Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 15th Iberoamerican Congress on Pattern Recognition, CIARP 2010, ... (Lecture Notes in Computer Science, 6419)

✍ Scribed by Isabelle Bloch (editor), Roberto M. Cesar Jr. (editor)

• Publisher: Springer
• Year: 2010
• Tongue: English
• Leaves: 588
• Category: Library

✦ Synopsis

Pattern recognition is a central topic in contemporary computer sciences, with continuously evolving topics, challenges, and methods, including machine learning, content-based image retrieval, and model- and knowledge-based - proaches, just to name a few. The Iberoamerican Congress on Pattern Recog- tion (CIARP) has become established as a high-quality conference, highlighting the recent evolution of the domain. These proceedings include all papers presented during the 15th edition of this conference, held in Sao Paulo, Brazil, in November 2010. As was the case for previous conferences, CIARP 2010 attracted parti- pants from around the world with the aim of promoting and disseminating - going research on mathematical methods and computing techniques for pattern recognition, computer vision, image analysis, and speech recognition, as well as their applications in such diverse areas as robotics, health, entertainment, space exploration, telecommunications, data mining, document analysis,and natural language processing and recognition, to name only a few of them. Moreover, it provided a forum for scienti?c research, experience exchange, sharing new kno- edge and increasing cooperation between research groups in pattern recognition and related areas. It is important to underline that these conferences have contributed sign- icantly to the growth of national associations for pattern recognition in the Iberoamerican region, all of them as members of the International Association for Pattern Recognition (IAPR).

✦ Table of Contents

Title
Preface
Organization
Table of Contents
Invited Talks
Recent Advances in Kernel Machines
Design of Pattern Classifiers Using Optimum-Path Forest with Applications in Image Analysis
Vision-Based Control of Robot Motion
Soft Computing, f-Granulation and Pattern Recognition
Computational Illumination
Color, Shape and Texture
Color Texture Analysis and Classification: An Agent Approach Based on Partially Self-avoiding Deterministic Walks
Introduction
Walks on Color Images and Texture Signature
Experiments and Results
Conclusion
References
Characterizing 3D Shapes Using Fractal Dimension
Introduction
Fractal Dimension Based Approach
Evaluation of the Shape Descriptor
Compared Approaches
Results
Conclusions
References
Multiresolution Histogram Analysis for Color Reduction
Introduction
Preliminaries
The Algorithm
Experimental Results
Concluding Remarks
References
Graphs and Hypergraphs
Graph of Words Embedding for Molecular Structure-Activity Relationship Analysis
Introduction
Embedding of Molecular Graphs
Formal Definitions
Edge Attributes Handling
Remarks
Experiments
Databases
Reference Systems
Results
Conclusions
References
A Hypergraph Reduction Algorithm for Joint Segmentation and Classification of Satellite Image Content
Introduction
Hypergraph Reduction and Properties
Application: Joint Segmentation and Superpixels Classification
Experimental Results
Conclusions and Perspectives
References
A Static Video Summarization Method Based on Hierarchical Clustering
Introduction
Clustering by Minimum Spanning Tree Approach
Video Summarization Using Minimum Spanning Tree
Experiments
Conclusion and Further Works
References
Biomedical Imaging
Histopathological Image Classification Using Stain Component Features on a pLSA Model
Introduction
Latent Topics Annotation Model
pLSA model
Representing Images as Bag of Words
Image Annotation Based on the pLSA Model
Local Region Description Based on the Staining Component Analysis
Color Normalization
Local Region Selection
Local Region Description
Model Evaluation
Experiments and Results
Data Set
Results and Discussion
Conclusions
References
Modified Expectation Maximization Algorithm for MRI Segmentation
Introduction
Methodology
