
Advances in Multimedia Information Processing -- PCM 2015
Description
Alles über E-Books | Antworten auf Fragen rund um E-Books, Kopierschutz und Dateiformate finden Sie in unserem Info- & Hilfebereich.
The two-volume proceedings LNCS 9314 and 9315, constitute the proceedings of the 16 th Pacific-Rim Conference on Multimedia, PCM 2015, held in Gwangju, South Korea, in September 2015.
The total of 138 full and 32 short papers presented in these proceedings was carefully reviewed and selected from 224 submissions. The papers were organized in topical sections named: image and audio processing; multimedia content analysis; multimedia applications and services; video coding and processing; multimedia representation learning; visual understanding and recognition on big data; coding and reconstruction of multimedia data with spatial-temporal information; 3D image/video processing and applications; video/image quality assessment and processing; social media computing; human action recognition in social robotics and video surveillance; recent advances in image/video processing; new media representation and transmission technologies for emerging UHD services.
More details
Other editions
Additional editions

Content
- Intro
- Preface
- Organization
- Contents - Part I
- Contents - Part II
- Image and Audio Processing
- Internal Generative Mechanism Based Otsu Multilevel Thresholding Segmentation for Medical Brain Images
- Abstract
- 1 Introduction
- 2 Otsu Thresholding
- 3 The Proposed Segmentation Algorithm
- 3.1 Segmentation Scheme
- 3.2 Internal Generative Mechanism
- 3.3 Regrouping the Controversial Pixels
- 4 Experimental Results and Analysis
- 4.1 Experimental Settings
- 4.2 Experimental Results
- 5 Conclusion
- Acknowledgements
- References
- Efficient Face Image Deblurring via Robust Face Salient Landmark Detection
- 1 Introduction
- 2 The Proposed Method
- 2.1 Motivation
- 2.2 Robust Face Landmark Detector Training
- 2.3 Salient Contour Detection
- 2.4 Blind Image Deblurring
- 3 Experimental Results
- 3.1 Experiments on Synthesised Dataset and Real Images
- 3.2 Computation Cost Comparison
- 3.3 Adaptation to Complex Face Poses
- 3.4 Rolling Guidance Face Deblurring
- 4 Conclusions
- References
- Non-uniform Deblur Using Gyro Sensor and Long/Short Exposure Image Pair
- Abstract
- 1 Introduction
- 2 The Proposed Algorithm
- 2.1 Non-uniform Blur Model
- 2.2 IMU Sensor and Camera Motion
- 2.3 The Initial Kernel Estimation Using Gyro Data
- 2.4 Kernel Refinement
- 2.5 Deconvolution
- 3 Experimental Results
- 4 Conclusion
- References
- Object Searching with Combination of Template Matching
- Abstract
- 1 Introduction
- 2 Conventional Methods
- 3 Proposed Method
- 3.1 Partition Search Area
- 3.2 Object Identification
- 3.3 Adaptive Combination Template Matching
- 4 Experiment Results
- 5 Conclusion
- Acknowledgement
- References
- Multimedia Content Analysis
- Two-Step Greedy Subspace Clustering
- 1 Introduction
- 1.1 Related Work on Subspace Clustering
- 1.2 Paper Contributions
- 2 Two-Step Greedy Subspace Clustering
- 2.1 First Step: Initial Subspace Construction
- 2.2 Second Step: Greedy Subspace Clustering
- 3 Experiments
- 3.1 Motion Segmentation
- 3.2 Face Clustering
- 4 Conclusion
- References
- Iterative Collection Annotation for Sketch Recognition
- 1 Introduction
- 2 Overview of Proposed Method
- 3 Sketch Representation and Similarity Measuring Model
- 4 Semi-Supervised Clustering
- 5 Supervision Information Establishment
- 6 Experiments and Results
- 7 Conclusion
- References
- Supervised Dictionary Learning Based on Relationship Between Edges and Levels
- 1 Introduction
- 2 Our Approach
