cvpr cvpr2013 cvpr2013-289 knowledge-graph by maker-knowledge-mining
Source: pdf
Author: Abed Malti, Richard Hartley, Adrien Bartoli, Jae-Hak Kim
Abstract: We propose a new approach for template-based extensible surface reconstruction from a single view. We extend the method of isometric surface reconstruction and more recent work on conformal surface reconstruction. Our approach relies on the minimization of a proposed stretching energy formalized with respect to the Poisson ratio parameter of the surface. We derive a patch-based formulation of this stretching energy by assuming local linear elasticity. This formulation unifies geometrical and mechanical constraints in a single energy term. We prevent local scale ambiguities by imposing a set of fixed boundary 3D points. We experimentally prove the sufficiency of this set of boundary points and demonstrate the effectiveness of our approach on different developable and non-developable surfaces with a wide range of extensibility.
Reference: text
sentIndex sentText sentNum sentScore
1 We extend the method of isometric surface reconstruction and more recent work on conformal surface reconstruction. [sent-2, score-1.184]
2 Our approach relies on the minimization of a proposed stretching energy formalized with respect to the Poisson ratio parameter of the surface. [sent-3, score-0.523]
3 We derive a patch-based formulation of this stretching energy by assuming local linear elasticity. [sent-4, score-0.48]
4 This formulation unifies geometrical and mechanical constraints in a single energy term. [sent-5, score-0.489]
5 We prevent local scale ambiguities by imposing a set of fixed boundary 3D points. [sent-6, score-0.247]
6 We experimentally prove the sufficiency of this set of boundary points and demonstrate the effectiveness of our approach on different developable and non-developable surfaces with a wide range of extensibility. [sent-7, score-0.43]
7 Introduction Monocular template-based 3D reconstruction and representation of nonrigid objects require models with the ability to assume a wide variety of shapes and to track complex motion. [sent-9, score-0.168]
8 The models must be able to recover a plausible deformed shape (or at least a set of discrete plausible deformed shapes) from noise-corrupted features while making the weakest possible assumptions about the observed shape. [sent-10, score-0.531]
9 On the one hand, previous approaches based on physical constraints address this problem in the case of isometric [6, 15], and conformal [9, 4] deformations. [sent-11, score-0.847]
10 On the other hand, statistical learning approaches [16, 13] have shown effectiveness only on isometric surfaces. [sent-12, score-0.34]
11 In this paper, we present a new physical-based approach for monocular template-based reconstruction of extensible and compressible surfaces with only one mechanical parameter. [sent-13, score-0.842]
12 As physical priors, our approach relies only on the Poisson ratio of the surface material assumed to be linear Hookean ∗This research has received funding from the EU’s FP7 through ERC grant 307483 FLEXABLE. [sent-14, score-0.566]
13 This Poisson parameter models the tendency of a material to be compressed in a transverse direction when it is stretched in the longitudinal direction [5]. [sent-19, score-0.475]
14 Our formulation is based on the principle that any extensible/compressible surface lies in the minimal stretching energy state subject to external applied constraints. [sent-20, score-0.731]
15 Accordingly, we formulate the reconstruction problem as being to estimate the shape that has minimal stretching energy given a set of boundary points, and is consistent with the measured image data. [sent-21, score-0.644]
16 The results are close to the real deformed surface as will be shown later in the experimental results section. [sent-22, score-0.401]
17 Related Work and Contribution Different types of constraints have been proposed and can be categorized as statistical or physical constraints. [sent-24, score-0.271]
18 Statistical constraints often model the deformation as a linear combination of basis vectors, which can be learned online either for human face reconstruction [11] or for generic shapes [15, 13, 18]. [sent-25, score-0.422]
19 Non-linear learning methods were applied in human tracking [14] and then extended for more generic surfaces [16]. [sent-26, score-0.243]
20 Physical constraints include spatial and temporal priors on the surface. [sent-27, score-0.152]