Mixture Model-Based Image Segmentation
ETM Algorithm
Results
Discussion
Conclusion
References
Generation of Synthetic Multifractal Realistic Surfaces Based on Natural Model and Lognormal Cascade: Application to MRI Classification
Introduction
Generation of the Natural Multifractal Model
Validation of the Model
Application: MRI Classification
The Proposed Method
Results and Discussion
Conclusions
References
Retrieval, Mining and Learning
Content-Based Emblem Retrieval Using Zernike Moments
Introduction
Previous Work on Zernike Moments
Zernike Moments as Image Descriptors
First Modification: Higher Order Weighting
Second Modification: PCA ZMs
Metrics for Comparing Zernike Moments
Results and Final Remarks
References
A New Algorithm for Training SVMs Using Approximate Minimal Enclosing Balls
Introduction
MEB Problems and Support Vector Machines
Approximate MEBs and Core Vector Machines
The New Algorithm for MEB-SVMs
Experiments and Conclusions
References
A Hybrid Face Recognition Approach Using GPUMLib
Introduction
GPU Machine Learning Library
Non-Negative Matrix Factorization
Multiple Back-Propagation
Face Recognition Based on a Hybrid Classifier
Image Pre-processing
Parts-Based Representation of the Images Yale Data Base
Building and Evaluating the Classifier
Results and Discussion
Conclusions and Future Work
References
Learning, Recognition and Clustering
Self-training for Handwritten Text Line Recognition
Introduction
Handwritten Text Line Recognition
Preprocessing
The HWR System
Self-training
Overview
Recognition Confidence
Confidence Thresholds
Experimental Evaluation
Conclusion
References
Improving the Dynamic Hierarchical Compact Clustering Algorithm by Using Feature Selection
Introduction
Feature Selection in Dynamic Hierarchical Compact Algorithm
F1-Travel Quality Measure
Experimental Results
Conclusions
References
Background Division, A Suitable Technique for Moving Object Detection
Introduction
Fresnel Reflection
Background Division
Results
Conclusions
References
Bayesian and Statistical Methods
Concept Formation Using Incremental Gaussian Mixture Models
Introduction
The Incremental Gaussian Mixture Model
Experimental Results
Conclusion
References
Integrating Phonological Knowledge in ASR Systems for Spanish Language
Introduction
Phonological Feature Extraction
Speech Recognizer Architecture
Experimental Evaluation
Database Description
Phonological Feature Extraction
Comparing Phonological and Acoustic Representation Spaces
Concluding Remarks
References
Inference of Restricted Stochastic Boolean GRN’s by Bayesian Error and Entropy Based Criteria
Introduction
Feature Selection
Yeast Cell-Cycle Stochastic Restricted Boolean Model
Experimental Results
Conclusion
References
Coding and Compression, Video, Tracking
Grid Smoothing: A Graph-Based Approach
Introduction
Graph-Based Image Representation
Optimisation-Based Approach to Grid Smoothing
General Framework
Convergence of the Cost Function
Simulations
Conclusions
References
A Quality Analysis on JPEG 2000 Compressed Leukocyte Images by Means of Segmentation Algorithms
Introduction
Materials and Methods
The Images
Compression with JPEG 2000 Codec
The Segmentation Algorithms and Distance Measures
Quantitative Measures
Results
Conclusions
References
Comments on Matrix-Based Secret Sharing Scheme for Images
Introduction
Description of $(2,n)$-Threshold Image Secret Sharing Scheme Proposed by Rey.M.D
Recovering the Original Image from a Single Share
Experimental Results
Conclusions
References
A Very Low Bit-Rate Minimalist Video Encoder Based on Matching Pursuits
Introduction
Related Work
Matching Pursuits
Optimized Orthogonal Matching Pursuits
K-SVD
Matching Pursuit Video Coding
Proposed Video Codec
Experimental Results
Conclusions and Future Work
References
An Unified Transition Detection Based on Bipartite Graph Matching Approach
Introduction
A Dissimilarity Measure
Our Method for Transition Detection
Experiments
Conclusion and Further Works
References
A New Dissimilarity Measure for Trajectories with Applications in Anomaly Detection
Introduction