- 2.1 Classical Dictionary Learning
- 2.2 Our Supervised Dictionary Learning
- 2.3 Optimization Algorithm
- 3 Experimental
- 3.1 Data Set
- 3.2 Comparison Methods and Evaluation Criteria
- 3.3 Experimental Results and Analysis
- 4 Conclusions
- References
- Adaptive Margin Nearest Neighbor for Person Re-Identification
- 1 Introduction
- 2 Large Margin Nearest Neighbor
- 3 Adaptive Margin Nearest Neighbor
- 4 Experiment
- 4.1 Experiment Setting
- 4.2 Parameter Selection
- 4.3 Evaluation on VIPeR and CUHK
- 5 Conclusion
- References
- Compressed-Domain Based Camera Motion Estimation for Realtime Action Recognition
- 1 Introduction
- 2 Proposed Method
- 2.1 Camera Model
- 2.2 Estimation of T
- 2.3 Estimation of a
- 2.4 Camera Motion Compensation
- 2.5 Feature Descriptor Extraction
- 3 Experimental Results
- 3.1 GME Evaluation
- 3.2 Feature Descriptor Evaluation
- 4 Conclusion
- References
- Image and Audio Processing
- On the Security of Image Manipulation Forensics
- Abstract
- 1 Introduction
- 2 Understanding and Evaluation of Forensics Security
- 2.1 Image Manipulation Forensics Model
- 2.2 Security Evaluation and Attacks
- 3 A Case Study with Resampling Forging Attack
- 4 Experimental Results
- 5 Conclusion
- Acknowledgements
- References
- A Sparse Representation-Based Label Pruning for Image Inpainting Using Global Optimization
- Abstract
- 1 Introduction
- 2 Proposed Label Pruning
- 2.1 Dictionary Construction for Two Target Region Cases
- 2.2 Active Label Selection by Label Pruning
- 3 Experimental Results
- 4 Conclusion
- References
- Interactive RGB-D Image Segmentation Using Hierarchical Graph Cut and Geodesic Distance
- 1 Introduction
- 2 Related Work
- 3 Interactive RGB-D Image Segmentation
- 3.1 Preliminary of Hierarchical Graph Cut
- 3.2 Scale Space Construction
- 3.3 Integration of Color Cue and Depth Cue
- 3.4 Upscaling Boundary Refinement
- 4 Experiments
- 4.1 Datasets and Experimental Settings
- 4.2 Segmentation Accuracy Evaluation
- 4.3 Running Time Evaluation
- 5 Conclusions
- References
- Face Alignment with Two-Layer Shape Regression
- Abstract
- 1 Introduction
- 2 Overview
- 3 Main Work
- 3.1 Key Feature Points of a Component
- 3.2 Two-Layer Geometric Constraint
- 3.3 Sub-shape Selection
- 4 Experimental Results
- 4.1 Comparison with Previous Works
- 5 Conclusion and Future Work
- Acknowledgments
- References
- 3D Panning Based Sound Field Enhancement Method for Ambisonics
- 1 Introduction
- 2 Ambisonics Method
- 3 3D Panning Method with Sound Pressure Constraint at Two Ears
- 4 New Signal Distribution Method
- 4.1 Extension of Loudspeakers Structure
- 4.2 Calculation of the Input Signal
- 4.3 Signal Redistribution
- 4.4 Final Signals
- 5 Experiments
- 5.1 Objective Tests
- 5.2 Subjective Tests
- 6 Conclusion
- References
- Multimedia Applications and Services
- Multi-target Tracking via Max-Entropy Target Selection and Heterogeneous Camera Fusion
- 1 Introduction
- 2 Our Method
- 2.1 Online Multi-target Tracking
- 2.2 Active Camera Scheduling
- 2.3 Static and Active Camera Tracklet Association
- 2.4 Final Trajectory Generation
- 3 Experiments
- 3.1 Experiment Setting
- 3.2 Results
- 4 Conclusion
- References
- Adaptive Multiple Appearances Model Framework for Long-Term Robust Tracking
- 1 Introduction
- 2 Related Works
- 3 The Framework of Adaptive Multiple Appearances Model Tracking
- 3.1 Dirichlet Process Mixture Model
- 3.2 Model Inference
- 3.3 AMAM Tracking
- 4 Experiments
- 4.1 The AMAM Modeling
- 4.2 Tracking System
- 5 Conclusion
- References
- On-line Sample Generation for In-air Written Chinese Character Recognition Based on Leap Motion Cont ...
