13108nam 22008295 450 99619826190331620200701105551.03-319-10605-810.1007/978-3-319-10605-2(CKB)3710000000219458(SSID)ssj0001338667(PQKBManifestationID)11704395(PQKBTitleCode)TC0001338667(PQKBWorkID)11344391(PQKB)10086514(DE-He213)978-3-319-10605-2(MiAaPQ)EBC6296464(MiAaPQ)EBC5610295(Au-PeEL)EBL5610295(OCoLC)889237498(PPN)180626345(EXLCZ)99371000000021945820140814d2014 u| 0engurnn|008mamaatxtccrComputer Vision -- ECCV 2014[electronic resource] 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part II /edited by David Fleet, Tomas Pajdla, Bernt Schiele, Tinne Tuytelaars1st ed. 2014.Cham :Springer International Publishing :Imprint: Springer,2014.1 online resource (XXVIII, 854 p. 357 illus.) Image Processing, Computer Vision, Pattern Recognition, and Graphics ;8690Bibliographic Level Mode of Issuance: Monograph3-319-10604-X Includes bibliographical references and index.Intro -- Foreword -- Preface -- Organization -- Table of Contents -- Learning and Inference (continued) -- Coarse-to-Fine Auto-Encoder Networks (CFAN) for Real-Time Face Alignment -- 1 Introduction -- 2 Related Works -- 2.1 Local Models with Regression Fitting -- 2.2 Deep Models -- 3 Coarse-to-Fine Auto-Encoder Networks -- 3.1 Method Overview -- 3.2 Global SAN -- 3.3 Local SANs -- 3.4 Discussions -- 4 Implementation Details -- 5 Experiments -- 5.1 Datasets and Methods for Comparison -- 5.2 Investigation of Each SAN in CFAN -- 5.3 Comparison on XM2VTS Dataset -- 5.4 Comparison on LFPW Dataset -- 5.5 Comparison on Helen Dataset -- 6 Conclusions and Future Works -- References -- From Manifold to Manifold: Geometry-Aware Dimensionality Reduction for SPD Matrices -- 1 Introduction -- 2 Related Work -- 3 Riemannian Geometry of SPD Manifolds -- 4 Geometry-Aware Dimensionality Reduction -- 4.1 Optimization on Grassmann Manifolds -- 4.2 Designing the Affinity Matrix -- 4.3 Discussion in Relation to Region Covariance Descriptors -- 5 Empirical Evaluation -- 5.1 Material Categorization -- 5.2 Action Recognition from Motion Capture Data -- 5.3 Face Recognition -- 6 Conclusions and Future Work -- References -- Pose Machines: Articulated Pose Estimation via Inference Machines -- 1 Introduction -- 2 Related Work -- 3 Pose Inference Machines -- 3.1 Background -- 3.2 Incorporating a Hierarchy -- 3.3 Context Features -- 3.4 Training -- 3.5 Stacking -- 3.6 Inference -- 3.7 Implementation -- 4 Evaluation -- 5 Discussion -- References -- Poster Session 2 -- Piecewise-Planar StereoScan: Structure and Motion from Plane Primitives -- 1 Introduction -- 1.1 Related Work -- 2 Background -- 2.1 Energy-Based Multi-Model Fitting -- 2.2 Semi-dense Piecewise Planar Stereo Reconstruction -- 3 Overview of the Approach -- 3.1 Semi-dense PPR from a Single Stereo Pair.3.2 PPR from a Stereo Sequence -- 4 Relative Pose Estimation -- 4.1 Relative Pose from 3 Plane Correspondences -- 4.2 Relative Pose Estimation in Case Ni Has Rank 2 -- 4.3 Relative Pose Estimation in Case Ni Has Rank 1 -- 4.4 Robust Algorithm for Computing the Relative Pose -- 5 Discrete-Continuous Bundle Adjustment -- 6 Experimental Results -- 7 Conclusions -- References -- Nonrigid Surface Registration and Completion from RGBD Images -- 1 Introduction -- 2 Related Work -- 3 Nonrigid Surface Registration and Completion -- 3.1 Nonrigid Patch-Based Surface Model -- 3.2 Nonrigid Registration as Inference in a CRF -- 3.3 Incorporating New Patches -- 4 Experimental Results -- 4.1 Comparison with the Baselines -- 4.2 Missing Data and Occlusions -- 4.3 Incorporating New Patches -- 4.4 Video Editing -- 5 Conclusion -- References -- Unsupervised Dense Object Discovery, Detection, Tracking and Reconstruction -- 1 Introduction -- 2 Overview -- 3 Discovery and Detection -- 4 Tracking and Reconstruction -- 4.1 Tracking -- 4.2 Reconstruction -- 4.3 Tracking and Reconstruction Interaction -- 5 System Integration -- 6 Results and Discussion -- 7 Failure Cases and Future Work -- 8 Conclusions -- References -- Know Your Limits: Accuracy of Long Range Stereoscopic Object Measurements in Practice -- 1 Introduction -- 2 Related Work -- 3 Long Range Object Stereo: Algorithm Overview -- 3.1 Local Differential Matching (LDM) -- 3.2 Joint Matching and Segmentation (SEG) -- 3.3 Total Variation Stereo (TV) -- 3.4 Semi-Global Matching (SGM) -- 4 Evaluation -- 4.1 Dataset -- 4.2 Performance Measures -- 5 Results and Analysis -- 6 Conclusions -- References -- As-Rigid-As-Possible Stereo under Second Order Smoothness Priors -- 1 Introduction -- 1.1 Background -- 1.2 Contribution -- 2 As-Rigid-As-Possible Stereo -- 2.1 Second-Order Smoothness Priors.2.2 As-Rigid-As-Possible Smoothness and Data Cost -- 2.3 Overall Energy -- 2.4 Interactions between the Two Priors -- 2.5 Optimization -- 3 Implementation -- 3.1 Initialize -- 3.2 Optimize ED -- 3.3 Optimize ES -- 3.4 Post-process -- 4 Experiments -- 4.1 Comparisons of Results -- 4.2 Running Time -- 5 Conclusion -- References -- Real-Time Minimization of the Piecewise Smooth Mumford-Shah Functional -- 1 Introduction -- 1.1 The Mumford-Shah Problem -- 1.2 Related Work -- 1.3 Contribution -- 2 Proposed Finite-Difference Discretization -- 3 Minimization Algorithm -- 3.1 Algorithm for Convex Regularizers R -- 3.2 Proposed Algorithm for the MS-Energy -- 4 Experiments -- 4.1 Energy in One-Dimensional Case -- 4.2 Comparison with Convex Relaxation -- 4.3 Comparison with Ambrosio-Tortorelli -- 4.4 Comparison with L0-Smoothing -- 4.5 Real-Time Unsupervised Image Segmentation -- 4.6 Real-Time Video Cartooning -- 5 Conclusion -- References -- A MAP-Estimation Framework for Blind Deblurring Using High-Level Edge Priors -- 1 Introduction -- 2 Our Blind Deconvolution Approach -- 2.1 Data Term Edata(k, x|I) -- 2.2 Blur Kernel Prior Term Ekernel(k) -- 2.3 Image Prior Term Eimg(x|e) -- 2.4 Edge Prior Term Eedge(e|x) -- 3 Geometric Parsing Prior for Blind Deconvolution -- 4 MAP-Estimation Inference -- 4.1 Optimizing over the Kernel k -- 4.2 Optimizing over the Latent Image -- 4.3 Optimizing over the Edge-Related Variables e -- 4.4 Multi-resolution Inference -- 5 Experimental Results -- 6 Conclusions -- References -- Efficient Color Constancy with Local Surface Reflectance Statistics -- 1 Introduction -- 2 Color Constancy with Local Surface Reflectance Estimation -- 2.1 Surface Reflectance Estimation in Local Region -- 2.2 Illuminant Estimation -- 3 Experimental Results -- 3.1 Real-World Image Set -- 3.2 An Indoor Image Dataset in Laboratory.3.3 SFU Grey Ball Image Datasets -- 3.4 SFU HDR Dataset -- 4 Discussion and Conclusion -- 5 Appendix -- References -- A Contrast Enhancement Framework with JPEG Artifacts Suppression -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 Structure-Texture Decomposition -- 3.2 Reducing Artifacts in the Texture Layer -- 3.3 Layer Recomposition -- 4 Results -- 5 Discussion and Conclusion -- References -- Radial Bright Channel Prior for Single Image Vignetting Correction -- 1 Introduction -- 2 Related Work -- 3 Radial Bright Channel Prior -- 4 Vignetting Correction Using RBC Prior -- 4.1 Simple Estimation of the Vignetting Function -- 4.2 Model-Based Vignetting Estimation with Outlier Handling -- 4.3 Restoring the Vignetting-Free Image -- 5 Results -- 5.1 Synthetic Examples -- 5.2 Real Examples -- 5.3 Computation Time -- 6 Discussion and Future Work -- References -- Tubular Structure Filtering by Ranking Orientation Responses of Path Operators -- 1 Introduction -- 2 Related Works -- 2.1 Differential Filters -- 2.2 Non-linear Filters -- 3 General Strategy -- 4 Ranking Orientation Responses of Path Operators -- 4.1 Path Operators -- 4.2 RPO-Based Filtering -- 4.3 Orientation Space Sampling -- 4.4 Cone-Oriented Robust Path Opening -- 4.5 Pointwise Rank Filtering -- 4.6 RORPO: A Filter Based on Ranking Orientations Responses Path Operator -- 4.7 Suppressing Artifacts Generated by Limit Cases -- 5 Experiments and Results -- 5.1 ComparedMethods