from the virtual cursor behavior. We developed two bimanual interaction techniques for 2D object manipulation that support the interaction scaling: separatedcursors and connected-cursors method. The separated-cursors technique differentiates the users hands between dominant hand and nondominant hand. The connected-cursors technique uses an additional midpoint cursor that is located between both virtual cursors. In an explorative study we evaluated the separated-cursors technique with and without interaction scaling. An interesting result is that the participants often needed less object selections to sort objects with interaction scaling compared to the scenario without interaction scaling.

Paper Nr: 107
Title:

AN EFFICIENT APPROACH TO RENDER 3D MESHES BY MANAGING MULTIRESOLUTION TRIANGLE STRIPS

Authors:

Francisco Ramos, Jesus Gumbau and Miguel Chover

Abstract: Visualization of large 3D scenes is a problem often solved by means of multiresolution modeling or level of detail. In this paper, we present a uniform resolution model that noticeably improves existing models, in terms of storage and visualization cost. The model is entirely based on optimized hardware primitives, triangle strips. Management of triangle strips coherence, on a multiresolution mesh, is key to achieving optimum performance. This model is able to take advantage of coherence in a software level as well as directly on the graphics hardware, integrating part of the model in that hardware. Use of stripification techniques, oriented to exploit vertex cache, has been taken into account to minimize vertex reprocessing. Comparisons to existing multiresolution model implementations show improvements of approximately 25% in storage space cost, 40% in level-of-detail extraction cost and visualization as much as 5 times better by applying hardware acceleration techniques.

Short Papers
Paper Nr: 4
Title:

WORKFLOW MODEL FOR COLLABORATIVE 3D INTERACTION - Application to Co-located Manipulation

Authors:

Christophe Domingues, Samir Otmane, Frédéric Davesne and Malik Mallem

Abstract: Collaborative Virtual Environments (CVEs) are complex environments where multiple users share the same objects to act together. The complexity of these environments is related to the interaction of the user group with shared items. In a shared world, several constraints appear, including those related to coordination and communication of users and user interaction management in virtual world towards objects and other users. This paper presents a useful workflow-based approach to manage 3D interactions and enhance presence in CVEs. This approach combines astutely different concepts from two research domains: CSCW and VR.

Paper Nr: 12
Title:

NEW HYBRID P2P COMMUNICATION MODELS FOR REMOTE TERRAIN INTERACTIVE VISUALIZATION SYSTEMS

Authors:

Ricardo Olanda, Mariano Pérez and Juan Manuel Orduña

Abstract: Over the last years, there has been a great development on real time terrain visualization applications using remote databases. One of the main problems that these applications must face is the system scalability. These applications usually use a client-server model that cannot support a large number of concurrent requests without using a considerable number of servers. In this paper, we propose a new hybrid P2P models for terrain interactive visualization systems. The comparative performance evaluation results show that the system throughput achieved by these strategies can be more than three times higher than the hybrid P2P strategy proposed in the literature, significantly improving the scalability of these systems.

Paper Nr: 34
Title:

AUGMENTED REALITY WITH AUTOSTEREOSCOPIC VISUALIZATION - A Comparative Study using an Autostereoscopic Display versus a Common Display

Authors:

Juan-J. Arino, M.-Carmen Juan, Santiago González-Gancedo, Ignacio Seguí and Roberto Vivó

Abstract: In this paper, we present a system that combines Augmented Reality and autostereoscopic visualization. We also report a study for comparing different aspects using an autostereoscopic display and a common display, in which 44 children aged from 8 to 10 years old have participated. From our study, statistically significant differences were found between both displays for the depth perception and for the sense of presence. Several correlations have also been found when children used the autostereoscopic display. In our study, the sense of presence is closely related with the depth perception; and the overall score of the game was also closely related with the depth perception and the sense of presence.