- Abstract
- 1 Introduction
- 2 Writing Trajectory Capturing
- 3 Proposed Method
- 3.1 Off-line Sample Generation
- 3.2 On-line Sample Generation
- 4 Experimental Results
- 5 Conclusion
- References
- Progressive Image Segmentation Using Online Learning
- 1 Introduction
- 2 Overview of Progressive Segmentation Method
- 3 Multi-level Image Representation
- 4 Online Segmentation
- 5 Experimental Result
- 6 Conclusion
- References
- A Study of Interactive Digital Multimedia Applications
- Abstract
- 1 Background
- 2 Unlimited Channel for Communication
- 3 Conclusion
- Acknowledgments
- References
- Video Coding and Processing
- Particle Filter with Ball Size Adaptive Tracking Window and Ball Feature Likelihood Model for Ball's 3D Position Tracking in Volleyball Analysis
- Abstract
- 1 Introduction
- 2 Proposal
- 2.1 Ball Size Adaptive Tracking Window
- 2.2 Volleyball Feature Likelihood Model
- 2.3 Anti-occlusion Likelihood Measurement Method
- 3 Experiment
- 3.1 Tracking Example and Evaluation Method
- 3.2 Result and Comparison Analysis
- 4 Conclusion
- Acknowledgment
- References
- Block-Based Global and Multiple-Reference Scheme for Surveillance Video Coding
- Abstract
- 1 Introduction
- 2 Analysis
- 3 The Proposed Scheme
- 3.1 Block-Based Reference Scheme
- 3.2 Multiple-Reference Scheme
- 3.3 Global Reference Scheme
- 3.4 Costs
- 4 Experimental Results
- 5 Conclusion
- Acknowledgements
- References
- Global Object Representation of Scene Surveillance Video Based on Model and Feature Parameters
- Abstract
- 1 Introduction
- 2 Global Coding Scheme of Scene Surveillance Video
- 2.1 The Generation Mechanism and Features of Global Redundancy
- 2.2 Scene Surveillance Video Global Coding Scheme
- 3 Global Object Representation Based on Model and Feature Parameters
- 3.1 Model and Shape Representation
- 3.2 Location and Pose Representation
- 3.3 Texture Parameters Representation
- 3.4 Illumination Parameters Representation
- 4 Experiments and Results
- 4.1 Experiment 1
- 4.2 Experiment 2
- 5 Conclusions
- References
- A Sparse Error Compensation Based Incremental Principal Component Analysis Method for Foreground Detection
- 1 Introduction
- 2 Foreground Detection via Sparse Error Compensation Based Incremental PCA
- 2.1 The Proposed Subspace Based Foreground Detection Model
- 2.2 Two-Step Optimization Algorithm
- 3 Experiments
- 4 Conclusion
- References
- Multimedia Representation Learning
- Convolutional Neural Networks Features: Principal Pyramidal Convolution
- Abstract
- 1 Introduction
- 2 Principal Pyramidal Convolution
- 3 Experiment
- 3.1 Datasets
- 3.2 Comparisons on Different Networks
- 3.3 Comparisons on Different Dimensions
- 4 Conclusion
- References
- Gaze Shifting Kernel: Engineering Perceptually- Aware Features for Scene Categorization
- 1 Introduction
- 2 Related Work
- 3 The Proposed Gaze Shifting Kernel
- 3.1 Low-Level and High-Level Descriptions of Graphlets
- 3.2 Sparsity-Constrained Graphlets Ranking
- 3.3 Gaze Shifting Kernel and SVM Training
- 4 Experimental Results and Analysis
- 4.1 Comparison with the State-of-the-Art
- 4.2 Parameters Analysis
- 4.3 Visualization Results
- 5 Conclusion
- References
- Two-Phase Representation Based Classification
- 1 Introduction
- 2 The Proposed TPLRMC
- 2.1 The Motivation of the TPLRMC
- 2.2 The First Phase of the TPLRMC
- 2.3 The Second Phase of the TPLRMC
- 2.4 Analysis of the TPLRMC
