WSCG 2010: Full Papers Proceedings
Permanent URI for this collection
Browse
Recent Submissions
Item Classification of regions extracted from scene images by morphological filters in text or non-text using decision tree(Václav Skala - UNION Agency, 2010) Luz Alves, Wonder Alexandre; Hashimoto, Ronaldo Fumio; Skala, VáclavWe present in this work a new method to classify regions extracted from scene images by morphological filters in text or nontext region using a decision tree. Our technique can be divided into three parts. Firstly, we extract a set of regions by a robust scheme based on morphological filters. Then, after a refinement, a set of text attributes is obtained for each region. In the last step, a decision tree is built in order to classify them as text or non-text regions. Experiments performed using images from the ICDAR public dataset show that this method is a good alternative for practical problems involving text location in scene images.
Item Facial Feature Detection and Tracking with a 3D Constrained(Václav Skala - UNION Agency, 2010) Yu, Meng; Tiddeman, Bernard P.; Skala, VáclavIn this paper, we describe a system for facial feature detection and tracking using a 3D extension of the Constrained Local Model (CLM) [Cris 06, Cris 08] algorithm. The use of a 3D shape model allows improved tracking through large head rotations. CLM uses a joint shape and texture appearance model to generate a set of region template detectors. A search is then performed in the global pose / shape space using these detectors. The proposed extension uses multiple appearance models from different viewpoints and a single 3D shape model. During fitting or tracking the current estimate of pose is used to select the appropriate appearance model. We demonstrate our results by fitting the model to image sequences with large head rotations. The results show that the proposed 3D constrained local model algorithm improves the performance of the original CLM algorithm for videos with large out-of-plane head rotations.Item Modeling with Subdivision Surfaces(Václav Skala - UNION Agency, 2010) Ullrich, T.; Schiefer, A.; Fellner, D. W.; Skala, VáclavSubdivision surfaces are an established modeling tool in computer graphics and computer-aided design. While the theoretical foundations of subdivision surfaces are well studied, the correlation between a control mesh and its subdivided limit surface still has some open-ended questions: Which topology should a control mesh have? Where should control vertices be placed? A modeler – human or software – is confronted with these questions and has to answer them. In this paper we analyze four characteristic situations. Each one consists of an analytical reference surface S and several variants of control meshes Ci. In order to concentrate on the topology of the control meshes, the geometrical positions of their control vertices have been determined and optimized automatically. As a result we identified the best topology of all Ci to represent the given surface S. Based on these results we derived heuristics to model with subdivision surfaces. These heuristics are beneficial for all modelers.Item Simulation of Massive Multibody Systems(Václav Skala - UNION Agency, 2010) Tasora, Alessandro; Negrut, Dan; Anitescu, Mihai; Mazhar, Hammad; Heyn, Toby David; Skala, VáclavWe describe an efficient method for the simulation of complex scenarios with millions of frictional contacts and mechanical constraints. To this end, the GPU processors of the modern graphic boards are used to solve the differential inclusion problem that represents the most challenging part of the multi–rigid–body problem. Thank to the massive parallelism offered by GPU boards, we are able to simulate sand, granular materials, soils and other complex physical scenarios with a large speedup respect to serial CPU–based algorithms.Item Feature preserving mesh smoothing algorithm based on local normal covariance(Václav Skala - UNION Agency, 2010) Svub, Miroslav; Krsek, Přemysl; Španěl, Michal; Štancl, Vít; Bartoň, Radek; Vaďura, Jiří; Skala, VáclavOur goal is to develop a smoothing algorithm, which would be feature preserving and simple to use without the need of extensive parameter tuning. Our method does the smoothing of vertices based on local neighbourhood character, which is modeled by a covariance matrix of neighbourhood triangle normals. The eigenvalues and eigenvectors of the covariance matrix are used for local weighting of the displacement vector of laplacian operator. This way the method is locally auto-tuned.Item Histograms of Oriented Gradients for 3D Object Retrieval(Václav Skala - UNION Agency, 2010) Scherer, Maximilian; Walter, Michael; Schreck, Tobias; Skala, Václav3D object retrieval has received much research attention during the last years. To automatically determine the similarity between 3D objects, the global descriptor