21 In [3] physical constraints are used as priors for a coarse-to-fine shape-basis statistical model. [sent-28, score-0.355]
22 A physical prior that has been studied is the isometry constraint [6], which requires that any surface geodesic distance is preserved after deformation. [sent-29, score-0.47]
23 This approach has proven its accuracy for paper-like surfaces and was recently extended for non-developable surfaces undergoing conformal deformation[9]. [sent-30, score-0.724]
24 More recently, [4] have studied the well-posedness of both isometric and conformal deformations. [sent-31, score-0.62]
25 Their applicability in real world deforming object remains limited (paper like surfaces, deforming balls under isotropic conditions). [sent-32, score-0.312]
26 Thus, the problem of monocular template-based 3D reconstruction of realistic deformations has not yet been tackled. [sent-33, score-0.431]
27 However, a SLAM (Simultaneous Localization And Mapping) method for elastic surfaces was attempted with fixed boundary conditions [1] and later extended to free boundary conditions [2]. [sent-34, score-0.564]
28 The surface is assumed to be homeomorphic to a disc. [sent-36, score-0.316]
29 It uses an FEM (Finite Element Method) to model the surface and approximate the deformation forces. [sent-38, score-0.365]
30 The surface and the deformation forces are both estimated using an EKF (Extended Kalman Filter). [sent-39, score-0.365]
31 It is important though to distinguish SLAM and monocular template-based reconstruction methods. [sent-40, score-0.315]
32 These assumptions simplify the problem since during tracking small deformations are sequentially integrated and added to the surface. [sent-42, score-0.147]
33 However, in monocular template-based approaches, the rate of deformation of the deformed surface may be very large and multiple solutions may appear. [sent-43, score-0.736]
34 This work goes beyond isometric and conformal deformations to consider elastic deformations for monocular template-based reconstruction. [sent-45, score-1.176]
35 First, we formalize the reconstruction problem of a generic surface in terms of the minimization of stretching energy. [sent-47, score-0.759]
36 From classic linear elasticity theory [5] we derive a patch-based formulation of the stretching energy of an elastic surface. [sent-48, score-0.793]
37 This formulation unifies geometrical and mechanical constrains in a single energy term. [sent-49, score-0.421]
38 Second, we show that we do not have global scale ambiguities and we prevent local scale ambiguities by imposing a set of fixed boundary 3D points. [sent-50, score-0.324]
39 We experimentally prove that this set of boundary condition sufficiently constrain the solution. [sent-51, score-0.166]
40 Third, we propose an iterative method to solve our extensible surface reconstruction problem. [sent-52, score-0.507]
41 Geometric Priors as Physical Constraints: Isometric and Conformal A smoothly deforming surface S homeomorphic to a disc, can be modelled as an embedding of a template Ω into R3. [sent-54, score-0.491]
42 It is described by a surface function ϕ of two variables (u, v) ∈ Ω : ϕ : Ω ⊂ R2 → R3 . [sent-55, score-0.219]
43 It is intuitively constrained by a set of point correspondences, represented by the 2D warp η, between the template and the input image. [sent-57, score-0.07]
44 The nature of the surface deformations are encoded in the differential prop- ×× erties of ϕ. [sent-58, score-0.335]
45 The Jacobian, denoted Jϕ, is a 3 2 matrix which measures and characterizes the lo,ca ils e ax t3en ×si o2n m mofa ttrhiex deformations. [sent-59, score-0.153]
46 Jϕ, gives the required tensor to distances on the deformed shape. [sent-61, score-0.237]
47 It is a 2 2 matrix which maps the loonc athl edid setafonrcmese dfr sohma Ωe. [sent-62, score-0.09]
48 aPr 2ev ×io 2u ms watroirxk whahisc hfoc muaspeds hone two types of mappings: isometric and conformal. [sent-64, score-0.296]
49 This property has been used as the main constraint in formalizing the problem of monocular template-based 3D reconstruction of isometric surfaces [6]. [sent-66, score-0.881]
50 This deformation assumption has been used as the main constraint in formalizing the problem of monocular template-based 3D reconstruction of conformal surfaces [9, 4]. [sent-70, score-1.055]
51 In our work, we link this geometric tensor to mechanical priors and then unify geometric and mechanical constraints in one equation, as will be seen in the next section. [sent-71, score-0.551]