- 3 Experiments
- 3.1 Databases
- 3.2 Experimental Results
- 4 Conclusions
- References
- Deep Feature Representation via Multiple Stack Auto-Encoders
- 1 Introduction
- 2 Our Method
- 2.1 The Basic Auto-Encoder
- 2.2 Building the Multiple Multi-level Auto-Encoders
- 2.3 The Layer-Wise Training and Fine Tuning
- 2.4 The Weight Assigned for Each Feature
- 2.5 Classification
- 3 Experiments
- 3.1 The MNIST
- 3.2 The CIFAR 10
- 4 Conclusion
- References
- Beyond HOG: Learning Local Parts for Object Detection
- 1 Introduction
- 2 Related Work
- 3 The Proposed Method
- 4 Experiment
- 4.1 Datasets and Details
- 4.2 Result and Discussion
- 5 Conclusion
- References
- Regular Poster Session
- Tuning Sparsity for Face Hallucination Representation
- Abstract
- 1 Introduction
- 2 Locally Weighted &hx2113
- 1 Regularization
- 3 Extension to &hx2113
- 1,2 for Regularizing Noisy Images
- 4 Experimental Results
- 4.1 Quantitative Evaluation
- 4.2 Comparisons of Subjective Results
- 5 Conclusions
- Acknowledgments
- References
- Visual Tracking by Assembling Multiple Correlation Filters
- Abstract
- 1 Introduction
- 2 Kernalized Correlation Filter
- 2.1 Linear Regression
- 2.2 Kernel Regression
- 3 Correlation Filter Fusion
- 3.1 Online Correlation Filter Update
- 3.2 Budgeting on Correlation Filters
- 4 Experiments
- 5 Conclusions
- References
- A Unified Tone Mapping Operation for HDR Images Including Both Floating-Point and Integer Data
- 1 Introduction
- 2 Preliminaries
- 2.1 Floating-Point HDR Image Formats
- 2.2 Global Tone Mapping Operation
- 3 Proposed Method
- 3.1 Unified TMO
- 3.2 Intermediate Format
- 3.3 Integer TMO for the Intermediate Format
- 3.4 Fixed-Point Arithmetic
- 4 Experimental and Evaluation Results
- 4.1 Comparison of Tone-Mapped LDR Images
- 4.2 Comparison of the Memory Usage
- 4.3 Comparison of the Processing Time
- 5 Conclusion
- References
- Implementation of Human Action Recognition System Using Multiple Kinect Sensors
- 1 Introduction
- 2 Proposed Human Action Recognition System
- 2.1 Multi-view Skeleton Integration
- 2.2 Snapshot Feature Extraction
- 2.3 Temporal Feature Extraction
- 2.4 Classification
- 3 Experiment and Results
- 4 Conclusion
- References
- Simplification of 3D Multichannel Sound System Based on Multizone Soundfield Reproduction
- Abstract
- 1 Introduction
- 1.1 Related Work
- 2 Problem Formulation
- 2.1 Multizone Soundfield Model
- 2.2 Formulation of Simplification from L- to (L-1)-Channel
- 2.3 Loudspeaker Weight Coefficients
- 3 Simulation and Error Analysis
- 3.1 Simplification Results
- 3.2 Simulation Results and Comparison Analysis
- 3.3 Subjective Results and Comparison Analysis
- 4 Conclusion and Future Work
- References
- Multi-channel Object-Based Spatial Parameter Compression Approach for 3D Audio
- Abstract
- 1 Introduction
- 2 Background
- 2.1 Directional Audio Coding (DirAC)
- 2.2 3D Audio Spatial Localization Quantization Method
- 2.3 The Existing Compression Approaches of Spatial Parameters
- 3 Proposed Spatial Parameter Compression Approach
- 3.1 Proposed Spatial Parameter Compression Scheme
- 3.2 Multi-channel Object-Based Spatial Parameter Compression Approach
- 4 Performance Evaluation
- 4.1 Objective Quality Evaluation
- 4.2 Subjective Quality Evaluation
- 5 Conclusions
- Acknowledgement
- References
- A FPGA Based High-Speed Binocular Active Vision System for Tracking Circle-Shaped Target