approach is very popular, and many competing methods for extracting global descriptors have been proposed to date. However, no single descriptor has yet shown to outperform all other descriptors on all retrieval benchmarks or benchmark classes. Instead, combinations of different descriptors usually yield improved performance over any single method. Therefore, enhancing the set of candidate descriptors is an important prerequisite for implementing effective 3D object retrieval systems. Inspired by promising recent results from image processing, in this paper we adapt the Histogram of Oriented Gradients (HOG) 2D image descriptor to the 3D domain. We introduce a concept for transferring the HOG descriptor extraction algorithm from 2D to 3D. We provide an implementation framework for extracting 3D HOG features from 3D mesh models, and present a systematic experimental evaluation of the retrieval effectiveness of this novel 3D descriptor. The results show that our 3D HOG implementation provides competitive retrieval performance, and is able to boost the performance of one of the best existing 3D object descriptors when used in a combined descriptor.Item A Multi-Cellular Orthographic Projection Approach to Image-Based Rendering(Václav Skala - UNION Agency, 2010) Saraiva, Carlos; Oliveira, João Fradinho; Pereira, João Madeiras; Skala, VáclavImage-based rendering (IBR) techniques take the approach of rendering new images based on existing ones, effectively separating the rendering complexity from the geometric complexity of the scene they represent. With the increasing complexity of scenes currently created, these techniques have gained renewed interest. This paper presents a multi-cellular IBR approach, which uses orthographic rather than perspective projected pre-computed images. Specifically we extend our single cell orthographically projected IBR approach to a multi-cell system, and show how the pixel error can be further reduced with the use of smaller and more numerous cells.Item A Narrow Band Level Set Method for Surface Extraction(Václav Skala - UNION Agency, 2010) Rosenthal, Paul; Molchanov, Vladimir; Linsen, Lars; Skala, VáclavLevel-set methods have become a valuable and well-established field of visualization over the last decades. Different implementations addressing different design goals and different data types exist. In particular, level sets can be used to extract isosurfaces from scalar volume data that fulfill certain smoothness criteria. Recently, such an approach has been generalized to operate on unstructured point-based volume data, where data points are not arranged on a regular grid nor are they connected in form of a mesh. Utilizing this new development, one can avoid an interpolation to a regular grid which inevitably introduces interpolation errors. However, the global processing of the level-set function can be slow when dealing with unstructured point-based volume data sets containing several million data points. We propose an improved level-set approach that performs the process of the level-set function locally. As for isosurface extraction we are only interested in the zero level set, values are only updated in regions close to the zero level set. In each iteration of the level-set process, the zero level set is extracted using direct isosurface extraction from unstructured point-based volume data and a narrow band around the zero level set is constructed. The band consists of two parts: an inner and an outer band. The inner band contains all data points within a small area around the zero level set. These points are updated when executing the level set step. The outer band encloses the inner band providing all those neighbors of the points of the inner band that are necessary to approximate gradients and mean curvature. Neighborhood information is obtained using an efficient kd-tree scheme, gradients and mean curvature are estimated using a four-dimensional least-squares fitting approach. Comparing ourselves to the global approach, we demonstrate that this local level-set approach for unstructured point-based volume data achieves a significant speed-up of one order of magnitude for data sets in the range of several million data points with equivalent quality and robustness.Item Heuristic Convergence Rate Improvements of the Projected Gauss–Seidel Method for Frictional Contact Problems(Václav Skala - UNION Agency, 2010) Poulsen, Morten; Niebe, Sarah; Erleben, Kenny; Skala, VáclavIn interactive physical simulation, contact forces are applied to prevent rigid bodies from penetrating and control slipping between bodies. Accurate contact force determination is a computationally hard problem. Thus, in practice one trades accuracy for performance. The result is visual artifacts such as viscous or damped contact response. In this paper, we present heuristics for improving performance for solving