(Dis)Similarity Measures for Trajectories
Proposed Approach
Experimental Results in Anomaly Detection
Conclusions
References
Modelling Postures of Human Movements
Introduction
System Overview
Selected Key-Frames
Modelling
Building Models
Construction of Database of Models
Shape Matching
Tests and Results
Conclusion
References
Detection and Tracking of Driver’s Hands in Real Time
Introduction
Technical Setup
Data Base Description
Detection and Tracking Methods
Global Threshold Detection (GDT)
Local Threshold Detection (LTD)
Local Threshold Detection with Tracking (LTD-T)
Experiments and Results
Conclusions and Future Work
References
Speech, Natural Language, Document
Speaker Verification in Noisy Environment Using Missing Feature Approach
Introduction
Mask Estimation
Cluster-Based Reconstruction
The Algorithm
Experiments and Results
Detection and Compensation of Unreliable Components
Speaker Verification Protocol
Results
Conclusions and Future Work
References
Fast $k$-NN Classifier for Documents Based on a Graph Structure
Introduction
Proposed Classifier
Preprocessing Phase
Classification Phase
Experimental Results
Conclusions
References
Comparative Analysis between Wavelets for the Identification of Pathological Voices
Introduction
Signal Pre-processing
Wavelet Packet Transform
SVM Classifier
Experiments and Results
Conclusions
References
Comparison of HMM and DTW for Isolated Word Recognition System of Punjabi Language
Introduction
System Model
End point Detection
LPC Feature Analysis
Vector Quantization
HMM and DTW for Isolated Word Recognition
Hidden Markov Model
Dynamic Time Warping
Experimental Performance
Performance Based on Punjabi Numerals
Performance Based on Randomly Selected Words
Conclusion
References
A Combination of Classifiers for the Pronominal Anaphora Resolution in Basque
Introduction
Pronominal Anaphora Resolution in Basque
Main Characteristics of Pronominal Anaphora in Basque
Determination of Feature Vectors
Generation of Training Instances
Experimental Setup
Learning Algorithms
Results for Single Classifiers
Experimental Results
Combination of Classifiers
Results Obtained
Conclusions and Future Work
References
Text Segmentation by Clustering Cohesion
Introduction
Related Work
ClustSeg: A Method for Text Segmentation
Searching for Topic Cohesion
Detecting the Topic Segment Boundaries
Detecting the Document Segment Boundaries
Evaluation
Conclusion
References
Multiple Clues for License Plate Detection and Recognition
Introduction
LPR Framework
License Plate Detection
Character Segmentation
Classification
Experimental Results
Conclusions and Perspectives
References
Image Filtering and Segmentation
Non Local Image Denoising Using Image Adapted Neighborhoods
Introduction
Non Local Means Denoising
Neighborhood Selection for Synthetic Images
PCA and NLM
NLM Using PCA Neighborhoods
Discussion, Conclusions and Future Work
References
A Quantitative Evaluation of Fixed-Pattern Noise Reduction Methods in Imaging Systems
Introduction
Striping Noise and Its Reduction
Nonuniformity and Its Correction
Quality Indexes
Results
Conclusions
References
Vessel Centerline Tracking in CTA and MRA Images Using Hough Transform
Introduction
Vessel Tracking Method
Initialization and Preprocessing
Vessel Detection by Hough Transform
Vessel Tangent Estimation
Bifurcations
Preliminary Results
Synthetic Images
Real Images
Discussion and Conclusions
References
Automatic Image Segmentation Optimized by Bilateral Filtering
Introduction
Methodology
Study Scenes
Bilateral Filtering
Segmentation Process with Self-calibration Framework
Segmentation Process with Definiens Developper Software
Results
Conclusions
References
Pansharpening of High and Medium Resolution Satellite Images Using Bilateral Filtering
Introduction
Bilateral Filter
Image Fusion with Bilateral Filter
Data and Methodology
Study Area
Quality Determination - Comparison with Other Techniques
Results
Conclusion
References
Color Image Segmentation by Means of a Similarity Function
Introduction
Previous Works
Description of the Method
Calculation of Average Hue
Results and Discussion
