Pattern Recognition
33rd DAGM Symposium, Frankfurt/Main, Germany, August 31 - September 2, 2011, Proceedings
Published on 2. September 2011
XII, 473 pages
978-3-642-23123-0 (ISBN)
Description
This book constitutes the refereed proceedings of the 33rd Symposium of the German Association for Pattern Recognition, DAGM 2011, held in Frankfurt/Main, Germany, in August/September 2011. The 20 revised full papers and 22 revised poster papers were carefully reviewed and selected from 98 submissions. The papers are organized in topical sections on object recognition, adverse vision conditions challenge, shape and matching, segmentation and early vision, robot vision, machine learning, and motion. The volume also includes the young researcher's forum, a section where a carefully jury-selected ensemble of young researchers present their Master thesis work.
More details
Other editions
Additional editions
Rudolf Mester | Michael Felsberg
Pattern Recognition
33rd DAGM Symposium, Frankfurt/Main, Germany, August 31 - September 2, 2011, Proceedings
Book
08/2011
1st Edition
Springer
€53.49
Shipment within 7-9 days
Content
- Title
- Preface
- Organization
- Table of Contents
- Poster Session 1
- A Bayesian Approach for Scene Interpretation with Integrated Hierarchical Structure
- Introduction
- Concept for Conditional Bayesian Network
- Realization Regarding Facade Image Interpretation
- Hierarchical Segmentation
- Features of Regions
- Classification of Regions
- Learning Probabilities of Hierarchy
- Conditional Bayesian Network
- Experiments
- Setup Up of Our Evaluation and Used Data
- Results
- Adaptation of the Concept
- Conclusions
- References
- Multi-view Active Appearance Models for the X-Ray Based Analysis of Avian Bipedal Locomotion
- Introduction and Related Work
- Active Appearance Models
- Training Step
- Model Fitting
- Application on Locomotion Data
- Multi-view Model
- Experiments and Results
- Resolving Anatomical Ambiguities
- Reduction of the Amount of Training Data
- Comparison to Manual Landmark Tracking
- Conclusions and Further Work
- References
- A Fully Implicit Framework for Sobolev Active Contours and Surfaces
- Introduction
- The Classical $L^2$ Framework
- Sobolev Spaces on Implicit Surfaces
- Sobolev Spaces
- Computing Sobolev Gradients
- Numerical Treatment
- Discussion of the Experiments
- Conclusion
- References
- Implicit Scene Context for Object Segmentation and Classification
- Introduction
- Contribution
- CRF Classification Framework
- Implicit Scene Context (ISC)
- Context Potential within CRF
- Experiments
- Classification of Objects in Different Scenes
- Parameter Assessment with Simulated Data
- Conclusions and Future Work
- References
- An Estimation Theoretical Approach to Ambrosio-Tortorelli Image Segmentation
- Introduction
- Related Work
- Bayesian Estimation Theory in a Nutshell
- AT Segmentation as an Estimation Problem
- A Probabilistic Interpretation of the AT Energy
- Numerical Issues
- Experimental Results
- Stochastic vs. Deterministic AT Segmentation
- Piecewise Smooth Images and the Crack-Tip Problem
- Summary
- References
- Combined Head Localization and Head Pose Estimation for Video-Based Advanced Driver Assistance Systems
- Introduction
- Previous Work
- Proposed System
- Methodology
- Training Procedure
- Head Search Domain
- Calculation and Normalization of Confidence Values
- Head Localization and Pose Estimation
- Experiments
- Experimental Setup
- Head Localization Performance
- Head Pose Estimation Performance
- Conclusion
- References
- Towards Cross-Modal Comparison of Human Motion Data
- Introduction
- Sensor Modalities
- Positional Motion Data
- Inertial Motion Data
- Feature Representations
- Virtual Sensors
- Local Acclerations
- Directions Relative to Acceleration
- Directions Relative to Gravity
- Cross-modal Comparison
- Class Confusion
- F-measure
- Conclusions
- References
- Indoor Calibration Using Segment Chains
- Introduction
- Segments Extraction
- Segments Merging
- Segment Chains Matching
- Dynamic Programing Formulation
- Camera Calibration
- Experiments
- Conclusion
- References
- Multilinear Model Estimation with $L^2$-Regularization
- Introduction
