11 results
Search Results
Now showing 1 - 10 of 11
Item Robust StructureāBased Shape Correspondence(Ā© 2019 The Eurographics Association and John Wiley & Sons Ltd., 2019) Kleiman, Yanir; Ovsjanikov, Maks; Chen, Min and Benes, BedrichWe present a robust method to find regionālevel correspondences between shapes, which are invariant to changes in geometry and applicable across multiple shape representations. We generate simplified shape graphs by jointly decomposing the shapes, and devise an adapted graphāmatching technique, from which we infer correspondences between shape regions. The simplified shape graphs are designed to primarily capture the overall structure of the shapes, without reflecting precise information about the geometry of each region, which enables us to find correspondences between shapes that might have significant geometric differences. Moreover, due to the special care we take to ensure the robustness of each part of our pipeline, our method can find correspondences between shapes with different representations, such as triangular meshes and point clouds. We demonstrate that the regionāwise matching that we obtain can be used to find correspondences between feature points, reveal the intrinsic selfāsimilarities of each shape and even construct pointātoāpoint maps across shapes. Our method is both time and space efficient, leading to a pipeline that is significantly faster than comparable approaches. We demonstrate the performance of our approach through an extensive quantitative and qualitative evaluation on several benchmarks where we achieve comparable or superior performance to existing methods.We present a robust method to find regionālevel correspondences between shapes, which are invariant to changes in geometry and applicable across multiple shape representations. We generate simplified shape graphs by jointly decomposing the shapes, and devise an adapted graphāmatching technique, from which we infer correspondences between shape regions. The simplified shape graphs are designed to primarily capture the overall structure of the shapes, without reflecting precise information about the geometry of each region, which enables us to find correspondences between shapes that might have significant geometric differences. Moreover, due to the special care we take to ensure the robustness of each part of our pipeline, our method can find correspondences between shapes with different representations, such as triangular meshes and point clouds.Item Incremental Labelling of Voronoi Vertices for Shape Reconstruction(Ā© 2019 The Eurographics Association and John Wiley & Sons Ltd., 2019) Peethambaran, J.; Parakkat, A.D.; Tagliasacchi, A.; Wang, R.; Muthuganapathy, R.; Chen, Min and Benes, BedrichWe present an incremental Voronoi vertex labelling algorithm for approximating contours, medial axes and dominant points (high curvature points) from 2D point sets. Though there exist many number of algorithms for reconstructing curves, medial axes or dominant points, a unified framework capable of approximating all the three in one place from points is missing in the literature. Our algorithm estimates the normals at each sample point through poles (farthest Voronoi vertices of a sample point) and uses the estimated normals and the corresponding tangents to determine the spatial locations (inner or outer) of the Voronoi vertices with respect to the original curve. The vertex classification helps to construct a pieceāwise linear approximation to the object boundary. We provide a theoretical analysis of the algorithm for points nonāuniformly (εāsampling) sampled from simple, closed, concave and smooth curves. The proposed framework has been thoroughly evaluated for its usefulness using various test data. Results indicate that even sparsely and nonāuniformly sampled curves with outliers or collection of curves are faithfully reconstructed by the proposed algorithm.We present an incremental Voronoi vertex labelling algorithm for approximating contours, medial axes and dominant points (high curvature points) from 2D point sets. Though there exist many number of algorithms for reconstructing curves, medial axes or dominant points, a unified framework capable of approximating all the three in one place from points is missing in the literature. Our algorithm estimates the normals at each sample point through poles (farthest Voronoi vertices of a sample point) and uses the estimated normals and the corresponding tangents to determine the spatial locations (inner or outer) of the Voronoi vertices with respect to the original curve. The vertex classification helps to construct a pieceāwise linear approximation to the object boundary. We provide a theoretical analysis of the algorithm for points nonāuniformly (εāsampling) sampled from simple, closed, concave and smooth curves.Item A Variational Approach to Registration with Local Exponential Coordinates(Ā© 2019 The Eurographics Association and John Wiley & Sons Ltd., 2019) Paman, Ashish; Rangarajan, Ramsharan; Chen, Min and Benes, BedrichWe identify a novel parameterization for the group of finite rotations (SO), consisting of an atlas of exponential maps defined over local tangent planes, for the purpose of computing isometric transformations