Computer Vision - ACCV 2010
10th Asian Conference on Computer Vision, Queenstown, New Zealand, November 8-12, 2010, Revised Selected Papers, Part II
Published on 28. February 2011
XXI, 726 pages
978-3-642-19309-5 (ISBN)
Description
The four-volume set LNCS 6492-6495 constitutes the thoroughly refereed post-proceedings of the 10th Asian Conference on Computer Vision, ACCV 2009, held in Queenstown, New Zealand in November 2010.
All together the four volumes present 206 revised papers selected from a total of 739 Submissions. All current issues in computer vision are addressed ranging from algorithms that attempt to automatically understand the content of images, optical methods coupled with computational techniques that enhance and improve images,
and capturing and analyzing the world's geometry while preparing the higher level image and shape understanding. Novel geometry techniques, statistical learning methods, and modern algebraic procedures are dealt with as well.
More details
Other editions
Additional editions
Ron Kimmel | Reinhard Klette | Akihiro Sugimoto
Computer Vision - ACCV 2010
10th Asian Conference on Computer Vision, Queenstown, New Zealand, November 8-12, 2010, Revised Selected Papers, Part II
Book
03/2011
1st Edition
Springer
€106.99
Shipment within 7-9 days
Content
- Title Page
- Preface
- Organization
- Table of Contents
- Posters on Day 1 of ACCV 2010
- Generic Object Class Detection Using Boosted Configurations of Oriented Edges
- Introduction
- Related Work
- Object Class Representation
- Edge Features
- Efficient Computation of Feature Values
- Learning
- Learning the Weak Classifiers
- Learning the Strong Classifier
- Learning the Cascade
- Detection
- Sliding Window
- Estimating the Aspect Ratio
- Experiments and Dataset
- Results
- Discussion
- Future Work
- Conclusions
- References
- Unsupervised Feature Selection for Salient Object Detection
- Introduction
- Graph Laplacian and Mixing Rate of Markov Process
- Feature Graphs and Irrelevant Features
- Irrelevant Feature Identification
- Experiments
- Conclusion
- References
- MRF Labeling for Multi-view Range Image Integration
- Introduction
- Related Work
- MRF Network Construction for Integration
- MAP-MRF Labeling for Integration
- Energy Minimisation Using Loopy Belief Propagation
- Experimental Results and Performance Analysis
- Conclusion and Future Work
- References
- Wave Interference for Pattern Description
- Introduction
- Related Work
- Properties of Interference Description
- Methods
- Background
- Interference Description
- Multi-frequency Analysis
- Descriptor Comparison
- Experiments and Results
- Verification of Gestalt Properties
- Application to Shape Matching and Retrieval
- Application to Pattern Recognition
- Discussion and Conclusions
- References
- Colour Dynamic Photometric Stereo for Textured Surfaces
- Introduction and Related Works
- Motivation
- Problem Formulation
- Proposed Solution
- Energy Functional
- Optimisation
- Experiments and Discussion
- Datasets
- Results
- Conclusion
- References
- Multi-Target Tracking by Learning Class-Specific and Instance-Specific Cues
- Introduction
- Overview of the Proposed Framework
- Detection and Association
- Tracking by Particle Filtering
- Global Pose Estimators
- Observation Model
- Dynamical Model
- Tracking Separated-Target
- Tracking Target-Group
- Experimental Results
- Conclusions
- References
- Modeling Complex Scenes for Accurate Moving Objects Segmentation
- Introduction
- General Model for Moving Object Segmentation
- Modeling the Complex Scenes
- Color Model
- Locality Model
- Temporal Coherence Model
- Experiments and Results
- Conclusions
- References
- Online Learning for PLSA-Based Visual Recognition
- Introduction
- PLSA-Based Visual Recognition
- Proposed Algorithm
- On-line Codebook Adaptation for PLSA
- On-Line EM for PLSA Learning
- Experiments
- Conclusions
- References
- Emphasizing 3D Structure Visually Using Coded Projection from Multiple Projectors
- Introduction
- Emphasizing 3D Structure Using Coded Projection
- Billboard Model for Space Representation
- Generating Projection Patterns for 3D Emphasis
- Representation of Negative Intensity
- Experiments
- Real Projector Experiments
- Accuracy Evaluation
- Constellation of Multiple Projectors