- Energy Minimization Formulation
- Introducing $L^2$-Regularization
- Evaluation for Synthetic Data
- Real World Applications
- Conclusion
- References
- Optimization of Quadrature Filters Based on the Numerical Integration of Improper Integrals
- Introduction
- Filter Optimization: A Continuous Reformulation
- Numerical Quadrature of Improper Integrals
- The One-Dimensional Case
- The Two-Dimensional Case
- The Three and Four Dimensional Cases
- Numerical Experiments
- Continuous versus Discrete Fourier Space Optimization
- A Two-Dimensional Example
- Computational Costs
- Conclusions
- References
- Real Time Head Pose Estimation from Consumer Depth Cameras
- Introduction
- Related Work
- Discriminative Random Regression Forests for Head Pose Estimation
- Training
- Head Pose Estimation
- Data Acquisition and Labeling
- Experiments
- Conclusions
- References
- Putting MAP Back on the Map
- Introduction
- The Image Labeling Problem
- Image Denoising
- Learning and Prediction
- Related Work
- Optimal Prediction
- Learning
- Maximum Likelihood and Maximum Pseudo-likelihood
- Maximum Margin
- Insights on Statistic Matching
- Experiments
- Synthetic Data
- Image Denoising
- Conclusions
- References
- Pose-Consistent 3D Shape Segmentation Based on a Quantum Mechanical Feature Descriptor
- Introduction
- Related Work
- Contribution
- The Wave Kernel Signature - A Quantum Mechanical Feature Descriptor
- Quantum Particles on Surfaces
- The Wave Kernel Signature
- Comparison of WKS and HKS
- Learning Pose-Invariant Shape Segmentation
- Experimental Results
- Computational Details
- Segmentation Results
- Robustness
- Shape Retrieval
- Conclusion
- References
- Object Recognition
- Multiple Instance Boosting for Face Recognition in Videos
- Introduction
- MIL - Boosting
- Gradient Boosting
- Loss Functions
- Solving MIL with Gradient Boosting
- Benchmark Datasets
- Face Recognition from Videos
- Conclusion
- References
- SHOG - Spherical HOG Descriptors for Rotation Invariant 3D Object Detection
- Introduction
- SHOG - Spherical Histograms of Oriented Gradients
- Experiments
- Conclusions
- References
- Pick Your Neighborhood - Improving Labels and Neighborhood Structure for Label Propagation
- Introduction
- Related Work
- Improving Neighborhood Structure for SSL
- Datasets and Representation
- Evaluation of Metric Learning for Label Propagation
- Selection of Training Data Based on Domain Knowledge
- Conclusion
- References
- Poster Session 2
- People Tracking Algorithm for Human Height Mounted Cameras
- Introduction
- Related Work
- Approach
- Background Subtraction
- Particle Filter
- Appearance Model
- Processing of Entries and Leavings
- Experiments
- Conclusion
- References
- Multi-person Localization and Track Assignment in Overlapping Camera Views
- Introduction
- Previous Work
- Multi-person Track Assignment
- Likelihood Formulation
- Likelihood Optimization
- Experiments
- Setup
- Evaluation
- Conclusion
- References
- Relaxed Exponential Kernels for Unsupervised Learning
- Introduction
- The Local Gaussian Assumption
- A Convolution Kernel for the Space X
- Isometric Embedding in a Hilbert Space H
- Kernels for Probability Distributions
- The Case of Gaussian Densities
- A Close Look at dJ and dB
- A Metric for Symmetric PD Matrices
- Relaxed Kernels for the Augmented Space X
- Related Work
- Experimental Results
- Conclusion
- References
- Using Landmarks as a Deformation Prior for Hybrid Image Registration
- Introduction
- Background
- Gaussian Processes
- Gaussian Process Regression
- Hybrid Registration Using a Landmark Prior
- Landmark Registration
- Combined Landmark and Image Registration
- A Note on the Implementation
- Results
- Conclusion
- References
- Improving Denoising Algorithms via a Multi-scale Meta-procedure
- Introduction
- Down-Scaling Has a Denoising Effect
- How to Denoise Lower Frequencies
- Multi-scale Denoising
- Experimental Evaluation and Results
- Conclusion
- References
- Adverse Vision Conditions Challenge
- Assessment of Visibility Quality in Adverse Weather and Illumination Conditions
- Introduction
- Principles of Keypoint-Based Fragment Matching
- From Image Matching to Visibility Conditions