52 Mechanical Priors as Physical Constraints: Our Stretching Energy Our idea is to model the material being deformed as made of some elastic material, and minimizing the deformation energy in trying to fit the surface to the data. [sent-73, score-0.898]
53 For isotropic materials, the deformation energy is computed in terms of the Young’s modulus E and the Poisson ratio ν [5]. [sent-74, score-0.448]
54 Background Consider a rod of some material with rest length L, which is longitudinally stretched with a force F. [sent-77, score-0.463]
55 Young’s modulus expresses the relationship of the force to the extension of the rod: F = E dL/L , (1) where dL/L is the relative extension of the rod, and E is Young’s modulus. [sent-78, score-0.183]
56 It is measured in Pascals (Newtons per square metre the square metres relating to the crosssection of the rod). [sent-79, score-0.049]
57 The Poisson ratio ν models the tendency of a material to become thinner when it is stretched. [sent-80, score-0.264]
58 To be exact, for a unit cube of material, – −ν =ΔΔyx=ΔΔxz Δx Δy Δz where , and are small changes in dimension when the cube is stretched (or compressed) in the X direction without constraints in the other directions, as shown in figure 2-a. [sent-82, score-0.358]
wordName wordTfidf (topN-words)
[('stretching', 0.34), ('conformal', 0.324), ('isometric', 0.296), ('surface', 0.219), ('monocular', 0.189), ('deformed', 0.182), ('elasticity', 0.178), ('surfaces', 0.165), ('extensible', 0.162), ('rod', 0.162), ('physical', 0.159), ('mechanical', 0.151), ('deformation', 0.146), ('poisson', 0.145), ('stretched', 0.138), ('elastic', 0.135), ('reconstruction', 0.126), ('deformations', 0.116), ('material', 0.116), ('deforming', 0.106), ('formalizing', 0.105), ('modulus', 0.105), ('energy', 0.1), ('homeomorphic', 0.097), ('young', 0.096), ('isometry', 0.092), ('priors', 0.084), ('unifies', 0.081), ('nicta', 0.078), ('boundary', 0.078), ('ambiguities', 0.077), ('cube', 0.076), ('constraints', 0.068), ('australian', 0.066), ('tendency', 0.065), ('slam', 0.065), ('compressed', 0.057), ('isotropic', 0.056), ('tensor', 0.055), ('sufficiency', 0.053), ('transverse', 0.053), ('adrien', 0.053), ('auvergne', 0.053), ('opi', 0.053), ('materials', 0.05), ('geometrical', 0.049), ('compressible', 0.049), ('ekf', 0.049), ('metres', 0.049), ('imposing', 0.048), ('force', 0.047), ('plausible', 0.046), ('bas', 0.046), ('athl', 0.046), ('longitudinal', 0.046), ('developable', 0.046), ('ttrhiex', 0.046), ('ron', 0.046), ('experimentally', 0.045), ('statistical', 0.044), ('prevent', 0.044), ('lat', 0.044), ('weakest', 0.044), ('balls', 0.044), ('dfr', 0.044), ('prove', 0.043), ('formalized', 0.042), ('thinner', 0.042), ('canberra', 0.042), ('unify', 0.042), ('shapes', 0.042), ('ratio', 0.041), ('formulation', 0.04), ('generic', 0.04), ('ils', 0.039), ('disc', 0.039), ('apr', 0.039), ('umr', 0.039), ('extended', 0.038), ('richard', 0.038), ('iot', 0.038), ('template', 0.038), ('xz', 0.037), ('mofa', 0.036), ('conditions', 0.035), ('formalize', 0.034), ('kalman', 0.034), ('jacobian', 0.033), ('australia', 0.032), ('france', 0.032), ('lies', 0.032), ('warp', 0.032), ('yx', 0.032), ('characterizes', 0.032), ('undergoing', 0.032), ('assumptions', 0.031), ('government', 0.031), ('expresses', 0.031), ('modelled', 0.031), ('funding', 0.031)]
simIndex simValue paperId paperTitle
same-paper 1 1.0000001 289 cvpr-2013-Monocular Template-Based 3D Reconstruction of Extensible Surfaces with Local Linear Elasticity
Author: Abed Malti, Richard Hartley, Adrien Bartoli, Jae-Hak Kim
Abstract: We propose a new approach for template-based extensible surface reconstruction from a single view. We extend the method of isometric surface reconstruction and more recent work on conformal surface reconstruction. Our approach relies on the minimization of a proposed stretching energy formalized with respect to the Poisson ratio parameter of the surface. We derive a patch-based formulation of this stretching energy by assuming local linear elasticity. This formulation unifies geometrical and mechanical constraints in a single energy term. We prevent local scale ambiguities by imposing a set of fixed boundary 3D points. We experimentally prove the sufficiency of this set of boundary points and demonstrate the effectiveness of our approach on different developable and non-developable surfaces with a wide range of extensibility.