contact force problems in interactive rigid body simulation. We formulate the contact force problem as a nonlinear complementarity problem, and discretize the problem using a splitting method and a minimum map reformulation. The resulting model is called the Projected Gauss–Seidel method. Quantitative research results are presented and can be used as a taxonomy for selecting a suitable heuristic when using the Projected Gauss–Seidel method.Item Simulating Atmospheric Pollution Weathering on Buildings(Václav Skala - UNION Agency, 2010) Mérillou, N.; Mérillou, S.; Ghazanfarpour, D.; Dischler, J. M.; Galin, M.; Skala, VáclavInteractions between polluted atmosphere and materials lead to an early aging of numerous buildings and monuments. This weathering process leads to important changes in appearance, from color blackening to smallscale geometric alterations: a black crust grows onto parts of affected surfaces, depending on the geometry of the object as well as on its environment. In this paper, we present a method to simulate this very important weathering process. Our method is physically inspired and provides full control to designers, keeping plausible results. First, specific polluted zones are detected according to their real physical classification. Then, the modifications of aspect of each zone are computed. Our results demonstrate that our model matches well the observed behavior of real-world monuments and buildings affected by atmospheric pollution.Item A Facade Tracking System for Outdoor Augmented Reality(Václav Skala - UNION Agency, 2010) Martins, José F.; Silva, Jorge A.; de Sousa, A. Augusto; Skala, VáclavWe describe a real-time facade tracking system that uses, as setup information, only two images of a facade, captured on the moment. No more previous information is needed, such as a facade 3D model, dimensions or aspect ratio. Feature points and their local descriptors are extracted from that pair of images and used during the detection and tracking of the facade. Additionally, parallax and topological information is also used in order to increase the overall robustness of the tracking process. Experiments show that the system can detect and track a wide variety of facades, including those that are not entirely planar, partially occluded or have few distinguishable visual landmarks. The reliance on on-the-spot information, alone, makes this system useful for Outdoor Augmented Reality applications, in an Anywhere Augmentation urban context.Item 3D Facial Animation for Mobile Devices(Václav Skala - UNION Agency, 2010) De Martino, José Mario; Leite, Tatiane Silvia; Skala, VáclavThis article presents the implementation of a 3D facial animation system for mobile devices. Due to the large processing and memory requirements for this type of application, its use on mobile devices was not possible until recently. Currently, however, with the increasing development of powerful hardware and with the spread of cellular telephony, 3D applications for these devices have become even more promising. The system presented in this article was implemented based on a previous one designed for desktop platforms. Both systems are driven by two kinds of input. The first one is an audio file containing the speech to be presented, and the second is a timed phonetic transcription specifying the phonemes of the utterance and their associated time spans. With the input data, the system converts phonemes into phonetic context dependent visemes that handle perseveratory and anticipatory coarticulation effects. The visemes are then used to modify the 3D polygonal mesh of the fa ce keeping the synchronization with the audio. The initial evaluation of the systems indicates that, despite the limitations of mobile platform, the implementation of video telephony using 3D facial animation is viable in modern mobile devices.Item Adaptive Real-Time Rendering of Planetary Terrains(Václav Skala - UNION Agency, 2010) Lerbour, Raphaël; Marvie, Jean-Eudes; Gautron, Pascal; Skala, VáclavAs virtual worlds applications get more and more demanding in terms of world complexity and rendering quality, rendering virtual terrains and planets in real-time introduces many challenges. In this paper, we provide a full-featured solution for rendering of arbitrary large terrain datasets in the context of client-server streaming. Our solution automatically adapts to arbitrary network bandwidths and client capabilities, ranging from high-end computers to mobile devices. We address the problem of rendering highly complex terrain databases comprising several hundred gigabytes of data, which therefore cannot be entirely loaded in memory nor rendered in real-time. The contributions of this paper solve important issues for high quality terrain rendering: adaptive texture mapping, inexpensive removal of geometry cracks and support of planetary terrains.Item A