Computer Vision -- ACCV 2014 : 12th Asian Conference on Computer Vision, Singapore, Singapore, November 1-5, 2014, Revised Selected Papers, Part I / / edited by Daniel Cremers, Ian Reid, Hideo Saito, Ming-Hsuan Yang

• Pubblicazione: Cham : , : Springer International Publishing : , : Imprint : Springer, , 2015
• Edizione: 1st ed. 2015.
• Descrizione fisica: 1 online resource (XX, 727 p. 293 illus.)
• Disciplina: 004
• Soggetto topico: Optical data processing; Pattern recognition; Artificial intelligence; Health informatics; Application software; Information storage and retrieval; Image Processing and Computer Vision; Pattern Recognition; Artificial Intelligence; Health Informatics; Information Systems Applications (incl. Internet); Information Storage and Retrieval
• Persona (resp. second.): CremersDaniel; ReidIan; SaitoHideo; YangMing-Hsuan
• Note generali: Bibliographic Level Mode of Issuance: Monograph
• Nota di contenuto: Intro -- Preface -- Organization -- Contents - Part I -- Recognition -- Deep Representations to Model User `Likes' -- 1 Introduction -- 2 Semantic Feature Representation -- 3 Proposed Approach -- 3.1 User-Specific Feature Selection -- 3.2 Learning Deep Bimodal Feature Representation -- 4 Experiments -- 4.1 Dataset -- 4.2 Results and Analysis -- 5 Conclusion -- References -- Submodular Reranking with Multiple Feature Modalities for Image Retrieval -- 1 Introduction -- 2 Related Works -- 3 Submodular Reranking -- 3.1 Preliminaries -- 3.2 Information Gain with Graphical Models -- 3.3 Relative Ranking Consistency -- 3.4 Optimization -- 4 Experiments -- 4.1 Experimental Setting -- 4.2 Results Comparisons -- 4.3 Parameter Analysis -- 5 Conclusions -- References -- Accurate Scene Text Recognition Based on Recurrent Neural Network -- 1 Introduction -- 2 Related Work -- 2.1 Scene Text Recognition -- 2.2 Recurrent Neural Network -- 3 Feature Preparation -- 4 Recurrent Neural Network Construction -- 5 Word Scoring with Lexicon -- 6 Experiments and Discussion -- 6.1 System Details -- 6.2 Experiments on ICDAR and SVT Datasets -- 6.3 Discussion -- 7 Conclusion -- References -- Massive City-Scale Surface Condition Analysis Using Ground and Aerial Imagery -- 1 Introduction -- 2 Related Work -- 3 Large-Scale Estimation of Land Surface Condition -- 3.1 Debris Detection -- 3.2 Projection of Debris Probabilities onto the Ground -- 3.3 Integration Using Gaussian Process Regression -- 4 Experimental Results -- 4.1 Our Data -- 4.2 Ablative Analysis -- 4.3 Extensions to City-Scale Vegetation Estimation -- 5 Conclusion -- References -- Can Visual Recognition Benefit from Auxiliary Information in Training? -- 1 Introduction -- 2 Related Work -- 3 A Latent Space Model: Addressing Missing-View-in-Test-Data -- 4 DCCA: Formulation -- 5 DCCA: Solution -- 6 Experiments. ; 6.1 Compared Methods and Datasets -- 6.2 NYU-Depth-V1-Indoor Scene Dataset -- 6.3 RGBD Object Dataset -- 6.4 Multi-Spectral Scene Dataset -- 6.5 Discussion -- 7 Conclusions -- References -- Low Rank Representation on Grassmann Manifolds -- 1 Introduction -- 2 LRR on Grassmann Manifold -- 2.1 Low-Rank Representation (LRR) -- 2.2 LRR on Grassmann Manifolds -- 3 Solution to LRR on Grassmann Manifold -- 4 Experiments -- 4.1 Data Preparation and Experiment Settings -- 4.2 MNIST Handwritten Digits Clustering -- 4.3 Dynamic Texture Clustering -- 5 Conclusion and Future Work -- References -- Poster Session 1 -- Learning Detectors Quickly with Stationary Statistics -- 1 Introduction -- 2 Background -- 2.1 Linear Discriminant Analysis with Stationarity -- 2.2 Correlation Filters -- 2.3 Related Work -- 3 Fast Estimation of the Toeplitz Covariance -- 4 From Toeplitz to Circulant -- 5 Multi-channel, Two-Dimensional