- Number of Projectors
- Conclusion
- References
- Object Class Segmentation Using Reliable Regions
- Introduction
- Related Work
- Proposed Approach
- Obtaining Multiple Segmentations
- Features from Regions
- Clustering
- Discarding Unreliable Segments
- Object Class Segmentation
- Experimental Results
- Comparison to Previous Work
- Conclusion
- References
- Specular Surface Recovery from Reflections of a Planar Pattern Undergoing an Unknown Pure Translation
- Introduction
- Translating Planar Pattern
- Recovering the Unknown Translation
- Specular Surface Recovery
- Experimental Analysis
- Synthetic Experiments
- Real Experiments
- Conclusions
- References
- Medical Image Segmentation Based on Novel Local Order Energy
- Introduction
- The Proposed Method
- Global-Homogenous and Local-Homogenous Methods
- Local Order Energy
- Level Set Formulation and Energy Minimization
- Experiment Results
- Conclusions and Future Works
- References
- Geometries on Spaces of Treelike Shapes
- Introduction
- Related Work
- Geometry of Treespace
- Representation of Trees
- The Space of Trees as a Quotient - A Singular Treespace
- Metrics on Treespace
- From Planar Trees to Spatial Trees
- Comparison of the QED and TED Metrics
- Computation and Complexity
- QED Implementation
- Experimental Results
- Synthetic Planar Trees of Depth 3
- Results in 3D: Pulmonary Airway Trees
- Conclusions and Future Work
- References
- Human Pose Estimation Using Exemplars and Part Based Refinement
- Introduction
- Overview of Our Framework
- Example Based Initialization
- Part Based Refinement
- Part Based Model
- Image Likelihood
- Iterative Pose Inference
- Experiments
- Experiment Data
- Experiments of Examplar Based Initialization
- Experiments of Pose Estimation
- Conclusion
- References
- Full-Resolution Depth Map Estimation from an Aliased Plenoptic Light Field
- Introduction
- Prior Work and Contributions
- Light Field Representation and Methodology Overview
- Virtual View Reconstructions
- Depth Estimation via Energy Minimization
- Obtaining Correspondences in Subimages and Views
- Brief Discussion on Sampling and Aliasing in Plenoptic Cameras
- Matching Sampled Views and Mosaicing
- Energy Terms for Full-Resolution Depth Maps
- Mosaiced Boundary Condition Consistency
- Regularization of the Depth Map
- Numerical Implementation
- Experiments
- Conclusions
- References
- Indoor Scene Classification Using Combined 3D and Gist Features
- Introduction
- Related Work
- The 3D Indoor Database
- Learning a Holistic Scene Representation
- Evaluation of Room Type Categorization
- Conclusion
- References
- Closed-Form Solutions to Minimal Absolute Pose Problems with Known Vertical Direction
- Introduction
- Problem Formulation
- Absolute Pose of a Calibrated Camera with Known Up Direction
- Absolute Pose of a Camera with Unknown Focal Length and Radial Distortion and Known Camera Up Direction
- Experiments
- Synthetic Data Set
- Real Data
- Conclusion
- References
- Level Set with Embedded Conditional Random Fields and Shape Priors for Segmentation of Overlapping Objects
- Introduction
- Level Set Methods with Embedded CRF
- Adaptive Kernel Density Estimation and Embedded Shape Priors
- Adaptive Kernel Density Estimation
- Embedded Shape Priors
- Construction of Shape Priors
- Implementation and Experimental Results
- Conclusion
- References
- Optimal Two-View Planar Scene Triangulation
- Introduction
- Planar Surface and Homography
- Triangulation for Known Plane and Cameras
- Optimal Planar Triangulation
- Triangulation for Unknown Plane and Cameras
- Experiments
- Concluding Remarks
- References
- Pursuing Atomic Video Words by Information Projection
- Introduction
- Formulation of the Atomic Video Word
- Atomic Video Words Pursuit
- Pursuing the Model of a Word by Information Projection
- Pursuing Atomic Video Words by Hybrid Modeling
- Experiments
- Dynamic Texture Segmentation by Implicit Representation
- Video Representation by Hybrid Vocabulary
- Video Scene Recognition
- Summary
- References
- A Direct Method for Estimating Planar Projective Transform
- Introduction
- Related Works
- 2D Projective Registration
- Special Case of Affine Transforms
- 1D Reduction
- Matching Cost Function Using Local Moments