and Quality Scores -- 5.2 Synthetic Images -- 5.3 Real Images -- 6 Discussion and Conclusion -- References -- Optimization-Based Artifact Correction for Electron Microscopy Image Stacks -- 1 Introduction -- 2 Artifact Correction Algorithm -- 2.1 Problem Formulation -- 2.2 Coarse-to-Fine Procedure -- 2.3 Removing the Blocking Effects -- 2.4 Parallelization -- 3 Evaluation on Electron Microscopy Data.3.1 Image Quality Evaluation -- 3.2 Segmentation Accuracy Evaluation -- 4 Electron Microscopy Results -- 4.1 Comparison of NIQE Scores -- 4.2 Comparison of Segmentation Accuracy -- 5 Lighting Correction of Time-Lapse Photography -- 6 Discussion -- References -- Metric-Based Pairwise and Multiple Image Registration -- 1 Introduction -- 1.1 Desired Properties in an Objective Function -- 1.2 Past and Current Literature -- 2 Metric-BasedImage Registration -- 2.1 Image Representation and Pairwise Registration -- 2.2 Gradient Method for Optimization Over Γ -- 2.3 Distance in the Quotient Space -- 3 Experiments -- 3.1 Pairwise Image Registration -- 3.2 Registering Multiple Images -- 3.3 Image Classification -- 4 Conclusion -- References -- Canonical Correlation Analysis on Riemannian Manifolds and Its Applications -- 1 Introduction -- 2 Canonical Correlation in Euclidean Space -- 3 Mathematical Preliminaries -- 4 A Model for CCA on Riemannian Manifolds -- 5 Optimization Schemes -- 5.1 An Augmented Lagrangian Method -- 5.2 Extensions to the Product Riemannian Manifold -- 6 Experiments -- 6.1 CCA on SPD Manifolds -- 6.2 Synthetic Experiments -- 6.3 CCA for Multi-modal Risk Analysis -- 7 Conclusion -- References -- Scalable 6-DOF Localization on Mobile Devices -- 1 Introduction -- 2 Overall Approach -- 3 Local and Global Pose Estimation -- 3.1 Local Pose Tracking Using SLAM -- 3.2 Server-Based Global Localization -- 4 Aligning the Local Map Globally -- 5 Experimental Evaluation -- 5.1 Comparison of the Proposed Alignment Strategies -- 5.2 Accuracy, Efficiency and Scalability of Pose Estimation -- 6 Conclusion and Future Work -- References -- On Mean Pose and Variability of 3D Deformable Models -- 1 Introduction -- 2 Related Work -- 3 Mean Pose Inference Model -- 3.1 Shape Space Parameterization -- 3.2 Mean Pose -- 3.3 Generative Model.3.4 Expectation-Maximization Inference.The seven-volume set comprising LNCS volumes 8689-8695 constitutes the refereed proceedings of the 13th European Conference on Computer Vision, ECCV 2014, held in Zurich, Switzerland, in September 2014. The 363 revised papers presented were carefully reviewed and selected from 1444 submissions. The papers are organized in topical sections on tracking and activity recognition; recognition; learning and inference; structure from motion and feature matching; computational photography and low-level vision; vision; segmentation and saliency; context and 3D scenes; motion and 3D scene analysis; and poster sessions.Image Processing, Computer Vision, Pattern Recognition, and Graphics ;8690Optical data processingPattern recognitionArtificial intelligenceComputer graphicsImage Processing and Computer Visionhttps://scigraph.springernature.com/ontologies/product-market-codes/I22021Pattern Recognitionhttps://scigraph.springernature.com/ontologies/product-market-codes/I2203XArtificial Intelligencehttps://scigraph.springernature.com/ontologies/product-market-codes/I21000Computer Graphicshttps://scigraph.springernature.com/ontologies/product-market-codes/I22013Optical data processing.Pattern recognition.Artificial intelligence.Computer graphics.Image Processing and Computer Vision.Pattern Recognition.Artificial Intelligence.Computer Graphics.006.37Fleet Davidedthttp://id.loc.gov/vocabulary/relators/edtPajdla Tomasedthttp://id.loc.gov/vocabulary/relators/edtSchiele Berntedthttp://id.loc.gov/vocabulary/relators/edtTuytelaars Tinneedthttp://id.loc.gov/vocabulary/relators/edtMiAaPQMiAaPQMiAaPQBOOK996198261903316Computer Vision -- ECCV 20142587608UNISA