Conclusions
References
Image Segmentation Using Quadtree-Based Similarity Graph and Normalized Cut
Introduction
Related Works
Graph Representation
Pixel Affinity Graph
Quadtree-Based Similarity Graph
Algorithm Overview
Experiments
Conclusion
References
A Genetic-Algorithm-Based Fusion System Optimization for 3D Image Interpretation
Introduction
Fusion System for 3D Image Interpretation
The Studied Fusion System
Importance of the Extraction Step
Genetic Algorithms
Principle of Genetic Algorithm
Application to Image Processing Parameter Adjustments
Illustration on 3D Tomography Image Interpretation
Conclusions
References
Feature Extraction, Shape, Texture, Geometry and Morphology
Quaternion Atomic Function Wavelet for Applications in Image Processing
Introduction
Atomic Functions
Quaternion Algebra
Quaternion Atomic Function
Quaternion Atomic Wavelet Function
Results
Conclusion
References
A Complex Network-Based Approach for Texture Analysis
Introduction
Texture as a Pixel Network
Network Characterization
Evaluation
Experiments
Results and Discussion
Conclusion
References
Enhancing Gabor Wavelets Using Volumetric Fractal Dimension
Introduction
Gabor Wavelets
Feature Extraction on the Gabor Space
Volumetric Fractal Dimension
Experimentation and Evaluation
Conclusions
References
Comparison of Shape Descriptors for Mice Behavior Recognition
Introduction
Shape Descriptors
Fourier Descriptors
Curvature
Zernike Moments
Chain Code
Experimental Results
Comparison of Methods
Conclusion and Future Works
References
Ridge Linking Using an Adaptive Oriented Mask Applied to Plant Root Images with Thin Structures
Introduction
Methodology
Adaptive Directional Mask
Ridge Direction and Geodesic Distance
Experiments
Conclusions
References
Parameter Estimation for Ridge Detection in Images with Thin Structures
Introduction
Ridge Detection
Proposed Detectors
The Frangi Filter
Steerable Filter
Parameter Estimation
Experimental Results
Conclusions
References
Experimental Comparison of Orthogonal Moments as Feature Extraction Methods for Character Recognition
Introduction
Orthogonal Moments
Experimental Results
Comparison of Feature Extraction Methods
Discriminating Capability of Orthogonal Moments
Effect of Moment Order in Classification Results
Conclusion
References
Face Segmentation and Recognition, Biometry
Improving Face Segmentation in Thermograms Using Image Signatures
Introduction
Overview of Face Segmentation in Thermal Infrared ImagES
Proposed Method
Experimental Results
Datasets
Experimental Results and Discussion
Conclusion
References
On Combining Local DCT with Preprocessing Sequence for Face Recognition under Varying Lighting Conditions
Introduction
Illumination Compensation Using DCT in Log Domain
The New Photometric Normalisation Method
Experimental Evaluation
Classifier Fusion
Conclusions
References
On Improving Dissimilarity-Based Classifications Using a Statistical Similarity Measure
Introduction
Related Work
Experimental Results
Conclusions
References
A Rotation Invariant Face Recognition Method Based on Complex Network
Introduction
Complex Networks
Measures
Face Recognition Using Complex Networks
Feature Vector Extraction
Face Recognition Experiments
Analysis of the Method Parameters
Comparing with Other Methods
Conclusion
References
Illumination Invariant Face Image Representation Using Quaternions
Introduction
Quaternion Theory
Image Applications
Image Quaternion Representation in the Frequency Domain
Experimental Evaluation
Verification Experiment
Identification Experiment
Conclusions and Future Works
References
A Contrario Detection of False Matches in Iris Recognition
Introduction
The A Contrario Framework
Iris Template Matching
Partitioning the Iris Template
Other Considerations
Performance Evaluation
Results
Conclusions
References
Statistical Approaches, Learning, Classification, Mining
A Functional Density-Based Nonparametric Approach for Statistical Calibration
Introduction
Functional Density-Based Nonparametric Regression
Definition of DBNR in a General Setting
Specification in the Gaussian Case