- Experimental Result
- Future Works and Conclusions
- References
- Robust Point Matching in HDRI through Estimation of Illumination Distribution
- Introduction
- Illumination Distribution with GMM
- Illumination Distribution in HDRI
- Shadow and Highlight Parts in HDRI
- Point Matching in HDR Images and Results
- Conclusions
- References
- Illumination-Robust Dense Optical Flow Using Census Signatures
- Introduction
- Dense Optical Flow in Real Scenes
- Related Work
- Dense Motion Estimation
- A General Numerical Solution Scheme
- Illumination-Invariant Motion Estimation
- Illumination Robustness with Local Compensation of the Mean
- Illumination Model
- Illumination-Invariant Census Based Residuals
- Results
- Evaluation of a Synthetic Scene from the Middlebury Data Set
- Real Scenes
- Conclusion
- References
- Efficient Stereo and Optical Flow with Robust Similarity Measures
- Introduction
- Related Work
- Background
- Method
- Robust Similarity Measures
- Optical Flow
- Results
- Robust Stereo Matching
- Efficient Dense Optical Flow
- Efficiency
- Conclusion
- References
- Shape and Matching
- Shape- and Pose-Invariant Correspondences Using Probabilistic Geodesic Surface Embedding
- Introduction
- Probabilistic Geodesic Surface Embedding
- Probabilistic Model
- Inference
- Results
- Data
- Evaluation
- Conclusions
- References
- Dense 3D Reconstruction of Symmetric Scenes from a Single Image
- Introduction
- Previous Work
- Symmetric Scene Geometry
- Estimating the Symmetry Plane
- Dense Depth Estimation
- Experiments
- Conclusion
- References
- Fingerprints for Machines - Characterization and Optical Identification of Grinding Imprints
- Introduction
- Related Works
- Feature-Based Profile Characterization
- Detection
- Description
- Comparison
- False Positive Profile Matches
- Experiments
- Evaluation of False Positive Detections
- Corrosion
- Optical Fingerprint Comparison
- Comparison to SIFT Matching
- Summary and Conclusion
- References
- Efficient and Robust Alignment of Unsynchronized Video Sequences
- Introduction
- Problem Formulation
- An IR Approach to Video Synchronization
- Indexing, Structure, and Retrieval
- Spatio-temporal Coherence
- Spatial Alignment and Synchro Refinement
- Efficiency
- Results
- Conclusions
- References
- Segmentation and Early Vision
- Time-Consistent Foreground Segmentation of Dynamic Content from Color and Depth Video
- Introduction
- Data Acquisition and Preprocessing
- Separation of Foreground and Background
- Foreground Extraction Based on Directed Histograms
- Grab-Cut Segmentation in a Video Volume
- Experimental Results
- Conclusion
- References
- Channel Coding for Joint Colour and Depth Segmentation
- Introduction
- Related Work
- Methods and Materials
- The Microsoft Kinect
- Fused Feature Vectors
- Superparamagnetic Clustering
- Channel Coding
- Experiments
- Data Sets
- Performance Evaluation
- Tested Methods
- Results
- Conclusions
- References
- Simultaneous Interpolation and Deconvolution Model for the 3-D Reconstruction of Cell Images
- Introduction
- Joint Variational Interpolation and Deconvolution
- Deconvolution Model
- Interpolation Model
- Joint Model
- Numerical Aspects
- Experimental Evaluation
- Conclusion and Future Work
- References
- Steerable Deconvolution Feature Detection as an Inverse Problem
- Introduction
- Steerable Deconvolution
- The Idea
- Regularization
- Probabilistic Interpretation and Optimization
- Complex Formulation
- Fourier Representation of the Steerable Convolution
- Regularizer in Fourier Representation
- A Combined Edge/Ridge Kernel
- Experiments
- Conclusion
- References
- Robot Vision
- Probabilistic Object Models for Pose Estimation in 2D Images
- Introduction
- Object Model
- General Form
- Use of Edge Segments
- Scene Observations
- Pose Estimation
- Evaluation
- Experimental Setup
- Rotating Object
- Cluttered Scenes
- Conclusions
- References
- Fusion of Audio- and Visual Cues for Real-Life Emotional Human Robot Interaction
- Introduction
- Emotions in Natural Interaction
- Bi-modal Emotion Recognition
- Visual Facial Expression Recognition
- Emotion Recognition from Speech
- Probabilistic Decision Level Fusion