comparative analysis of B-spline deformation models in 3D shape matching(Václav Skala - UNION Agency, 2010) Langerak, T. R.; Song, Y.; Skala, VáclavLattice-based B-spline driven deformation has become a useful method in shape matching applications. Here, the challenge is to find a configuration of the B-spline deformation model that effectuates a deformation that spatially aligns one shape (the source) to another shape (the target). Literature study indicates that few B-Spline deformation based algorithms were implemented that target polygonal meshes. In contrast, in the field of medical image registration, B-spline deformation has been extensively applied in matching shapes that use a voxel-based shape representation. For exploring the opportunities of applying these voxel-based methods to the shape matching of polygonal meshes, in this paper we propose to match polygon meshes by transforming them to voxel models and apply established techniques from the field of medical imaging. Two voxel-based methods are selected and implemented: Global Optimization methods, which globally optimize the B-spline model, and Markov Random Field methods, which locally optimize the B-spline model. These methods are compared to parameterized B-spline-based shape matching methods previously proposed by the authors. These methods directly match polygon meshes. Results indicate that the proposed methodology outperforms the parameterized approach in terms of accuracy and computation time and therefore is a promising alternative to existing methods.Item Enhanced Visual Depth Cues for Collocated Visuo-Haptic Augmented Reality(Václav Skala - UNION Agency, 2010) Knörlein, B.; Székely, G.; Harders, M.; Skala, VáclavOur current research focuses on the application of visuo-haptic augmented reality in medical training. The setup developed in this context enables collocated haptic interaction with scene objects. In order to allow natural manipulations, provision of appropriate depth cues becomes a crucial factor. Therefore, we have included fast occlusion handling and shadow synthesis in our augmented environment. The occlusion map is initialized using a plane sweep approach, followed by an edge-based optimization via a Mumford-Shah functional. For obtaining the depth map three head mounted cameras are used and a leftright consistency check is performed to provide robustness against half occlusions. Shadowing is implemented via shadow mapping, considering both real and virtual objects. All steps have been implemented on GPU shaders and are performed in real-time.Item Geo-Spatial Hypermedia based on Augmented Reality(Václav Skala - UNION Agency, 2010) Kim, Sung-Soo; Kim, Kyong-Ho; Jang, Sie-Kyung; Lim, Jin-Mook; Wohn, Kwang-Yun; Skala, VáclavIn order to produce convincing geographic information, it is not only necessary to provide real-world videos according to the user’s geographic locations but also non-spatial data in the videos. We present a novel approach for efficient linking heterogeneous data of the same objective nature, such as 2D maps, 3D virtual environments and videos with GPS data. Our approach is motivated by the observation that we can get the non-spatial data in a video by transforming the video search space into the 3D virtual world search space which contain the non-spatial attributes according to the remotely sensed GPS data. Our proposed system consists of two primary components: a client layer that implements the augmentation algorithm for geo-spatial video in exploiting server resources, and a server layer designed to provide the non-spatial attributes according to the client’s requests. To implement attribute query in a video, we present an easily implementable data model that serves well as a foundation for point query in 2D. In order to apply this to telematics applications such as car navigation systems, we propose a live-video processing method using augmented reality technology according to the user’s locations along navigation path. Experimental results demonstrate that our approach is effective in retrieving geospatial video clips and nonspatial data.Item Real-Time Gaze Tracking with a Consumer-Grade Video Camera(Václav Skala - UNION Agency, 2010) Keil, Andrea; Albuquerque, Georgia; Berger, Kai; Magnor, Marcus Andreas; Skala, VáclavEye gaze can be a rich source of information to identify particular interests of human users. Eye gaze tracking has been largely used in different research areas in the last years, as for example in psychology, visual system design and to leverage the user interaction with computer systems. In this paper, we present an IR-based gaze tracking framework that can be easily coupled to common user applications and allows for real-time gaze estimation. Compared to other gaze tracking systems, our system uses only affordable consumer-grade hardware and still achieves fair accuracy. To evaluate the