2 0.26372975 423 cvpr-2013-Template-Based Isometric Deformable 3D Reconstruction with Sampling-Based Focal Length Self-Calibration
Author: Adrien Bartoli, Toby Collins
Abstract: It has been shown that a surface deforming isometrically can be reconstructed from a single image and a template 3D shape. Methods from the literature solve this problem efficiently. However, they all assume that the camera model is calibrated, which drastically limits their applicability. We propose (i) a general variational framework that applies to (calibrated and uncalibrated) general camera models and (ii) self-calibrating 3D reconstruction algorithms for the weak-perspective and full-perspective camera models. In the former case, our algorithm returns the normal field and camera ’s scale factor. In the latter case, our algorithm returns the normal field, depth and camera ’s focal length. Our algorithms are the first to achieve deformable 3D reconstruction including camera self-calibration. They apply to much more general setups than existing methods. Experimental results on simulated and real data show that our algorithms give results with the same level of accuracy as existing methods (which use the true focal length) on perspective images, and correctly find the normal field on affine images for which the existing methods fail.
3 0.17214312 44 cvpr-2013-Area Preserving Brain Mapping
Author: Zhengyu Su, Wei Zeng, Rui Shi, Yalin Wang, Jian Sun, Xianfeng Gu
Abstract: Brain mapping transforms the brain cortical surface to canonical planar domains, which plays a fundamental role in morphological study. Most existing brain mapping methods are based on angle preserving maps, which may introduce large area distortions. This work proposes an area preserving brain mapping method based on MongeBrenier theory. The brain mapping is intrinsic to the Riemannian metric, unique, and diffeomorphic. The computation is equivalent to convex energy minimization and power Voronoi diagram construction. Comparing to the existing approaches based on Monge-Kantorovich theory, the proposed one greatly reduces the complexity (from n2 unknowns to n ), and improves the simplicity and efficiency. Experimental results on caudate nucleus surface mapping and cortical surface mapping demonstrate the efficacy and efficiency of the proposed method. Conventional methods for caudate nucleus surface mapping may suffer from numerical instability; in contrast, current method produces diffeomorpic mappings stably. In the study of cortical sur- face classification for recognition of Alzheimer’s Disease, the proposed method outperforms some other morphometry features.
4 0.14033861 226 cvpr-2013-Intrinsic Characterization of Dynamic Surfaces
Author: Tony Tung, Takashi Matsuyama
Abstract: This paper presents a novel approach to characterize deformable surface using intrinsic property dynamics. 3D dynamic surfaces representing humans in motion can be obtained using multiple view stereo reconstruction methods or depth cameras. Nowadays these technologies have become capable to capture surface variations in real-time, and give details such as clothing wrinkles and deformations. Assuming repetitive patterns in the deformations, we propose to model complex surface variations using sets of linear dynamical systems (LDS) where observations across time are given by surface intrinsic properties such as local curvatures. We introduce an approach based on bags of dynamical systems, where each surface feature to be represented in the codebook is modeled by a set of LDS equipped with timing structure. Experiments are performed on datasets of real-world dynamical surfaces and show compelling results for description, classification and segmentation.
5 0.1246952 208 cvpr-2013-Hyperbolic Harmonic Mapping for Constrained Brain Surface Registration
Author: Rui Shi, Wei Zeng, Zhengyu Su, Hanna Damasio, Zhonglin Lu, Yalin Wang, Shing-Tung Yau, Xianfeng Gu
Abstract: Automatic computation of surface correspondence via harmonic map is an active research field in computer vision, computer graphics and computational geometry. It may help document and understand physical and biological phenomena and also has broad applications in biometrics, medical imaging and motion capture. Although numerous studies have been devoted to harmonic map research, limited progress has been made to compute a diffeomorphic harmonic map on general topology surfaces with landmark constraints. This work conquer this problem by changing the Riemannian metric on the target surface to a hyperbolic metric, so that the harmonic mapping is guaranteed to be a diffeomorphism under landmark constraints. The computational algorithms are based on the Ricci flow method and the method is general and robust. We apply our algorithm to study constrained human brain surface registration problem. Experimental results demonstrate that, by changing the Riemannian metric, the registrations are always diffeomorphic, and achieve relative high performance when evaluated with some popular cortical surface registration evaluation standards.