- Evaluation on Real-Life Data
- Conclusion and Outlook
- References
- Machine Learning
- Training of Sparsely Connected MLPs
- Introduction
- Sparse Generative Models for SMLP Initialization
- Non-Negative Matrix Factorization with Sparseness Constraints
- Extension for Fast Inference of Sparse Code Words
- Sparsely Connected MLPs
- Architecture
- Training Algorithm
- Computational Complexity of Classification
- Experiments on Handwritten Digits
- Comparison with Alternative Approaches
- Conclusions
- References
- Minimizing Calibration Time for Brain Reading
- Introduction
- Scenario
- Methods
- Evaluation
- Conclusion
- References
- Agnostic Domain Adaptation
- Introduction
- Prior Work
- Problem Setup
- Generalization Bound
- Algorithm
- Experiments
- Conclusion
- References
- Motion
- Will the Pedestrian Cross? Probabilistic Path Prediction Based on Learned Motion Features
- Introduction
- Previous Work
- Proposed System
- Experiments
- Conclusion
- References
- Simultaneous Reconstruction and Tracking of Non-planar Templates
- Introduction
- Related Work
- Method
- Regularization
- Evaluation
- Quantitative Evaluation
- Qualitative Evaluation
- Discussion and Future Work
- Conclusion
- References
- Multi-target Tracking in Crowded Scenes
- Introduction
- Related Work
- Multi-target Tracking Based on Dense Detection-Scores
- Generation of Tracklets
- Multi-target Tracking from Tracklets
- Experiments
- Conclusions
- References
- Efficient and Robust Shape Matching for Model Based Human Motion Capture
- Introduction
- Related Work
- Tracking System
- Motion Capture with String Matching
- Linear Assignment Problem
- String Representation of Shapes
- The Wagner-Fischer Algorithm
- Experiments
- Computational Time
- ICP vs. Proposed Method
- Conclusions
- References
- Young Researcher's Forum
- Image Comparison on the Base of a Combinatorial Matching Algorithm
- Introduction
- Interest Points and Features
- Matching Algorithm
- Reformulation as Markov Random Field
- Experiments
- Conclusion
- References
- Large Displacement Optical Flow for Volumetric Image Sequences
- Introduction
- Variational Model
- Descriptor Matching
- Symmetrical Optical Flow
- Results
- Synthetic Image Sequence
- Zebrafish Embryo Sequence
- Runtime
- References
- Visual Motion Capturing for Kinematic Model Estimation of a Humanoid Robot
- Introduction
- Visual Motion Capture System
- Efficient Rigid Body Detection
- Kinematic Model Estimation and Joint Angle Calculation
- Ball Joint Model Estimation
- Hinge Joint Model Estimation
- Joint Angle Calculation
- Error Evaluation
- Conclusion
- Future Work
- References
- Object Recognition System Guided by Gaze of the User with a Wearable Eye Tracker
- Introduction
- Proposed Application
- Real-Time Gaze Based Object Recognition
- Eye Tracking Method
- Basic Object Recognition Method
- Fixation Guided Object Recognition
- Gaze-Based Ground Truth Processing
- Gaze Detection Based on Recognition Results
- Experiments and Results
- Conclusion
- References
- Spectral Clustering of ROIs for Object Discovery
- Introduction and Related Work
- Spectral Clustering Techniques
- Nonlinear Component Analysis
- Normalized Cuts Spectral Clustering
- Object Detection and Categorization of ROIs
- Experimental Results
- Conclusions
- References
- Robust Classification and Semi-supervised Object Localization with Gaussian Processes
- Introduction and Related Work
- Object Localization with Hyperparameter Optimization in a Gaussian Processes Framework
- Experimental Results
- Conclusion and Future Work
- References
- Color Image Segmentation Based on an Iterative Graph Cut Algorithm Using Time-of-Flight Cameras
- Introduction
- Related Work
- Color Image Segmentation
- Data Acquisition
- Trimap Generation
- Color Clustering
- Segmentation by Energy Minimization
- Experimental Results
- Segmentation Performance
- Summary and Conclusion
- References
- Application of Multi-modal Features for Terrain Classification on a Mobile System
- Introduction
- Hardware Setup
- Related Work
- Multi-modal Features
- Terrain Classification with a Markov Random Field
- Experiments
- Conclusion
- References
- Author Index