Signals -- 5.1 Toeplitz Covariance Matrix -- 5.2 Circulant Covariance Matrix -- 5.3 From Toeplitz to Circulant -- 6 Solving Toeplitz Systems -- 6.1 Direct Methods -- 6.2 Iterative Methods -- 6.3 An Effective Heuristic -- 7 Empirical Study -- 7.1 Detection Performance -- 7.2 Time and Memory -- 8 Conclusion -- References -- Age Estimation Based on Complexity-Aware Features -- 1 Introduction -- 2 Related Work -- 3 Features Used for Face Description -- 3.1 Gradient Features -- 3.2 Gabor Filters -- 4 Learning the Features Using Realadaboost with Complexity Penalty Terms -- 5 Experiments -- 5.1 Experiments Setup -- 5.2 Experimental Results -- 5.3 Analysis -- 6 Conclusion -- References -- Efficient On-the-fly Category Retrieval Using ConvNets and GPUs -- 1 Introduction -- 2 Evaluating Large-Scale Object Category Retrieval -- 2.1 Evaluation Protocol -- 2.2 Experimental Scenarios -- 3 Retrieval Performance over Image Representations -- 3.1 Results and Analysis. ; 3.2 Implementation Details -- 4 On-the-fly Architecture -- 4.1 System Performance -- 5 Conclusion -- References -- A Latent Clothing Attribute Approach for Human Pose Estimation -- 1 Introduction -- 2 Related Work -- 3 HPE with Latent Clothing Attributes -- 3.1 Feature Representation -- 3.2 Structured Learning with Latent SVM -- 3.3 Inference -- 4 Experiments -- 4.1 Datasets -- 4.2 Baselines and Metric -- 4.3 Results -- 5 Conclusion -- References -- NOKMeans: Non-Orthogonal K-means Hashing -- 1 Introduction -- 2 Background -- 2.1 Notation -- 2.2 Related Work -- 3 Proposed Algorithm -- 3.1 Formulation -- 3.2 Computational Complexity -- 3.3 Discussion -- 4 Experiments -- 4.1 Performance Measurements -- 4.2 Parameter Selection -- 4.3 Results -- 5 Conclusion -- References -- Visual Vocabulary with a Semantic Twist -- 1 Introduction -- 2 Semantic Vocabulary for Object Retrieval -- 2.1 Semantic Vocabulary -- 2.2 Matching Patches: SemanticSIFT -- 2.3 Product Vocabulary and Speedup -- 2.4 Reduction in Memory Requirements -- 2.5 Accounting for Segmentation Uncertainty -- 2.6 Challenges -- 3 Fast Semantic Segmentation -- 3.1 Soft Segments -- 3.2 Labelling -- 3.3 Segmentation Results -- 4 Experimental Setup and Retrieval Results -- 4.1 Evaluation, Datasets and Baseline -- 4.2 Retrieval Results -- 5 Conclusions and Future Work -- References -- Context Based Re-ranking for Object Retrieval -- 1 Introduction -- 2 Related Work -- 3 Context Generation -- 3.1 Random Space Partition -- 3.2 Expansion of Contexts -- 4 Context Based Re-ranking -- 4.1 Computing the Context Factor for Re-ranking -- 5 Experimental Results -- 5.1 Experimental Setup -- 5.2 Evaluation of Random Space Partition -- 5.3 Effects of Context Expansion -- 5.4 Evaluation of Context Re-ranking -- 5.5 Comparison to State-of-the-Art -- 6 Conclusion -- References. ; Adaptive Structural Model for Video Based Pedestrian Detection -- 1 Introduction -- 2 Related Work -- 3 Preliminaries -- 3.1 Bilinear DPM -- 3.2 Cost-Flow Based Data Association -- 4 Adaptive Structural Model -- 4.1 Adaptive DPM and Regularization -- 4.2 Loss Function -- 4.3 Adaptive Optimization -- 5 Experiment -- 5.1 Different Methods for Video Based Detection -- 5.2 The Convergence -- 5.3 Comparisons with State-of-the-Art Methods -- 6 Conclusion -- References -- Fusion of Auxiliary Imaging Information for Robust, Scalable and Fast 3D Reconstruction -- 1 Introduction -- 2 Related Work -- 3 A Global Approach by Iteratively Optimizing Potential Inliers -- 3.1 Step 1: Pre-processing -- 3.2 Step 2: Robust Iterative Rotation Estimation -- 3.3 Step 3: Robust Iterative Scene Reconstruction -- 4 Experiments -- 4.1 Comparison Methods and Comparison Criteria -- 4.2 Results and Analysis -- 5 Conclusion -- References -- What Visual Attributes Characterize an Object Class? -- 1 Introduction -- 2 Related Work -- 3 Unsupervised Visual Attributes Learning -- 3.1 Visual Prototypes Generating -- 3.2 Visual Attributes Learning -- 3.3 Visual Attributes Ranking -- 4 Experiments -- 4.1 Datasets -- 4.2 Experiment Settings -- 4.3 Attribute Learning Results -- 4.4 Object Categorization on AwA -- 4.5 Object Categorization on PASCAL VOC 2007 -- 5 Conclusion -- References -- Accurate Object Detection with Location Relaxation and Regionlets Re-localization -- 1 Introduction -- 2 Our Approach -- 2.1 Bottom-Up Object Proposal -- 2.2 Top-Down Supervised Object Search -- 2.3 Regionlets Object Re-localization -- 3 Experiments -- 3.1 Location Relaxation Search -- 3.2 Regionlets Re-localization -- 3.3 Run-Time Speed -- 4 Conclusions -- References -- Unsupervised Feature Learning for RGB-D Image Classification -- 1 Introduction -- 2 Related Work. ; 2.1 Work Related to Deep Neural Networks for Image Classification -- 2.2 Work Related to RGB-D Image Classification -- 3 Deep R2ICA Framework -- 3.1 Data Preprocessing -- 3.2 Filter Learning and Feature Encoding -- 3.3 Spatial Pooling and Normalization -- 3.4 Implementation Details -- 4 Experiments -- 4.1 RGB-D Object Recognition Benchmark [15] -- 4.2 2D3D Object Dataset -- 4.3 NYU Depth V1 Indoor Scene Benchmark -- 5 Conclusion -- References -- Non-maximum Suppression for Object Detection by Passing Messages Between Windows -- 1 Introduction -- 2 Related Work -- 3 A Message-Passing Approach for NMS -- 3.1 Affinity Propagation: Binary Formulation and Inference -- 3.2 Adapting Affinity Propagation for NMS -- 3.3 Structured Learning for Affinity Propagation -- 4 Experiments on Object Class Detection -- 4.1 Implementation Details -- 4.2 Results -- 5 Experiments on Generic Object Detection -- 5.1 Implementation Details -- 5.2 Results -- 6 Discussion -- References -- Stable Radial Distortion Calibration by Polynomial Matrix Inequalities Programming -- 1 Introduction -- 2 Camera Radial Distortion -- 2.1 Extrapolation Issues of Radial Distortion Calibration -- 3 Polynomials and PMI Programming -- 3.1 Polynomials and Polynomial Matrices -- 3.2 Polynomials Positive on Finite Intervals -- 3.3 Polynomial Matrix Inequalities -- 4 Shape Optimization for Radial Distortion Calibration -- 4.1 Unconstrained Radial Distortion Calibration -- 4.2 Barrel Distortion and the Polynomial Model -- 4.3 Pincushion Distortion and the Division Model -- 4.4 Zero-Crossing Problem of the Rational Model -- 4.5 Shape Optimization in Camera Calibration Procedure -- 5 Experiments -- 6 Conclusion -- References -- Pedestrian Verification for Multi-Camera Detection -- 1 Introduction -- 2 Related Work -- 3 Base Detector -- 3.1 Low-Level Detection -- 3.2 Mid-Level Aggregation. ; 4 Pedestrian Verification.
• Sommario/riassunto: The five-volume set LNCS 9003--9007 constitutes the thoroughly refereed post-conference proceedings of the 12th Asian Conference on Computer Vision, ACCV 2014, held in Singapore, Singapore, in November 2014. The total of 227 contributions presented in these volumes was carefully reviewed and selected from 814 submissions. The papers are organized in topical sections on recognition; 3D vision; low-level vision and features; segmentation; face and gesture, tracking; stereo, physics, video and events; and poster sessions 1-3.
• Titolo autorizzato: Computer Vision -- ACCV 2014
• ISBN: 3-319-16865-7
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Lingua di pubblicazione: Inglese
• Record Nr.: 9910483222303321
• Lo trovi qui: Univ. Federico II
• Serie: Image Processing, Computer Vision, Pattern Recognition, and Graphics ; ; 9003