6 0.11123529 111 cvpr-2013-Dense Reconstruction Using 3D Object Shape Priors
7 0.10436491 113 cvpr-2013-Dense Variational Reconstruction of Non-rigid Surfaces from Monocular Video
8 0.10033174 443 cvpr-2013-Uncalibrated Photometric Stereo for Unknown Isotropic Reflectances
9 0.10027285 298 cvpr-2013-Multi-scale Curve Detection on Surfaces
10 0.099062361 465 cvpr-2013-What Object Motion Reveals about Shape with Unknown BRDF and Lighting
11 0.097571276 306 cvpr-2013-Non-rigid Structure from Motion with Diffusion Maps Prior
12 0.094648644 424 cvpr-2013-Templateless Quasi-rigid Shape Modeling with Implicit Loop-Closure
13 0.091750666 230 cvpr-2013-Joint 3D Scene Reconstruction and Class Segmentation
14 0.089002676 61 cvpr-2013-Beyond Point Clouds: Scene Understanding by Reasoning Geometry and Physics
15 0.088834114 286 cvpr-2013-Mirror Surface Reconstruction from a Single Image
17 0.085227996 97 cvpr-2013-Correspondence-Less Non-rigid Registration of Triangular Surface Meshes
18 0.075488597 303 cvpr-2013-Multi-view Photometric Stereo with Spatially Varying Isotropic Materials
19 0.07443919 231 cvpr-2013-Joint Detection, Tracking and Mapping by Semantic Bundle Adjustment
20 0.074215285 372 cvpr-2013-SLAM++: Simultaneous Localisation and Mapping at the Level of Objects
topicId topicWeight
[(0, 0.123), (1, 0.139), (2, -0.003), (3, 0.037), (4, 0.012), (5, -0.098), (6, -0.041), (7, -0.009), (8, 0.027), (9, -0.019), (10, -0.01), (11, -0.007), (12, -0.147), (13, -0.042), (14, 0.078), (15, -0.067), (16, 0.07), (17, 0.073), (18, 0.007), (19, 0.074), (20, -0.115), (21, -0.03), (22, 0.02), (23, 0.069), (24, -0.13), (25, 0.027), (26, -0.01), (27, 0.002), (28, -0.01), (29, -0.034), (30, 0.048), (31, -0.02), (32, 0.042), (33, -0.001), (34, 0.031), (35, 0.041), (36, 0.001), (37, -0.078), (38, 0.026), (39, -0.046), (40, 0.092), (41, 0.017), (42, -0.011), (43, 0.008), (44, 0.001), (45, 0.027), (46, 0.01), (47, -0.021), (48, 0.083), (49, -0.015)]
simIndex simValue paperId paperTitle
same-paper 1 0.96970719 289 cvpr-2013-Monocular Template-Based 3D Reconstruction of Extensible Surfaces with Local Linear Elasticity
Author: Abed Malti, Richard Hartley, Adrien Bartoli, Jae-Hak Kim
Abstract: We propose a new approach for template-based extensible surface reconstruction from a single view. We extend the method of isometric surface reconstruction and more recent work on conformal surface reconstruction. Our approach relies on the minimization of a proposed stretching energy formalized with respect to the Poisson ratio parameter of the surface. We derive a patch-based formulation of this stretching energy by assuming local linear elasticity. This formulation unifies geometrical and mechanical constraints in a single energy term. We prevent local scale ambiguities by imposing a set of fixed boundary 3D points. We experimentally prove the sufficiency of this set of boundary points and demonstrate the effectiveness of our approach on different developable and non-developable surfaces with a wide range of extensibility.