- Abstract
- 1 Introduction
- 2 Active Vision System
- 3 Target Tracking
- 3.1 Prior Knowledge
- 3.2 Gradients in Fixed Directions
- 3.3 Circle Fitting
- 3.4 Improvement
- 4 FPGA Implementation
- 5 3-D Localization and Pan-Tilt Control
- 6 Experiment
- 6.1 Hardware Environment
- 6.2 Target Tracking Experiment
- 7 Conclusions
- Acknowledgements
- References
- The Extraction of Powerful and Attractive Video Contents Based on One Class SVM
- Abstract
- 1 Introduction
- 2 Video Summarization of Powerful Contents Based on OCSVM
- 2.1 Extraction of Key Frames and Features
- 2.2 Powerful Frames Selection with One Class SVM
- 3 Experiments and Discussion
- 4 Conclusion
- References
- Blur Detection Using Multi-method Fusion
- 1 Introduction
- 2 Motivation
- 3 The Methodology
- 3.1 Multi-method Fusion via CRF
- 3.2 Locality-Aware Multi-method Fusion
- 4 Experiments
- 4.1 Experimental Settings
- 4.2 Results and Discussion
- 5 Conclusion
- References
- Motion Vector and Players' Features Based Particle Filter for Volleyball Players Tracking in 3D Space
- Abstract
- 1 Introduction
- 2 Proposal
- 2.1 Motion Vector Prediction Model
- 2.2 Players' Features Based Likelihood Model
- 3 Experiment and Result
- 4 Conclusion
- Acknowledgement
- References
- A Novel Edit Propagation Algorithm via L0 Gradient Minimization
- 1 Introduction
- 2 Related Works
- 3 The L0 Propagation Method
- 3.1 Algorithm Framework
- 3.2 Affinity Matrix Approximation
- 3.3 Constrain Parameters
- 4 Experiments
- 4.1 Implemention
- 4.2 Recoloring
- 4.3 Tonal Values Adjustments
- 5 Discussions and Conclusions
- References
- Improved Salient Object Detection Based on Background Priors
- 1 Introduction
- 2 Improved Saliency Detection Based on Background Priors
- 2.1 Pre-processing
- 2.2 Initial Saliency Map Calculation
- 2.3 Saliency Map Refinement
- 3 Experiments
- 3.1 Performance Evalation on ASD and MSRA Datasets
- 3.2 Effectiveness of Saliency Map Refinement
- 4 Conclusion
- References
- Position-Patch Based Face Hallucination via High-Resolution Reconstructed-Weights Representation
- 1 Introduction
- 2 High-Resolution Reconstruction Weights
- 2.1 High-Resolution Reconstruction Weights
- 3 High-Resolution Reconstructed-Weights Representation
- 3.1 Estimate of the HR Reconstruction Weights
- 3.2 Face Hallucination via HRR
- 4 Experimental Results
- 4.1 Experiment Settings
- 4.2 Results Comparison
- 4.3 Influence of Parameters
- 5 Conclusion
- References
- Real-Time Rendering of Layered Materials with Linearly Filterable Reflectance Model
- 1 Introduction
- 2 Related Work
- 3 Proposed System
- 3.1 Overview
- 3.2 Surface Reflection
- 3.3 Subsurface Reflection
- 4 Results
- 5 Conclusions
- References
- Hybrid Lossless-Lossy Compression for Real-Time Depth-Sensor Streams in 3D Telepresence Applications
- 1 Introduction
- 2 Related Work
- 3 Compression Approach
- 3.1 System Overview
- 3.2 Analysis of Depth-Bit Assignment
- 3.3 Optimal X264 Encoding Parameters for Depth Image Streams
- 4 Results and Discussion
- 4.1 Evaluation of Depth-Bit Assignment-Methods
- 4.2 Evaluation of x264 Encoding Parameter Settings
- 4.3 Depth Compression for Real-Time 3D Reconstruction
- 5 Conclusion
- References
- Marginal Fisher Regression Classification for Face Recognition
- Abstract
- 1 Introduction
- 2 Linear Regression Classification
- 3 Marginal Fisher Regression Classification