Paper Nr: 41
Title:

AN IMMERSIVE VIRTUAL ENVIRONMENT FOR PHANTOM LIMB PAIN REHABILITATION

Authors:

Steve Pettifer, Toby Howard, Ben Blundell, David Edwards and Ilan Lieberman

Abstract: Phantom Limb Pain is a debilitating condition that affects a significant percentage of patients after loss of an arm or leg. These patients experience chronic pain and other unpleasant sensations in the missing limb, and the pain resists treatment. Previous research has demonstrated that pain levels can be reduced in some patients when they are immersed in a virtual environment that presents a 3D computer graphics visualisation of their missing limb, the movements of which are controlled by sensors attached to the remaining limb. In this paper we describe a novel approach to the implementation of such a system, using the Kinect game device for limb motion tracking, in conjunction with wireless motion sensors worn by the patient. We present some preliminary, but very encouraging, results based on an informal trial with a patient.

Paper Nr: 42
Title:

TOWARDS A MIXED REALITY LEARNING ENVIRONMENT IN THE CLASSROOM

Authors:

Santiago González-Gancedo, M.-Carmen Juan, Ignacio Seguí, Noemí Rando and Juan Cano

Abstract: In this paper, we present a novel study that emphasizes the use of Augmented Reality (AR) as a natural complement for the Virtual Reality Learning Environment (VRLE) model, towards a general acceptance of Mixed Reality Learning Environment (MRLE) in the classroom. Handheld devices help this scheme serving as general purpose computers available for use by other applications. AR has not been explored deeply enough to have full acceptance of use in the classroom. We present an application in which a tablet PC was used to evaluate our game, working with multimodal interaction provided by a tactile screen and an accelerometer. It can be played in two modes: combining AR and non-AR (NAR), and only NAR. Seventy-three children of primary school tested the system. For the learning outcomes, there were no statistically significant differences between both modes, but the AR mode enhanced highly user satisfaction and engagement. This confirms our hypothesis that AR can be an excellent complement to VRLE for the use in the classroom.

Paper Nr: 50
Title:

A KINECT-BASED AUGMENTED REALITY SYSTEM FOR INDIVIDUALS WITH AUTISM SPECTRUM DISORDERS

Authors:

Xavier Casas, Gerardo Herrera, Inmaculada Coma and Marcos Fernández

Abstract: In this paper an Augmented Reality system for teaching key developmental abilities for individuals with ASD is described. The system has been designed as an augmented mirror where users can see themselves in a mirror world augmented with virtual objects. Regarding the content the tool is designed with the aim of facilitating the acquisition of certain skills in children with ASD. From a technical point of view, it has been necessary to solve two problems to develop the tool. The first one has been how to capture the user’s movements, without wearing invasive devices, to animate their virtual avatar. A Kinect device has been chosen. The other problem has been to mix in the augmented virtual scene real objects located at different depths. An OSG shader has been developed, of which details will given in this paper.

Paper Nr: 59
Title:

VOLUME RENDERING STRATEGIES ON MOBILE DEVICES

Authors:

José M. Noguera, Juan-Roberto Jiménez, Carlos J. Ogáyar and Rafael J. Segura

Abstract: This paper proposes and compares several methods for interactive volume rendering in mobile devices. This kind of devices has several restrictions and limitations both in performance and in storage capacity. The paper reviews the suitability of some existing direct volume rendering methods, and proposes a novel approach that takes advantage of the graphics capabilities of modern OpenGL ES 2.0 enabled devices. Several experiments have been carried out to test the behaviour of the described method.

Paper Nr: 66
Title:

AN AUGMENTED REALITY SYSTEM BASED ON LIGHT FIELDS

Authors:

A. Quintana, R. Quirós, I. Remolar and E. Camahort

Abstract: The development of augmented reality systems that combine virtual elements with the real world is currently increasing. This paper presents an augmented reality system that renders the virtual elements from models based on the light fields. The use of these models allow us to obtain higher level of photorealism than representations that render polygons. Using light fields also allows us to keep constant the rendering time. The presented system has been implemented using the Open Source library ARToolKit and a spherical model of the light field based on the direction-and-point parameterization (DPP) with some associated depth information. The system has been validated using different light fields, and it has been compared its performance with a classic version of the ARToolKit library based on VRML files. The presented augmented reality system can be applied to the visual inspection of synthetic objects of great complexity or based on real images.