2 0.83099508 226 cvpr-2013-Intrinsic Characterization of Dynamic Surfaces
Author: Tony Tung, Takashi Matsuyama
Abstract: This paper presents a novel approach to characterize deformable surface using intrinsic property dynamics. 3D dynamic surfaces representing humans in motion can be obtained using multiple view stereo reconstruction methods or depth cameras. Nowadays these technologies have become capable to capture surface variations in real-time, and give details such as clothing wrinkles and deformations. Assuming repetitive patterns in the deformations, we propose to model complex surface variations using sets of linear dynamical systems (LDS) where observations across time are given by surface intrinsic properties such as local curvatures. We introduce an approach based on bags of dynamical systems, where each surface feature to be represented in the codebook is modeled by a set of LDS equipped with timing structure. Experiments are performed on datasets of real-world dynamical surfaces and show compelling results for description, classification and segmentation.
3 0.8276884 208 cvpr-2013-Hyperbolic Harmonic Mapping for Constrained Brain Surface Registration
Author: Rui Shi, Wei Zeng, Zhengyu Su, Hanna Damasio, Zhonglin Lu, Yalin Wang, Shing-Tung Yau, Xianfeng Gu
Abstract: Automatic computation of surface correspondence via harmonic map is an active research field in computer vision, computer graphics and computational geometry. It may help document and understand physical and biological phenomena and also has broad applications in biometrics, medical imaging and motion capture. Although numerous studies have been devoted to harmonic map research, limited progress has been made to compute a diffeomorphic harmonic map on general topology surfaces with landmark constraints. This work conquer this problem by changing the Riemannian metric on the target surface to a hyperbolic metric, so that the harmonic mapping is guaranteed to be a diffeomorphism under landmark constraints. The computational algorithms are based on the Ricci flow method and the method is general and robust. We apply our algorithm to study constrained human brain surface registration problem. Experimental results demonstrate that, by changing the Riemannian metric, the registrations are always diffeomorphic, and achieve relative high performance when evaluated with some popular cortical surface registration evaluation standards.
4 0.80519509 44 cvpr-2013-Area Preserving Brain Mapping
Author: Zhengyu Su, Wei Zeng, Rui Shi, Yalin Wang, Jian Sun, Xianfeng Gu
Abstract: Brain mapping transforms the brain cortical surface to canonical planar domains, which plays a fundamental role in morphological study. Most existing brain mapping methods are based on angle preserving maps, which may introduce large area distortions. This work proposes an area preserving brain mapping method based on MongeBrenier theory. The brain mapping is intrinsic to the Riemannian metric, unique, and diffeomorphic. The computation is equivalent to convex energy minimization and power Voronoi diagram construction. Comparing to the existing approaches based on Monge-Kantorovich theory, the proposed one greatly reduces the complexity (from n2 unknowns to n ), and improves the simplicity and efficiency. Experimental results on caudate nucleus surface mapping and cortical surface mapping demonstrate the efficacy and efficiency of the proposed method. Conventional methods for caudate nucleus surface mapping may suffer from numerical instability; in contrast, current method produces diffeomorpic mappings stably. In the study of cortical sur- face classification for recognition of Alzheimer’s Disease, the proposed method outperforms some other morphometry features.
5 0.7986263 298 cvpr-2013-Multi-scale Curve Detection on Surfaces
Author: Michael Kolomenkin, Ilan Shimshoni, Ayellet Tal
Abstract: This paper extends to surfaces the multi-scale approach of edge detection on images. The common practice for detecting curves on surfaces requires the user to first select the scale of the features, apply an appropriate smoothing, and detect the edges on the smoothed surface. This approach suffers from two drawbacks. First, it relies on a hidden assumption that all the features on the surface are of the same scale. Second, manual user intervention is required. In this paper, we propose a general framework for automatically detecting the optimal scale for each point on the surface. We smooth the surface at each point according to this optimal scale and run the curve detection algorithm on the resulting surface. Our multi-scale algorithm solves the two disadvantages of the single-scale approach mentioned above. We demonstrate how to realize our approach on two commonly-used special cases: ridges & valleys and relief edges. In each case, the optimal scale is found in accordance with the mathematical definition of the curve.