- Determine the Range for h31, h32
- Refinement
- Experiments and Results
- Results
- Conclusion and Future Work
- References
- Spatial-Temporal Motion Compensation Based Video Super Resolution
- Introduction
- Problem Description
- Proposed Algorithm
- Registration by TV-L1 Optic Flow
- Spatial-Temporal Motion Compensation Based Super Resolution Fusion Framework (STMC)
- Experimental Result and Discussion
- Conclusion
- References
- Learning Rare Behaviours
- Introduction
- Related Work
- Classify and Detect Rare Behaviours by MC-LDA
- Video Clip Representation
- Multi-Class Delta Latent Dirichlet Allocation (MC-LDA)
- Learning: Modelling Rare Behaviours with Weak Supervision
- Inference: Classification of Known Rare Behaviours
- Synthesising Topics: Detection of Unknown Rare Behaviours
- Experiments
- Learning Rare Topics in Simulated Data
- Learning Rare Behaviours in Public Space CCTV Videos
- Conclusion
- References
- Character Energy and Link Energy-Based Text Extraction in Scene Images
- Introduction
- Related Work
- Proposed Approach
- Character Energy Based Text Detection
- Link Energy Based Text Grouping
- Experimental Results
- Conclusions
- References
- A Novel Representation of Palm-Print for Recognition
- Introduction
- The Discrete Cosine Transform
- Extraction of Region of Interest and Enhancement
- Feature Extraction
- Matching
- Experiments and Results
- Performance and Tuning Parameters
- Results
- Speed
- Conclusions
- References
- Real-Time Robust Image Feature Description and Matching
- Introduction
- Related Work
- Scale Invariant Feature Transform (SIFT)
- Speeded-Up Robust Features (SURF)
- Compact Signatures
- Simple Circular Accelerated Robust Features (SCARF)
- Keypoint Orientation Estimation
- Descriptor Structure
- Experimental Evaluation
- Matching Performance
- CPU Time and Memory Requirements
- Conclusions
- References
- A Biologically-Inspired Theory for Non-axiomatic Parametric Curve Completion
- Introduction
- Previous Work
- Curve Completion in the Tangent Bundle T(I)=R2S1
- Admissibility in T(I)
- ``Least Action'' Completion in T(I)
- Minimum Length Curve in the Tangent Bundle
- Theoretical Analysis
- Numerical Solution
- Dependency on Scale and Other Visual Properties
- Experimental Results
- References
- Geotagged Image Recognition by Combining Three Different Kinds of Geolocation Features
- Introduction
- Related Work
- Overview
- Methods
- Data Collection
- BoF Features
- Color Histogram
- Raw Location Vectors
- Aerial Photo Features
- Geo-information Textual Features
- Date Features
- Multiple Kernel Learning
- Experiments
- 28-Category Dataset
- Combinations of Features
- Evaluation
- Experimental Results
- Conclusion
- References
- Modeling Urban Scenes in the Spatial-Temporal Space
- Introduction
- The Approach
- Building Segmentation
- Visibility Reasoning of Buildings
- Temporal Order Recovering
- Experimental Results
- Conclusions
- References
- Family Facial Patch Resemblance Extraction
- Introduction
- Background
- Family Classification and Resemblance Extraction
- Facial Patch Template
- Features and Classifier
- Mixture of Experts
- Selection of Important Patches
- Experimental Results and Discussion
- Family Datasets
- Family Classification
- Family Facial Patch Resemblance
- Conclusion and Future Work
- References
- 3D Line Segment Detection for Unorganized Point Clouds from Multi-view Stereo
- Introduction
- Traditional Line Matching Algorithm
- The Approach
- A New Matching Model for 2D Line Segment
- 3D Line Segment Generation
- False Matching Removal
- Experimental Results and Analysis
- Simple Planar Structure Data
- Complex Building with Occlusions
- More Challenging Data
- Analysis
- Conclusion
- References
- Multi-View Stereo Reconstruction with High Dynamic Range Texture
- Introduction
- Background
- Scene Radiance and Camera Response Function
- HDR Recovering
- The Approach
- Multi-view Multi-exposure Image Capture
- Film Response Recovery and Image Calibration
- HDR Texture 3D Model Reconstruction
- Results
- Recovery of Camera Response Curve
- Reconstruction of Geometry
- Recovery of HDR Texture
- Conclusion
- References
- Feature Quarrels: The Dempster-Shafer Evidence Theory for Image Segmentation Using a Variational Framework