usability of our gaze tracking system, we performed a user study with persons of different genders and ethnicities.Item Modeling trajectories of free moving objects with smooth flow fields(Václav Skala - UNION Agency, 2010) Nykolaychuk, Mykhaylo; Rössl, Christian; Theisel, Holger; Richter, Klaus; Skala, VáclavWe present a new method for smooth global analysis and representation of image-based motion data with topological methods for flow fields, useful for further activity analysis. A video sensor captures motion of subjects and yields discrete samples of velocity vectors in the image domain. We show how to construct a smooth, continuous, globally defined bidirectional flow field which approximates the directional part of given samples. We then apply a topological analysis of this continuous flow field which yields a segmentation into regions of similar flow behavior, i.e., regions of similar motion. This segmentation (topology) can be seen as global model of typical motions and used for further analysis like detecting atypical motions. We tested our method empirically and provide results for two different scenarios with human motion and traffic motion, respectively.Item A Simple and Effective CAPTCHA by Exploiting the Orientation of Sub-images Cropped from Whole-size Photos(Václav Skala - UNION Agency, 2010) Chung, Woo-Keun; Ji, Seung-Hyun; Kim, Jong-Woo; Cho, Hwan-Gue; Skala, VáclavAutomated detection of image orientation has previously been studied as an important problem in intelligent image processing and computer vision. For this problem, numerous methods and tools have been developed by adopting approaches such as objects segmentation, color feature analysis and machine learning e.g., Support Vector Machines(SVMS). But conversely, the difficulty of image orientation can be used to examine the robustness of a CAPTCHA(Completely Automated Public Turing test to Tell Computers and Human Apart). The automated image orientation problem previously only had been studied and solved using typical photos which almost include important semantic cues such as people, bright sky, dark ground and vertical edges. In this paper we propose a simple prototype CAPTCHA, which exploits the hardness of orienting sub-images cropped from a whole digital photo. Our CAPTCHA takes 8 sub-images from base-photos and rotates them randomly. Then we present them to the user, who is required to find the correct orientations of the 8 sub-images. The true orientation is easily obtained since most current high-end digital cameras have an automatic mechanism to store its orientation in EXIF. Thus we can simply and easily obtain the image orientation without applying complicated computation. For our experiment, we have collected about 1850 base photos that provide more than 100,000 different sub-images. Experiment showed that the accuracy of our CAPTCHA with humans is about 95%. We think this sub-image orientation is hard to solve by an automated procedure since all previous machine learning procedures have only considered whole photos with enough semantic cues, rather than partial image segments. Another advantage of our system is that user interaction is simpler(there are four choices) and more intuitive than a common text-based system or the previous image orientation method with arbitrary rotation. Experiment showed that common users performed at most two rotations for each sub-image. The total time to complete orienting the 8 sub-image orientation was less than 15 seconds which is significantly shorter than that of previous image-based CAPTCHAs.Item Selection of an Optimal Set of Discriminative and Robust Local Features with Application to Traffic Sign Recognition(Václav Skala - UNION Agency, 2010) Höferlin, Benjamin; Heidemann, Gunther; Skala, VáclavToday, discriminative local features are widely used in different fields of computer vision. Due to their strengths, discriminative local features were recently applied to the problem of traffic sign recognition (TSR). First of all, we discuss how discriminative local features are applied to TSR and which problems arise in this specific domain. Since TSR has to cope with highly structured and symmetrical objects, which are often captured at low resolution, only a small number of features can be matched correctly. To alleviate these issues, we provide an approach for the selection of discriminative and robust features to increase the matching performance by speed, recall, and precision. Contrary to recent techniques that solely rely on density estimation in feature space to select highly discriminative features, we additionally address the question of features’ retrievability and positional stability under scale changes as well as their reliability to viewpoint variations. Finally, we combine the proposed methods to obtain a small set of robust features that have excellent matching properties.