6 0.77528805 435 cvpr-2013-Towards Contactless, Low-Cost and Accurate 3D Fingerprint Identification
8 0.73975307 423 cvpr-2013-Template-Based Isometric Deformable 3D Reconstruction with Sampling-Based Focal Length Self-Calibration
9 0.7269361 97 cvpr-2013-Correspondence-Less Non-rigid Registration of Triangular Surface Meshes
10 0.63950878 286 cvpr-2013-Mirror Surface Reconstruction from a Single Image
11 0.59639782 465 cvpr-2013-What Object Motion Reveals about Shape with Unknown BRDF and Lighting
12 0.59381419 432 cvpr-2013-Three-Dimensional Bilateral Symmetry Plane Estimation in the Phase Domain
13 0.5868386 141 cvpr-2013-Efficient Computation of Shortest Path-Concavity for 3D Meshes
14 0.56020284 443 cvpr-2013-Uncalibrated Photometric Stereo for Unknown Isotropic Reflectances
15 0.55357808 52 cvpr-2013-Axially Symmetric 3D Pots Configuration System Using Axis of Symmetry and Break Curve
16 0.52797967 303 cvpr-2013-Multi-view Photometric Stereo with Spatially Varying Isotropic Materials
17 0.52687013 218 cvpr-2013-Improving the Visual Comprehension of Point Sets
18 0.50700378 90 cvpr-2013-Computing Diffeomorphic Paths for Large Motion Interpolation
19 0.49038306 327 cvpr-2013-Pattern-Driven Colorization of 3D Surfaces
20 0.48593467 354 cvpr-2013-Relative Volume Constraints for Single View 3D Reconstruction
topicId topicWeight
[(3, 0.324), (10, 0.102), (16, 0.019), (26, 0.033), (33, 0.244), (67, 0.041), (69, 0.042), (87, 0.111)]
simIndex simValue paperId paperTitle
same-paper 1 0.81370938 289 cvpr-2013-Monocular Template-Based 3D Reconstruction of Extensible Surfaces with Local Linear Elasticity
Author: Abed Malti, Richard Hartley, Adrien Bartoli, Jae-Hak Kim
Abstract: We propose a new approach for template-based extensible surface reconstruction from a single view. We extend the method of isometric surface reconstruction and more recent work on conformal surface reconstruction. Our approach relies on the minimization of a proposed stretching energy formalized with respect to the Poisson ratio parameter of the surface. We derive a patch-based formulation of this stretching energy by assuming local linear elasticity. This formulation unifies geometrical and mechanical constraints in a single energy term. We prevent local scale ambiguities by imposing a set of fixed boundary 3D points. We experimentally prove the sufficiency of this set of boundary points and demonstrate the effectiveness of our approach on different developable and non-developable surfaces with a wide range of extensibility.
2 0.74580491 9 cvpr-2013-A Fast Semidefinite Approach to Solving Binary Quadratic Problems
Author: Peng Wang, Chunhua Shen, Anton van_den_Hengel
Abstract: Many computer vision problems can be formulated as binary quadratic programs (BQPs). Two classic relaxation methods are widely used for solving BQPs, namely, spectral methods and semidefinite programming (SDP), each with their own advantages and disadvantages. Spectral relaxation is simple and easy to implement, but its bound is loose. Semidefinite relaxation has a tighter bound, but its computational complexity is high for large scale problems. We present a new SDP formulation for BQPs, with two desirable properties. First, it has a similar relaxation bound to conventional SDP formulations. Second, compared with conventional SDP methods, the new SDP formulation leads to a significantly more efficient and scalable dual optimization approach, which has the same degree of complexity as spectral methods. Extensive experiments on various applications including clustering, image segmentation, co-segmentation and registration demonstrate the usefulness of our SDP formulation for solving large-scale BQPs.
3 0.73716903 226 cvpr-2013-Intrinsic Characterization of Dynamic Surfaces
Author: Tony Tung, Takashi Matsuyama
Abstract: This paper presents a novel approach to characterize deformable surface using intrinsic property dynamics. 3D dynamic surfaces representing humans in motion can be obtained using multiple view stereo reconstruction methods or depth cameras. Nowadays these technologies have become capable to capture surface variations in real-time, and give details such as clothing wrinkles and deformations. Assuming repetitive patterns in the deformations, we propose to model complex surface variations using sets of linear dynamical systems (LDS) where observations across time are given by surface intrinsic properties such as local curvatures. We introduce an approach based on bags of dynamical systems, where each surface feature to be represented in the codebook is modeled by a set of LDS equipped with timing structure. Experiments are performed on datasets of real-world dynamical surfaces and show compelling results for description, classification and segmentation.