- Introduction
- Contribution
- Paper Organization
- Background
- Image Segmentation Using a Variational Framework
- Dempster-Shafer Evidence Theory
- Dempster-Shafer Evidence Theory for Variational Image Segmentation
- Experimental Results
- Conclusion
- References
- Gait Analysis of Gender and Age Using a Large-Scale Multi-view Gait Database
- Introduction
- Related Work
- The Gait Database
- Multi-view Synchronous Gait Capturing System
- Large-Scale Gait Database for Gender and Age
- Gait Feature Extraction
- Preliminary Experiments on Gender and Age Classification
- Method
- Classification of the Four Classes
- Classification of Specific Combinations of Classes
- Feature Analysis of Gender and Age Classes
- Discussion
- Conclusion
- References
- A System for Colorectal Tumor Classification in Magnifying Endoscopic NBI Images
- Introduction
- Motivation
- Related Work
- Feature Representation and Classification Method
- Image Representation
- Classifier
- Experimental Results
- Dataset
- Evaluation
- Experimental Results
- Conclusions
- References
- A Linear Solution to 1-Dimensional Subspace Fitting under Incomplete Data
- Introduction
- A Linear Solution
- Synthetic Evaluation
- Real-Image Application
- Projective Factorization
- Missing Observations
- Real-Image Sequences
- Conclusions
- References
- Efficient Clustering Earth Mover's Distance
- Introduction
- Problem Formulation and Related Work
- Implementation
- Experimental Results
- Summary and Outlook
- References
- Appendix
- One-Class Classification with Gaussian Processes
- Introduction
- Classification via Gaussian Processes
- The Regression Case
- From Regression to Classification
- One-Class Classification with Gaussian Process Priors
- Predictive Mean Generalizes Parzen Estimators
- Predictive Variance Models a Gaussian in Feature Space
- Support Vector Data Description as Baseline Method
- One-Class Classification for Visual Object Recognition
- Experiments
- Evaluation of One-Class Classification Methods
- Performance with an Increasing Number of Training Examples
- Influence of an Additional Smoothness Parameter
- Qualitative Evaluation
- Conclusions and Further Work
- References
- A Fast Semi-inverse Approach to Detect and Remove the Haze from a Single Image
- Introduction
- Related Work
- The Optical Model
- Haze Detection
- Our Approach
- Airlight Color (A) Estimation
- Layer-Based Dehazing
- Discussion
- Haze Detection
- Haze Removal
- Evaluating Dehazing Methods
- Conclusions
- References
- Salient Region Detection by Jointly Modeling Distinctness and Redundancy of Image Content
- Introduction
- Probabilistic Clustering by GMM
- Cluster Saliency Initialization
- Distinctness Estimation
- Redundancy Estimation
- Initial Cluster Saliency
- Cluster Saliency Propagation
- Graph-Based Propagation
- Pixel Saliency
- Experiment Evaluation
- Conclusion
- References
- Unsupervised Selective Transfer Learning for Object Recognition
- Introduction
- Related Work
- Kernel Adaptation for Knowledge Transfer
- Knowledge Transfer for Object Recognition
- Experiments
- Conclusion
- References
- A Heuristic Deformable Pedestrian Detection Method
- Introduction
- Heuristic Deformable Pedestrian Detection
- S1 Units
- S2 Units
- S3 Units
- Empirical Studies
- INRIA Human Dataset
- Pedestrian Poses Database
- Efficiency Analysis
- Discussion
- Conclusion
- References
- Gradual Sampling and Mutual Information Maximisation for Markerless Motion Capture
- Introduction
- Our Appearance-Based Body Template
- Proposed Illumination-Invariant and Error-Oriented Evaluation
- Maximisation of Mutual Information
- Gradual Sampling Scheme for Annealed Particle Filter
- Experiments and Discussion
- Conclusion and Future Work
- References
- Temporal Feature Weighting for Prototype-Based Action Recognition
- Introduction
- Temporal Action Learning
- Features
- Learning a Prototype-Based Representation
- Learning Temporal Weights
- Recognition Using Temporal Weights
- Experimental Results
- Benchmark Datasets
- Analysis of Prototype-Based Learning
- Comparison to State-of-the-Art
- Conclusion
- References
- PTZ Camera Modeling and Panoramic View Generation via Focal Plane Mapping
- Introduction
- Related Work
- Proposed Camera Model
- X-Y to Pan-Tilt Mapping