in registration problems that arise in machine vision applications. Together with a simple representation for translations, the resulting system of coordinates for rigid body motions is proper, free from singularities, is unrestricted in the magnitude of motions that can be represented and poses no difficulties in computer implementations despite their multiāchart nature. Crucially, such a parameterization helps to admit varied types of data sets, to adopt dataādependent error functionals for registration, seamlessly bridges correspondence and pose calculations, and inspires systematic variational procedures for computing optimal solutions. As a representative problem, we consider that of registering point clouds onto implicit surfaces without introducing any discretization of the latter. We derive coordinateāfree stationarity conditions, compute consistent linearizations, provide algorithms to compute optimal solutions and examine their performance with detailed examples. The algorithm generalizes naturally to registering curves and surfaces onto implicit manifolds, is directly adaptable to handle the familiar problem of pairwise registration of point clouds and allows for incorporating scale factors during registration.We identify a novel parameterization for the group of finite rotations (SO), consisting of an atlas of exponential maps defined over local tangent planes, for the purpose of computing isometric transformations in registration problems that arise in machine vision applications. Together with a simple representation for translations, the resulting system of coordinates for rigid body motions is proper, free from singularities, is unrestricted in the magnitude of motions that can be represented and poses no difficulties in computer implementations despite their multiāchart nature. Crucially, such a parameterization helps to admit varied types of data sets, to adopt dataādependent error functionals for registration, seamlessly bridges correspondence and pose calculations, and inspires systematic variational procedures for computing optimal solutions. As a representative problem, we consider that of registering point clouds onto implicit surfaces without introducing any discretization of the latter. We derive coordinateāfree stationarity conditions, compute consistent linearizations, provide algorithms to compute optimal solutions and examine their performance with detailed examples. The algorithm generalizes naturally to registering curves and surfaces onto implicit manifolds, is directly adaptable to handle the familiar problem of pairwise registration of point clouds and allows for incorporating scale factors during registration.Item TexNN: Fast Texture Encoding Using Neural Networks(Ā© 2019 The Eurographics Association and John Wiley & Sons Ltd., 2019) Pratapa, S.; Olson, T.; Chalfin, A.; Manocha, D.; Chen, Min and Benes, BedrichWe present a novel deep learningābased method for fast encoding of textures into current texture compression formats. Our approach uses stateāofātheāart neural network methods to compute the appropriate encoding configurations for fast compression. A key bottleneck in the current encoding algorithms is the search step, and we reduce that computation to a classification problem. We use a trained neural network approximation to quickly compute the encoding configuration for a given texture. We have evaluated our approach for compressing the textures for the widely used adaptive scalable texture compression format and evaluate the performance for different block sizes corresponding to 4 Ć 4, 6 Ć 6 and 8 Ć 8. Overall, our method (TexNN) speeds up the encoding computation up to an order of magnitude compared to prior compression algorithms with very little or no loss in the visual quality.We present a novel deep learningābased method for fast encoding of textures into current texture compression formats. Our approach uses stateāofātheāart neural network methods to compute the appropriate encoding configurations for fast compression. A key bottleneck in the current encoding algorithms is the search step, and we reduce that computation to a classification problem. We use a trained neural network approximation to quickly compute the encoding configuration for a given texture.We have evaluated our approach for compressing the textures for the widely used adaptive scalable texture compression format and evaluate the performance for different block sizes corresponding to 4 Ć 4, 6 Ć 6 and 8 Ć 8.Item Efficient Computation of Smoothed Exponential Maps(Ā© 2019 Eurographics ā The European Association for Computer Graphics and John Wiley & Sons Ltd, 2019) Herholz, Philipp; Alexa, Marc; Chen, Min and Benes, BedrichMany applications in geometry processing require the computation of local parameterizations on a surface mesh at interactive rates. A popular approach is to compute local exponential maps, i.e. parameterizations that preserve distance and angle to the origin of the map. We extend the computation of geodesic distance by heat diffusion to also determine angular information for the geodesic curves. This approach has two important benefits compared to fast approximate as well as exact forward tracing of the distance function: First, it allows generating