A Simulation Study
Data Generation
Simulation Results
A Study of Tecator Dataset
Conclusions
References
Feature Extraction Using Circular Statistics Applied to Volcano Monitoring
Introduction
Volcano Monitoring
Seismic Signals
Mathematical Background
The Analytic Signal
Circular Statistics
Wavelet Transform
Results
Conclusions and Future Work
References
Improving the Accuracy of the Optimum-Path Forest Supervised Classifier for Large Datasets
Introduction
Optimum-Path Forest Classifier
Optimum-Path Forest Classifier with Outliers Detection: OPF-OD
Experiments
Datasets Used
Sizes of $Z_1$ and $Z_2$
Results
Conclusions
References
Assessment of a Modified Version of the EM Algorithm for Remote Sensing Data Classification
Introduction
The Standard EM for GMM
E-Step
M-Step
Convergence
Modifications to the Standard EM Algorithm
Initialization
Probabilities Estimation
Classification
Results
Conclusion
References
A Sequential Minimal Optimization Algorithm for the All-Distances Support Vector Machine
Introduction
Background
SMO Algorithm for the AD-SVM
Stopping Criteria
Selection Strategy
Optimization Step
F$_t$'s Update
Algorithm Structure
Experiments and Conclusions
References
Aerosol Optical Thickness Retrieval from Satellite Observation Using Support Vector Regression
Introduction
Methodology
Data Sets
AERONET-MODIS Combination
Support Vector Regression
Experimental Results
Conclusion and Future Works
References
An Overproduce-and-Choose Strategy to Create Classifier Ensembles with Tuned SVM Parameters Applied to Real-World Fault Diagnosis
Introduction
Model-Free Approach to Fault Diagnosis
Feature-Based Classifier Ensembles
Feature Selection
Best Selected Feature Subsets Ensemble Creation
Experimental Results
Studied Classification Approaches
Misalignment Predictor Overproduced SVMs
$5\times2$ Cross-Validation Estimation Results
Conclusions and Future Work
References
Multi-Objective Semi-Supervised Feature Selection and Model Selection Based on Pearson’s Correlation Coefficient
Introduction
Semi-Supervised Learning
Semi-Supervised Feature Selection
Multi-Objective Learning
Multi-Objective Semi-Supervised Feature Selection (MOBJ-SSFS)
Results and Discussions
Conclusions
References
Introducing ROC Curves as Error Measure Functions: A New Approach to Train ANN-Based Biomedical Data Classifiers
Introduction
ROC Az Based Error Measures
ANNs Trained with Simulated Annealing
ROC Az Error with Simulated Annealing
The Biomedtk Framework
Experimental Setup
Results and Discussion
Conclusions
References
Partition Selection Approach for Hierarchical Clustering Based on Clustering Ensemble
Introduction
Partition Selection Based on Cluster Ensemble
Computational Complexity Analysis
Experimental Results
Conclusions
References
The Imbalanced Problem in Morphological Galaxy Classification
Introduction
Related Work
Imbalanced Data Sets
Automated Galaxy Classification
The Methods
Naive Bayes Classifier
C4.5
Radial Basis Function Networks
Random Forest
Support Vector Machines
SMOTE
Experimental Results
Conclusions
References
Exploiting Contextual Information for Image Re-ranking
Introduction
A Re-ranking Method Based on Contextual Information
Contextual Information Representation
Exploiting Contextual Information for Image Re-ranking
Experimental Evaluation
Impact of Parameters
Experimental Results
Conclusions
References
Assessing the Role of Spatial Relations for the Object Recognition Task
Introduction
Visual Description of Images
Computing Visual Similarity
Our Proposed Spatial Descriptor
The Spatial Descriptor
The Spatial Relationship Similarity
Experiments
Matching Strategy
Experiment Description and Results
Conclusions
References
Automatic Representation of Semantic Abstraction of Geographical Data by Means of Classification
Introduction
Definitions
Ontologies
Data-Representation Ontology (DRO)
Structure of the DRO
Principals Steps for Automatic Generation of DRO
Semantic Abstraction of Geographical Data
Classification of Data Representation Nodes (DRN)
Dissimilarity Measure
Experimental Results
Conclusions
References
Author Index