- Automatic Calculation of Focal Length
- Variation with Zoom
- Experiments
- Panorama Generation
- Pan-Tilt Recall
- Validity Across Zoom Levels
- Use in Tracking with Active PTZ Cameras
- Key Contributions and Advantages
- Conclusion and Future Work
- References
- Horror Image Recognition Based on Emotional Attention
- Introduction
- Image and Emotion Introduction
- Emotional Attention
- Emotional Saliency Map Framework
- Emotional Saliency Map Implementation
- Horror Image Recognition Based on Emotional Attention
- Attention based Bag-of-Words
- Feature Extraction
- Emotional Histogram Construction
- Experiments
- Data Set Preparation and Error Measure
- Experiments for Emotional Saliency Computation
- Horror Image Recognition
- Conclusion
- References
- Spatial-Temporal Affinity Propagation for Feature Clustering with Application to Traffic Video Analysis
- Introduction
- Affinity Propagation
- Spatial-Temporal Affinity Propagation
- Motivation
- Derivation
- Experiments
- Vehicle Detection and Tracking
- Virtual Loop Detector
- Conclusion
- References
- Minimal Representations for Uncertainty and Estimation in Projective Spaces
- Minimal Representation of Uncertainty
- Minimal Representation for Uncertain Points in 2D and 3D
- Minimal Representation for 3D Lines
- Examples
- Conclusions and Outlook
- References
- Personalized 3D-Aided 2D Facial Landmark Localization
- Introduction
- Related Work
- Method
- Framework
- Matching Problem Formulation
- Personalized 3D Facial Shape Model
- Hierarchical DAISY Representation of a Facial Landmark
- Data-Driven SAmple Consensus (DDSAC)
- Gallery/Probe Processing
- Experiments
- Multi-PIE
- UHDB14 Database
- Conclusions
- References
- A Theoretical and Numerical Study of a Phase Field Higher-Order Active Contour Model of Directed Networks
- Introduction
- Network Models
- Undirected Networks
- Directed Networks
- Theoretical Stability Analysis
- Energy of the Bar
- Stability of the Bar
- Parameter Settings in Practice
- Experiments
- Geometric Evolutions
- Results on Real Images
- Conclusion
- References
- Appendix
- Sparse Coding on Local Spatial-Temporal Volumes for Human Action Recognition
- Introduction
- Related Work
- Approach
- Local Descriptor and Sampling Strategy
- Sparse Coding Scheme for Action Recognition
- Dictionary Construction Based on Transfer Learning
- Multi-class Linear SVM
- Experiments
- Parameter Settings
- Performance on KTH Dataset
- Performance on the UCF Sports Dataset
- Dictionary Construction Analysis
- Conclusion
- References
- Multi-illumination Face Recognition from a Single Training Image per Person with Sparse Representation
- Introduction
- Preliminaries
- SRC
- Quotient Image
- SRC for One Training Image per Person
- Motivation for Synthesizing Virtual Samples
- Our Approach of Recognition with One Training Image per Person
- Experiments
- Gray Scale Images Verification
- Recognition with Color Images
- A Cross-Database Comparison
- Conclusions
- References
- Human Detection in Video over Large Viewpoint Changes
- Introduction
- Related Work
- Intra-frame and Inter-frame Comparison Features
- Granular Space
- Intra-frame and Inter-frame Comparison Features (I2CFs)
- Heuristic Learning I2CFs
- EMC-Boost
- Three Components CC/MC/SC
- Learning Algorithms of CC/MC/SC
- A General EMC-Boost
- Multiple Video Sampling
- Overview of Our Approach
- Experiments
- Conclusion
- References
- Adaptive Parameter Selection for ImageSegmentation Based on Similarity Estimation of Multiple Segmenters
- Introduction
- Parameter Selection Methods
- Framework for Unsupervised Parameter Selection
- Proposed Framework
- Segmentation Algorithms Used in Prototypical Implementation
- Similarity Measure Used in Prototypical Implementation
- Dealing with Similarity Matrix
- Handling Several Free Parameters
- Experimental Results
- Discussion
- Conclusion
- References
- Cosine Similarity Metric Learning for Face Verification
- Cosine Similarity Metric Learning
- Cosine Similarity
- Metric Learning Formulation
- Objective Function
- The Algorithm and Its Complexity
- Application to Face Verification
- LFW Dataset
- Face Verification Pipeline
- How to Choose A0 in CSML?
- Experimental Results
- Conclusion
- References
- Author Index