Paper Nr: 67
Title:

STEREOSCOPIC VISION IN DESKTOP AUGMENTED REALITY - User Performance in the Presence of Conflicting Depth Cues

Authors:

Gustavo Rovelo, Francisco Abad, M.-Carmen Juan and Emilio Camahort

Abstract: The use of stereoscopy as a depth cue in current computer graphics technology is increasingly popular. On the other hand, in a Desktop Augmented Reality (AR) environment, a web cam and a set of printed markers allow the user to interact with an augmented view of the environment in the computer display. So, it seems a logical step for Desktop Augmented Reality systems to take advantage of the available stereoscopic hardware to improve the realism of the augmented scenes. The main problem of Desktop AR applications is that they will be used in different and unprepared environments (wide variety of hardware, inconsistent and unpredictable lighting conditions, etc.) Therefore, it is crucial to understand how adverse conditions affect user experience. Furthermore, Desktop AR applications usually present conflicts in the depth cues presented to the users: viewpoint offset and incorrect occlusions between real and virtual objects. We use the within subjects experimental approach to evaluate user performance in an Augmented Reality game. Our goal is to find if stereoscopic graphics help to reduce the impact of the other depth cues conflicts inherent to Desktop AR applications.

Paper Nr: 84
Title:

AN ARCHITECTURAL PATTERN FOR X3D-BASED VIRTUAL ENVIRONMENTS - An Object-oriented Approach

Authors:

George Anastassakis and Themis Panayiotopoulos

Abstract: X3D is an open, royalty-free, scene-graph-based standard for 3D virtual worlds that has been used by virtual environment developers many times until today. As a virtual environment development tool, it has numerous appealing features; however, it also has certain characteristics that may, under certain circumstances, create significant problems to developers. With a motivation to communicate our own experience to, and contribute to the efforts of, other researchers who are using or consider using it, we evaluate X3D in a virtual environment developer's perspective and identify potential problems with its usage. Also, we present an architectural pattern aimed at addressing those problems effectively, with a focus on transparency, standards-compliance, reusability and extendibility. In conclusion, we present a case study of the application of the proposed pattern in a fully-implemented intelligent virtual environment system.

Paper Nr: 88
Title:

ON-LINE 3D BODY MODELLING FOR AUGMENTED REALITY

Authors:

Luis Almeida, Paulo Menezes and Jorge Dias

Abstract: Building 3D body models is an important task for virtual and augmented reality applications in telerehabilitation, education, 3DTV, entertainment and tele-presence. We propose a real-time full 3D reconstruction system that combines visual features and shape-based alignment using low cost depth sensor and video cameras targeting three-dimensional conferencing applications. With this approach we overcome the classic video based reconstruction problem in low-texture or repeated pattern regions. Alignment between successive frames is computed by jointly optimizing over both appearances and shape matching. Appearance-based alignment is done over 2D SURF features annotated with 3D position. Shape-based alignment is performed using the motion transformation estimation between consecutive annotated 3D point clouds through a linear method. A solution to avoid wrong annotated 3D matched points is proposed. 3D mesh model representation is used to lower the processed data and create a 3D representation that is independent of view-point.

Paper Nr: 100
Title:

MSQUASH - An Exergame using the Playstation Move Controller

Authors:

Mauro F. M. Ferrão, Kelly O'Hara and Frutuoso G. M. Silva

Abstract: The exergames became very popular in last years because of the appearance of the motion controllers with accelerometers but also because this type of games promotes the physical exercise. Msquash is a 3D game that simulates a squash field where the player can play using the Move controller. The game was developed with the objective to create an exergame using the Move motion controller and also to compare the Move controller performance with Wii Remote. Thus the idea is providing a fun and engaging experience for players while they do physical exercise also. The Msquash game was developed using OGRE and OpenCV libraries with Playstation Move controller and Playstation eye camera. This paper presents a brief description of the Msquash game and some results based on tests performed with users to evaluate the interaction method used and its gameplay.

Paper Nr: 105
Title:

A COMPARATIVE STUDY OF IMPLEMENTED COLLISION DETECTION STRATEGIES

Authors:

Félix Paulano, Juan J. Jiménez, Rubén Pulido and Carlos J. Ogayar

Abstract: Many computer interactive environments require to calculate collision detections between geometrical models, usually triangle meshes. In this paper, a comparative study of some collision detection strategies is realized. For this purpose, the compared algorithms have been implemented. Some of these algorithms also allow realizing other procedures, such as calculating distances or overlapping triangles. Hence, the main features of each algorithm have been explained. The compared strategies have been tested considering, in addition to the runtime, the pre-processing time and the memory usage. Finally, the results have been compared in order to extract the benefits and constraints of each strategy, and proposing some possible applications for each strategy.