Author: Keyang Shi, Keze Wang, Jiangbo Lu, Liang Lin
Abstract: Driven by recent vision and graphics applications such as image segmentation and object recognition, assigning pixel-accurate saliency values to uniformly highlight foreground objects becomes increasingly critical. More often, such fine-grained saliency detection is also desired to have a fast runtime. Motivated by these, we propose a generic and fast computational framework called PISA Pixelwise Image Saliency Aggregating complementary saliency cues based on color and structure contrasts with spatial priors holistically. Overcoming the limitations of previous methods often using homogeneous superpixel-based and color contrast-only treatment, our PISA approach directly performs saliency modeling for each individual pixel and makes use of densely overlapping, feature-adaptive observations for saliency measure computation. We further impose a spatial prior term on each of the two contrast measures, which constrains pixels rendered salient to be compact and also centered in image domain. By fusing complementary contrast measures in such a pixelwise adaptive manner, the detection effectiveness is significantly boosted. Without requiring reliable region segmentation or post– relaxation, PISA exploits an efficient edge-aware image representation and filtering technique and produces spatially coherent yet detail-preserving saliency maps. Extensive experiments on three public datasets demonstrate PISA’s superior detection accuracy and competitive runtime speed over the state-of-the-arts approaches.
5 0.69626456 91 cvpr-2013-Consensus of k-NNs for Robust Neighborhood Selection on Graph-Based Manifolds
Author: Vittal Premachandran, Ramakrishna Kakarala
Abstract: Propagating similarity information along the data manifold requires careful selection of local neighborhood. Selecting a “good” neighborhood in an unsupervised setting, given an affinity graph, has been a difficult task. The most common way to select a local neighborhood has been to use the k-nearest neighborhood (k-NN) selection criterion. However, it has the tendency to include noisy edges. In this paper, we propose a way to select a robust neighborhood using the consensus of multiple rounds of k-NNs. We explain how using consensus information can give better control over neighborhood selection. We also explain in detail the problems with another recently proposed neighborhood selection criteria, i.e., Dominant Neighbors, and show that our method is immune to those problems. Finally, we show the results from experiments in which we compare our method to other neighborhood selection approaches. The results corroborate our claims that consensus ofk-NNs does indeed help in selecting more robust and stable localities.
6 0.69367868 225 cvpr-2013-Integrating Grammar and Segmentation for Human Pose Estimation
7 0.67080498 244 cvpr-2013-Large Displacement Optical Flow from Nearest Neighbor Fields
8 0.66741812 365 cvpr-2013-Robust Real-Time Tracking of Multiple Objects by Volumetric Mass Densities
9 0.66668236 44 cvpr-2013-Area Preserving Brain Mapping
10 0.66574168 222 cvpr-2013-Incorporating User Interaction and Topological Constraints within Contour Completion via Discrete Calculus
11 0.66370404 71 cvpr-2013-Boundary Cues for 3D Object Shape Recovery
12 0.66246939 298 cvpr-2013-Multi-scale Curve Detection on Surfaces
13 0.66213328 155 cvpr-2013-Exploiting the Power of Stereo Confidences
14 0.66188562 147 cvpr-2013-Ensemble Learning for Confidence Measures in Stereo Vision
15 0.66107547 242 cvpr-2013-Label Propagation from ImageNet to 3D Point Clouds
16 0.6610322 98 cvpr-2013-Cross-View Action Recognition via a Continuous Virtual Path
17 0.65958571 39 cvpr-2013-Alternating Decision Forests
18 0.65957648 19 cvpr-2013-A Minimum Error Vanishing Point Detection Approach for Uncalibrated Monocular Images of Man-Made Environments
19 0.6593678 188 cvpr-2013-Globally Consistent Multi-label Assignment on the Ray Space of 4D Light Fields
20 0.65902287 61 cvpr-2013-Beyond Point Clouds: Scene Understanding by Reasoning Geometry and Physics