smoother maps, avoiding discontinuities. Second, exploiting the factorization of the global LaplaceāBeltrami operator of the mesh and using recent localized solution techniques, the computation is more efficient even compared to fast approximate solutions based on Dijkstra's algorithm.Many applications in geometry processing require the computation of local parameterizations on a surface mesh at interactive rates. A popular approach is to compute local exponential maps, i.e. parameterizations that preserve distance and angle to the origin of the map. We extend the computation of geodesic distance by heat diffusion to also determine angular information for the geodesic curves. This approach has two important benefits compared to fast approximate as well as exact forward tracing of the distance function: First, it allows generating smoother maps, avoiding discontinuities. Second, exploiting the factorization of the global LaplaceāBeltrami operator of the mesh and using recent localized solution techniques, the computation is more efficient even compared to fast approximate solutions based on Dijkstra's algorithm.Item Appearance Modelling of Living Human Tissues(Ā© 2019 Eurographics ā The European Association for Computer Graphics and John Wiley & Sons Ltd, 2019) Nunes, Augusto L.P.; Maciel, Anderson; Meyer, Gary W.; John, Nigel W.; Baranoski, Gladimir V.G.; Walter, Marcelo; Chen, Min and Benes, BedrichThe visual fidelity of realistic renderings in Computer Graphics depends fundamentally upon how we model the appearance of objects resulting from the interaction between light and matter reaching the eye. In this paper, we survey the research addressing appearance modelling of living human tissue. Among the many classes of natural materials already researched in Computer Graphics, living human tissues such as blood and skin have recently seen an increase in attention from graphics research. There is already an incipient but substantial body of literature on this topic, but we also lack a structured review as presented here. We introduce a classification for the approaches using the four types of human tissues as classifiers. We show a growing trend of solutions that use first principles from Physics and Biology as fundamental knowledge upon which the models are built. The organic quality of visual results provided by these approaches is mainly determined by the optical properties of biophysical components interacting with light. Beyond just picture making, these models can be used in predictive simulations, with the potential for impact in many other areas.The visual fidelity of realistic renderings in Computer Graphics depends fundamentally upon how we model the appearance of objects resulting from the inter action between light and matter reaching the eye. In this paper, we survey the research addressing appearance modelling of living human tissue. Among the many classes of natural materials already researched in Computer Graphics, living human tissues such as blood and skin have recently seen an increase in attention from graphics research. There is already an incipient but substantial body of literature on this topic, but we also lack a structured review as presented here. We introduce a classification for the approaches using the four types of human tissues as classifiers. We show a growing trend of solutions that use first principles from Physics and Biology as fundamental knowledge upon which the models are built.Item Markerless Multiview Motion Capture with 3D Shape Model Adaptation(Ā© 2019 Eurographics ā The European Association for Computer Graphics and John Wiley & Sons Ltd, 2019) Fechteler, P.; Hilsmann, A.; Eisert, P.; Chen, Min and Benes, BedrichIn this paper, we address simultaneous markerless motion and shape capture from 3D input meshes of partial views onto a moving subject. We exploit a computer graphics model based on kinematic skinning as template tracking model. This template model consists of vertices, joints and skinning weights learned a priori from registered fullābody scans, representing true human shape and kinematicsābased shape deformations. Two dataādriven priors are used together with a set of constraints and cues for setting up sufficient correspondences. A Gaussian mixture modelābased pose prior of successive joint configurations is learned to softāconstrain the attainable pose space to plausible human poses. To make the shape adaptation robust to outliers and nonāvisible surface regions and to guide the shape adaptation towards realistically appearing human shapes, we use a meshāLaplacianābased shape prior. Both priors are learned/extracted from the training set of the template model learning phase. The output is a model adapted to the captured subject with respect to shape and kinematic skeleton as well as the animation parameters to resemble the observed movements. With example applications, we demonstrate the benefit of such footage. Experimental evaluations on publicly available datasets show the achieved natural appearance and accuracy.: In this paper, we address simultaneous markerless motion and shape capture from 3D input meshes of partial views onto a moving subject. We exploit a computer graphics model based on kinematic skinning as template tracking