Posters
Paper Nr: 26
Title:

NAVIGATION AND INTERACTION IN URBAN ENVIRONMENTS USING WEBGL

Authors:

María Dolores Robles-Ortega, Lidia Ortega, Francisco R. Feito and Manuel J. González

Abstract: The process of rendering and interacting with large scenes in web systems is still an open problem in 3D urban environments. In this paper we propose a prototype to visualize a city model in a client-server architecture using open-source technologies like WebGL and X3DOM. Moreover, free navigation around the scene is allowed and users are able to obtain additional information when interacting with buildings and street furniture. To achieve this objective, aMySQL geodatabase has been designed to store both geometric and non-geometric urban information. Therefore, the extra data about the urban elements is obtained through queries in the database. The communication process between MySQL and the X3D model is performed by Ajax.

Paper Nr: 33
Title:

TELARSCOPE: A GAMEBASED TOURISTIC AUGMENTED REALITY TELESCOPE

Authors:

Dominik Grüntjens, Daniel Arndt, Jenny Beschorner, Jonas Dietterle, Jens Epe, Paul Kobold, Tobias Stüttem and Stefan Müller

Abstract: TelARscope is a gamebased touristic augmented reality telescope. Its interaction and exploration concept is sufficient for a heterogeneous user group such as tourists. It enables not only to watch augmented content overlayed at all times statically but also to explore it completely interactively. We evaluated the telescope with 91 real tourists.

Paper Nr: 43
Title:

SIMULATING CHARACTERS FOR OBSERVATION

Authors:

A. Albin-Clark, T. L. J. Howard and B. Anderson

Abstract: Observation of young children is common in educational settings but student practitioners have infrequent access so limited amounts of observable activity can ever be captured. We have developed a software application using research-based models of child development to support observations made by pedagogical practitioners. The prototype system – Observation – which we believe is the first to employ 3D interactive computer graphics for visualising early childhood play, is available for download and evaluation.

Paper Nr: 56
Title:

AN INEXPENSIVE 3D STEREO HEAD-TRACKED DUAL VIEWING SINGLE DISPLAY

Authors:

Armando de la Re, Eduardo Martorell, Gustavo Rovelo, Emilio Camahort and Francisco Abad

Abstract: Dual view displays allow two users to collaborate on the same task. Specialized computer systems have been designed to provide each user with a private view of the scene. This paper presents two hardware configurations that support independent stereo viewing for two users sharing a common display. Both prototypes are built using off-the-shelf parts: a 120Hz computer monitor and a 120Hz projector. The result is a functional, responsive and inexpensive dual-view 3D stereo display system where both viewers see stereo on the same display. We demonstrate our systems’ features with three demos: a competitive game, where each user has a private stereoscopic view of the game table; a two-user multimedia player, where each user can watch and hear a different stereoscopic video on the same display and a head-tracked dual stereo 3D model viewer that provides each user with a correct perspective of the model. The latter demostrator also provides basic gesture-based interaction.

Paper Nr: 61
Title:

AN URBAN ONTOLOGY TO GENERATE COLLABORATIVE VIRTUAL ENVIRONMENTS FOR MUNICIPAL PLANNING AND MANAGEMENT

Authors:

Tiago Martins, Pedro Brandão Silva, António Coelho and A. Augusto Sousa

Abstract: The municipal planning and management tasks are generally performed based on text documents or through digital maps provided by geographic information systems (GIS). However, most municipal GIS follow different data models, leading to interoperability problems when there is a need to combine data from different sources. Furthermore, most of the time these tasks are performed in a collaborative way between the municipal technicians, emerging some difficulties in decision making due to the three-dimensional nature of urban space. Thus, this paper describes a methodology which can integrate multiple sources of real data from diverse municipal GIS, in a unified data model based on the CityGML specification. This model is mapped onto an urban ontology oriented for procedural modeling, which, in turn, produces the three-dimensional models of the urban environments. The system developed operates in a client-server approach, where the server is responsible for mapping the urban information to the unified data model and the client represents the procedural modeling technology that generates the urban environment in three-dimensional format, allowing users to interact and amplify the existing urban information.