- 4 Experimental Results
- 4.1 Experiment on FERET Dataset
- 4.2 Experiment on PIE Dataset
- 4.3 Experiment on AR Dataset
- 5 Conclusion and Future Work
- Acknowledgements
- References
- Temporally Adaptive Quantization Algorithm in Hybrid Video Encoder
- Abstract
- 1 Introduction
- 2 The Temporally Adaptive Quantization Algorithm
- 3 The Proposed CDA Based delta - rho Model
- 3.1 The Proposed delta - rho Model
- 3.2 The Improved Quantization Control Algorithm
- 4 Simulation Results and Analysis
- 5 Conclusions
- Acknowledgment
- References
- Semi-automatic Labeling with Active Learning for Multi-label Image Classification
- Abstract
- 1 Introduction
- 2 Label Correlation Based Sampling Strategy
- 3 Semi-automatic Labeling with Active Learning
- 3.1 Automatic Labeling Strategy
- 3.2 Complete Algorithm
- 4 Experiments
- 4.1 On Image Datasets
- 4.2 On Non-image Datasets
- 5 Conclusions
- Acknowledgement
- References
- A New Multi-modal Technique for Bib Number/Text Detection in Natural Images
- Abstract
- 1 Introduction
- 2 Proposed Technique
- 2.1 Text Candidate Region Detection
- 2.2 Multi-modal Method for Text Detection/Recognition
- 3 Experimental Results
- 3.1 Experiments on Text Candidate Region Detection
- 3.2 Validating Multi-modality Through Text Detection
- 3.3 Validating Multi-modality Through Recognition
- 4 Conclusion and Future Work
- Acknowledgment
- References
- A New Multi-spectral Fusion Method for Degraded Video Text Frame Enhancement
- Abstract
- 1 Introduction
- 2 Proposed Methodology
- 2.1 Multi-spectral Images for Reducing Degradation Effect
- 2.2 Multi-spectral Fusion-1 for Text Frame Enhancement
- 2.3 Multi-spectral Fusion-2 for Text Frame Enhancement
- 3 Experimental Results
- 3.1 Experiments on Measuring Quality of the Enhanced Frame
- 3.2 Validating Enhancement Through Text Detection
- 3.3 Validating Enhancement Through Recognition
- 4 Conclusion
- Acknowledgment
- References
- A Robust Video Text Extraction and Recognition Approach Using OCR Feedback Information
- Abstract
- 1 Introduction
- 2 Related Work
- 3 Video Text Segmentation
- 4 Text Extraction
- 5 Best Extraction Schemes Choosing
- 6 Experimental Results
- 6.1 Performance of Text Segmentation
- 6.2 Performance of Text Extraction
- 6.3 Recognition Performance with Best Scheme Choosing
- 7 Conclusions
- References
- Color and Active Infrared Vision: Estimate Infrared Vision of Printed Color Using Bayesian Classifier and K-Nearest Neighbor Regression
- 1 Introduction
- 2 Related Works
- 3 Proposed Methods
- 3.1 Prediction by Bayesian Classifier
- 3.2 Regression by K-Nearest Neighbors
- 4 Experimental Results
- 4.1 Prediction by Bayesian Classifier
- 4.2 Regression by K-Nearest Neighbors
- 5 Conclusion
- References
- Low Bitrates Audio Bandwidth Extension Using a Deep Auto-Encoder
- Abstract
- 1 Introduction
- 2 Proposed Methodology
- 2.1 Motivation
- 2.2 Proposed BWE Method
- 3 Prediction of Fine Structure Using Deep Auto-Encoder
- 3.1 Auto-Encoders
- 3.2 Prediction of Fine Structure
- 4 Experiments and Evaluation
- 4.1 Training Auto-Encoders
- 4.2 The Results of Implement
- 4.3 Performance Evaluation
- 5 Conclusions
- Acknowledgments
- References
- Part-Aware Segmentation for Fine-Grained Categorization
- 1 Introduction
- 2 Hybrid Part Localization
- 3 Part-Aware Segmentation
- 3.1 Definitions
- 3.2 Optimization
- 4 Experiments
- 4.1 Dataset
