Pattern Recognition Applications and Methods: International Conference, ICPRAM 2013 Barcelona, Spain, February 15-18, 2013 Revised Selected Papers (Advances ... Intelligent Systems and Computing Book 318)

This book contains the extended and revised versions of a set of selected papers from the 2nd International Conference on Pattern Recognition (ICPRAM 2013), held in Barcelona, Spain, from 15 to 18 February, 2013. ICPRAM was organized by the Institute for Systems and Technologies of Information, Control and Communication (INSTICC) and was held in cooperation with the Association for the Advancement of Artificial Intelligence (AAAI). The hallmark of this conference was to encourage theory and practice to meet in a single venue. The focus of the book is on contributions describing applications of Pattern Recognition techniques to real-world problems, interdisciplinary research, experimental and/or theoretical studies yielding new insights that advance Pattern Recognition methods. Front Matter....Pages i-xv Front Matter....Pages 1-1 A Two-Part Approach to Face Recognition: Generalized Hough Transform and Image Descriptors....Pages 3-16 Improved Boosting Performance by Explicit Handling of Ambiguous Positive Examples....Pages 17-37 Discriminative Dimensionality Reduction for the Visualization of Classifiers....Pages 39-56 Online Unsupervised Neural-Gas Learning Method for Infinite Data Streams....Pages 57-70 The Path Kernel: A Novel Kernel for Sequential Data....Pages 71-84 A MAP Approach to Evidence Accumulation Clustering....Pages 85-100 Feature Discretization with Relevance and Mutual Information Criteria....Pages 101-118 Multiclass Semi-supervised Learning on Graphs Using Ginzburg-Landau Functional Minimization....Pages 119-135 Probabilistic Discriminative Dimensionality Reduction for Pose-Based Action Recognition....Pages 137-152 Graph Cut Based Segmentation of Predefined Shapes: Applications to Biological Imaging....Pages 153-170 Artificial Neural Network Modeling of Relative Humidity and Air Temperature Spatial and Temporal Distributions Over Complex Terrains....Pages 171-187 Front Matter....Pages 189-189 Beyond SIFT for Image Categorization by Bag-of-Scenes Analysis....Pages 191-207 Unsupervised Learning of Semantics of Object Detections for Scene Categorization....Pages 209-224 Supervised Learning of Anatomical Structures Using Demographic and Anthropometric Information....Pages 225-240 Wikifying Novel Words to Mixtures of Wikipedia Senses by Structured Sparse Coding....Pages 241-255 Measuring Linearity of Planar Curves....Pages 257-271 Video Segmentation Framework Based on Multi-kernel Representations and Feature Relevance Analysis for Object Classification....Pages 273-283 Quality-Based Super Resolution for Degraded Iris Recognition....Pages 285-300 Generic Biometry Algorithm Based on Signal Morphology Information: Application in the Electrocardiogram Signal....Pages 301-310 Erratum to: A MAP Approach to Evidence Accumulation Clustering....Pages E1-E2 Back Matter....Pages 311-312 A MAP Approach to Evidence Accumulation Clustering1 Introduction; 2 Probabilistic Model; 3 Optimization Algorithm; 3.1 Computation of a Search Direction; 3.2 Computation of an Optimal Step Size; 3.3 Complexity; 4 Related Work; 5 Experiments and Results; 5.1 UCI and Synthetic Data; 5.2 Text Data; 6 Conclusions; References; Feature Discretization with Relevance and Mutual Information Criteria; 1 Introduction; 1.1 Our Contribution; 2 Background; 2.1 Entropy and Mutual Information; 2.2 Feature Discretization; 2.3 Unsupervised Discretization; 2.4 Supervised Discretization; 3 Proposed Methods Online Unsupervised Neural-Gas Learning Method for Infinite Data Streams1 Introduction; 2 Related Work; 3 Proposed Algorithm (AING); 3.1 General Behaviour; 3.2 AING Distance Threshold; 3.3 AING Merging Process; 4 Experimental Evaluation; 4.1 Experiments on Synthetic Data; 4.2 Experiments on Real Datasets; 5 Conclusions and Future Work; References; The Path Kernel: A Novel Kernel for Sequential Data; 1 Introduction; 2 Kernels and Sequences; 2.1 Sequence Similarity Measures; 3 The Path Kernel; 3.1 Efficient Computation; 3.2 Ground Kernel Choice; 4 Experiments; 5 Conclusions; References Preface; Organization; Contents; Part I Theory and Methods; A Two-Part Approach to Face Recognition: Generalized Hough Transform and Image Descriptors; 1 Introduction; 2 Method; 2.1 Modified GHT; 2.2 Gradient Distance Descriptor; 3 Results and Discussion; 4 Conclusions; References; Improved Boosting Performance by Explicit Handling of Ambiguous Positive Examples; 1 Introduction; 1.1 Relation to Bootstrapping Methods; 1.2 Contributions; 2 Relation to Previous Work; 3 Boosting Theory; 3.1 Convex-Loss Boosting Algorithms; 3.2 Robust Boosting Algorithms; 4 A Two-Pass Exclusion Extension 3.1 Relevance-Based LBG3.2 Mutual Information Discretization; 4 Experimental Evaluation; 4.1 Comparison Between Our Approaches; 4.2 Comparison with Existing Methods; 5 Conclusions; References; Multiclass Semi-supervised Learning on Graphs Using Ginzburg-Landau Functional Minimization; 1 Introduction; 2 Data Segmentation with the Ginzburg-Landau Model; 2.1 Application of Diffuse Interface Models to Graphs; 3 Multiclass Extension; 3.1 Generalized Difference Function; 3.2 Computational Algorithm; 4 Results; 4.1 Synthetic Data; 4.2 Image Segmentation; 4.3 Benchmark Sets; 5 Conclusions 4.1 Inverted Cascade5 Experiments; 6 Results; 6.1 Comparison of Boosting Algorithms; 6.2 Bootstrapping Methods in Relation to Outlier Exclusion; 7 Discussion and Future Work; 8 Conclusions; References; Discriminative Dimensionality Reduction for the Visualization of Classifiers; 1 Introduction; 2 Supervised Visualization Based on the Fisher Information; 2.1 Computation of the Class Probabilities; 2.2 Approximation of Minimum Path Integrals; 3 Training a Discriminative Visualization Mapping; 4 Visualization of Classifiers; 5 Conclusions; References