Paper Nr: 74
Title:

A GIS DATA UPDATING METHOD WITH VIDEO

Authors:

Ruijun Liu, Myriam Servières and Guillaume Moreau

Abstract: We present a method for extracting and updating Geographic Information System (GIS) data from video. This paper uses a ground-view video sequence as input data. In the first step, feature points are extracted from video, which provide the structure information for architecture facade reconstruction; after discarding outliers, we create the footprint and reconstruct a 3D model. The second step is the back-projection process. In the third step, we correct the footprint and update it by using user knowledge. Finally, this information is used to update GIS.

Paper Nr: 78
Title:

AUDIOVISUAL RECORDING SYSTEM FOR E-LEARNING APPLICATIONS

Authors:

Al. L. Ronzhin

Abstract: Paper presents the audiovisual recording system, which has been implemented in the SPIIRAS smart room and provides conduct of e-learning meetings. The techniques of video tracking and sound source localization are implemented for recording AVI files of speaker messages in the room. Video processing of streams from five cameras serves for registration participants in fixed chair positions, tracking main speaker and recording view to the audience. During the experiments AVI files with speaker’s messages were recorded in a discussion work mode. After manual checking it was detected that 90% of files are files with speaker’s speech and 10% of files are false files with noises. The verification of the system was accomplished on the functional level and also the estimations of detection quality of participants, and camera pointing on speaker and speaker message had acceptable rate.

Paper Nr: 97
Title:

ICA-BASED ACTION RECOGNITION FOR HUMAN-COMPUTER INTERACTION IN DISTURBED BACKGROUNDS

Authors:

Wei-Yao Chiu and Du-Ming Tsai

Abstract: In this paper, we propose an independent component analysis (ICA) based scheme for action recognition. The proposed method does not require the feature design or modeling process, and it is computationally very fast for real-time applications. Experimental results show that it is robust on disturbed backgrounds where both foreground and background objects are moving simultaneously in the scene. The experiments also reveal that the proposed method is very effective for human-computer interaction.

Paper Nr: 106
Title:

TEACHING MATH AND PHYSICS BY DECONSTRUCTING GRAPHICS

Authors:

Amit Shesh

Abstract: The foundations of computer graphics theory lie in mathematics, whereas most effects that add realism to a virtual graphical world are a direct simulation of the rules of physics. Complementary to the role of math and physics in learning about computer graphics, this paper proposes using graphics which has widespread appeal to directly encourage learning math and physics. We aim to use graphics to kindle interest in learning basic math, physics and by extension, computer science at the high school and college level, by directly and immediately matching theoretical concepts to how they are practically used in familiar graphics effects. We propose to achieve this by creating a virtual navigable environment of arbitrary complexity that, upon simple interaction, automatically de-constructs its constituent objects and various visual effects to reveal how they were created as well as relevant physics such as light and object behavior. We concentrate on two technical problems (a) setting up a visually appealing environment with ease for users like instructors who are not necessarily familiar with technical details of graphics and (b) de-constructing, illustrating and viewing various effects as automatically as possible so that users like students and teachers concentrate on the concepts and not how best to illustrate them.

Paper Nr: 108
Title:

GESTURE RECOGNITION - Control of a Computer with Natural Head Movements

Authors:

Kornél Bertók and Attila Fazekas

Abstract: The topic of this article is a basic research considering a human-computer interaction. The system is still under construction, however its basis – facial features tracking and head pose estimation – is ready to use, thus it could bring a head gesture controlled system into reality. We present an approach to control applications with head movements. We construct an Active Appearance Model (AAM) for facial feature extraction. Based on the landmarks of AAM shape, the head pose is modelled in the three-dimensional space by three Euler angles of rotation around the three axes orthogonal to each other, and three orthogonal translation directions. Regarding all the above mentioned facts, we are capable of recognizing gestures like head pointing, nodding, and shaking, blinking, opening and closing mouth.