- 4.2 Part Localization Results
- 4.3 Segmentation Results
- 4.4 Recognition Results
- 5 Conclusion
- References
- Improved Compressed Sensing Based 3D Soft Tissue Surface Reconstruction
- Abstract
- 1 Introduction
- 2 Surface Reconstruction
- 2.1 RBF Interpolation
- 2.2 CS Reconstruction Algorithm
- 3 Experiment Results
- 3.1 Data Sets
- 3.2 Comparison of Different Measurement Matrix
- 3.3 Comparison of LI and RBFI
- 3.4 Comparison with Other Methods
- 4 Conclusions
- Acknowledgment
- References
- Constructing Learning Maps for Lecture Videos by Exploring Wikipedia Knowledge
- 1 Introduction
- 2 Construction of Video Map by Exploring Domain Knowledge from Wikipedia
- 2.1 Extraction of Concept Words
- 2.2 Representing Video Content with Revised TF-IDF
- 2.3 Construction of Video Map by Maximum Spanning Tree
- 3 Construction of Concept Map by Integrating Wikipedia Knowledge and Lecture Videos
- 3.1 Constructing Undirected Concept Map with Wikipedia Articles
- 3.2 Constructing Directed Concept Map by Discovering the Prerequisite Relationships Between Concepts
- 4 Experiments
- 4.1 Evaluation of Video Maps
- 4.2 Evaluation of Concept Maps
- 5 Conclusions
- References
- Object Tracking via Combining Discriminative Global and Generative Local Models
- 1 Introduction
- 2 Proposed Algorithm
- 2.1 Discriminative Global Model
- 2.2 Generative Local Model
- 2.3 Template Update
- 3 Experiments
- 3.1 Implementation Details
- 3.2 Performance Evaluation
- 4 Conclusion
- References
- Tracking Deformable Target via Multi-cues Active Contours
- 1 Introduction
- 2 Proposed Method
- 2.1 Contour Based Meanshift Target Locating
- 2.2 Appearance Model Combing Global and Local Layers
- 2.3 Dynamic Shape Model
- 2.4 Multi-cues Active Contours and Curve Evolution
- 3 Experimental Results
- 3.1 Experimental Setup
- 3.2 Qualitative and Quantitative Analysis
- 4 Conclusion
- References
- Person Re-identification via Attribute Confidence and Saliency
- 1 Introduction
- 2 The Proposed Approach
- 2.1 Classifiers Training
- 2.2 Attribute Confidence and Saliency Calculation
- 2.3 Attribute Confidence and Saliency Matching Method
- 3 Experiments and Results
- 3.1 Dataset and Evaluation Protocol
- 3.2 Results and Discussions
- 4 Conclusions
- References
- Light Field Editing Based on Reparameterization
- 1 Introduction
- 2 Related Work
- 3 Light Field Editing Framework
- 3.1 Light Field Reparameterization
- 3.2 Downsampling-Upsampling Propagation Framework
- 4 Results
- 5 Conclusion
- References
- Interactive Animating Virtual Characters with the Human Body
- Abstract
- 1 Introduction
- 2 Related Work
- 3 System Pipeline
- 4 Methods
- 4.1 Embedded Deformation [4]
- 4.2 Construct Deformation Graph
- 4.3 Optimization
- 5 Experiment
- 5.1 Experimental Setup
- 5.2 Animating Disproportionate Avatar
- 5.3 Animating Non-humanoid Avatar
- 6 Conclusion
- References
- Visual Understanding and Recognition on Big Data
- Fast Graph Similarity Search via Locality Sensitive Hashing
- 1 Introduction
- 2 The Proposed Method
- 2.1 Vectorial Representation
- 2.2 Fast Similarity Search
- 2.3 Retrieval Framework
- 2.4 Complexity
- 3 Experimental Evaluation
- 3.1 Experimental Setting
- 3.2 Experimental Results
- 4 Conclusions
- References
- Text Localization with Hierarchical Multiple Feature Learning
- Abstract
- 1 Introduction
- 2 Character Localization
- 2.1 Structure Features