model. This template model consists of vertices, joints and skinning weights learned a priori from registered fullābody scans, representing true human shape and kinematicsābased shape deformations. Two dataādriven priors are used together with a set of constraints and cues for setting up sufficient correspondences. A Gaussian mixture modelābased pose prior of successive joint configurations is learned to softāconstrain the attainable pose space to plausible human poses. To make the shape adaptation robust to outliers and nonāvisible surface regions and to guide the shape adaptation towards realistically appearing human shapes, we use a meshāLaplacianābased shape prior. Both priors are learned/extracted from the training set of the template model learning phase.Item Selective Padding for PolycubeāBased Hexahedral Meshing(Ā© 2019 The Eurographics Association and John Wiley & Sons Ltd., 2019) Cherchi, G.; Alliez, P.; Scateni, R.; Lyon, M.; Bommes, D.; Chen, Min and Benes, BedrichHexahedral meshes generated from polycube mapping often exhibit a low number of singularities but also poorāquality elements located near the surface. It is thus necessary to improve the overall mesh quality, in terms of the minimum scaled Jacobian (MSJ) or average SJ (ASJ). Improving the quality may be obtained via global padding (or pillowing), which pushes the singularities inside by adding an extra layer of hexahedra on the entire domain boundary. Such a global padding operation suffers from a large increase of complexity, with unnecessary hexahedra added. In addition, the quality of elements near the boundary may decrease. We propose a novel optimization method which inserts sheets of hexahedra so as to perform selective padding, where it is most needed for improving the mesh quality. A sheet can pad part of the domain boundary, traverse the domain and form singularities. Our global formulation, based on solving a binary problem, enables us to control the balance between quality improvement, increase of complexity and number of singularities. We show in a series of experiments that our approach increases the MSJ value and preserves (or even improves) the ASJ, while adding fewer hexahedra than global padding.Item A Survey on DataāDriven 3D Shape Descriptors(Ā© 2019 The Eurographics Association and John Wiley & Sons Ltd., 2019) Rostami, R.; Bashiri, F. S.; Rostami, B.; Yu, Z.; Chen, Min and Benes, BedrichRecent advances in scanning device technologies and improvements in techniques that generate and synthesize 3D shapes have made 3D models widespread in various fields including medical research, biology, engineering, etc. 3D shape descriptors play a fundamental role in many 3D shape analysis tasks such as point matching, establishing pointātoāpoint correspondence, shape segmentation and labelling, and shape retrieval to name a few. Various methods have been proposed to calculate succinct and informative descriptors for 3D models. Emerging dataādriven techniques use machine learning algorithms to construct accurate and reliable shape descriptors. This survey provides a thorough review of the dataādriven 3D shape descriptors from the machine learning point of view and compares them in different criteria. Also, a comprehensive taxonomy of the existing descriptors is proposed based on the exploited machine learning algorithms. Advantages and disadvantages of each category have been discussed in detail. Besides, two alternative categorizations from the data type and the application perspectives are presented. Finally, some directions for possible future research are also suggested.Recent advances in scanning device technologies and improvements in techniques that generate and synthesize 3D shapes have made 3D models widespread in various fields including medical research, biology, engineering, etc. 3D shape descriptors play a fundamental role in many 3D shape analysis tasks such as point matching, establishing pointātoāpoint correspondence, shape segmentation and labelling, and shape retrieval to name a few. Various methods have been proposed to calculate succinct and informative descriptors for 3D models. Emerging dataādriven techniques use machine learning algorithms to construct accurate and reliable shape descriptors.Item GradientāGuided Local Disparity Editing(Ā© 2019 The Eurographics Association and John Wiley & Sons Ltd., 2019) Scandolo, Leonardo; Bauszat, Pablo; Eisemann, Elmar; Chen, Min and Benes, BedrichStereoscopic 3D technology gives visual content creators a new dimension of design when creating images and movies. While useful for conveying emotion, laying emphasis on certain parts of the scene, or guiding the viewer's attention, editing stereo content is a challenging task. Not respecting comfort zones or adding incorrect depth cues, for example depth inversion, leads to a poor viewing experience. In this paper, we present a solution for editing stereoscopic content that allows an artist to impose disparity constraints and removes resulting depth conflicts using an optimization scheme. Using our approach, an artist only needs to focus on important highālevel indications that are automatically made consistent with the entire scene while avoiding contradictory depth cues and respecting viewer comfort.