- 2.2 HOG Feature
- 2.3 CNN-Based Features
- 3 Text Line Formation
- 4 String Splitting
- 5 Experimental Results
- 6 Conclusions
- Acknowledgments
- References
- Recognizing Human Actions by Sharing Knowledge in Implicit Action Groups
- 1 Introduction
- 2 The Proposed Method
- 2.1 Fisher Vector
- 2.2 Exploring the Implicit Group Structure
- 2.3 Model Learning
- 3 Experiment
- 3.1 Experimental Setup
- 3.2 Experiments on HMDB51 Dataset
- 4 Conclusion
- References
- Human Parsing via Shape Boltzmann Machine Networks
- 1 Introduction
- 2 Model Design
- 2.1 Model Structure for Human Parsing
- 2.2 Multi Channel Segmentation by Shape Boltzmann Machine Network
- 2.3 Similarity Measurement of the Curve and Curve Correction
- 2.4 Overlap Regions and Missing Regions
- 3 Experiments
- 3.1 Dataset and Implements
- 3.2 Results and Performances
- 4 Conclusion
- References
- Depth-Based Stereoscopic Projection Approach for 3D Saliency Detection
- Abstract
- 1 Introduction
- 2 Depth-Based Stereoscopic Projection Approach
- 2.1 Three-Dimensional Reconstruction and Stereographic Projection
- 2.2 Processing Based on Characteristics of Projected Images
- 2.3 Generating Depth Saliency Map and 3D Saliency Map
- 3 The Experimental Results and Analysis
- 4 Conclusion
- References
- Coding and Reconstruction of Multimedia Data with Spatial-Temporal Information
- Revisiting Single Image Super-Resolution Under Internet Environment: Blur Kernels and Reconstruction Algorithms
- 1 Introduction
- 2 Evaluated SISR Methods
- 2.1 A Fast and Effective SISR Method Based on Mixture of Experts
- 3 Experimental Settings
- 3.1 Blur Kernels and Scaling Factors
- 3.2 Datasets and Features
- 4 Evaluation Results
- 4.1 SISR Methods w.r.t. Mismatched Blur Kernels
- 4.2 SISR Methods w.r.t. Mismatched Blur Kernels
- 5 Conclusions
- References
- Prediction Model of Multi-channel Audio Quality Based on Multiple Linear Regression
- Abstract
- 1 Introduction
- 2 Objective Measurement of Spatial Audio Quality
- 2.1 Motivation of the Prediction Model
- 2.2 Extracting the Objective Parameters
- 2.3 Design of Subjective Listening Test
- 3 The Specific Structure of the Model
- 3.1 Data Pre-processing
- 3.2 Principle Components Extraction
- 3.3 Quality Measurement with MLR
- 4 Performance Analysis
- 4.1 Training and Testing Data Set
- 4.2 Algorithm Evaluation
- 5 Conclusions
- References
- Physical Properties of Sound Field Based Estimation of Phantom Source in 3D
- 1 Introduction
- 2 Proposed Physical Properties Based Method
- 2.1 Verification of Panning Law Based on Physical Properties of Sound Field
- 2.2 Estimation of Phantom Source for Symmetric Arrangement Case
- 2.3 Estimation of Phantom Source for Asymmetric Arrangement Case
- 3 Experiment and Analysis
- 3.1 Objective Experiments
- 3.2 Subjective Experiment
- 4 Conclusion
- References
- Non-overlapped Multi-source Surveillance Video Coding Using Two-Layer Knowledge Dictionary
- Abstract
- 1 Introduction
- 2 Global Object Redundancy
- 2.1 Analysis of Global Object Redundancy
- 2.2 Two-Layer Knowledge Dictionary for Eliminating Global Object Redundancy
- 3 Proposed Method
- 3.1 Two-Layer Dictionary Learning
- 3.2 Dictionary-Based Coding Scheme for Moving Vehicles
- 3.3 Overview of the Coding Framework
- 4 Experimental Results
- 5 Conclusion
- References
- Global Motion Information Based Depth Map Sequence Coding
- 1 Introduction
- 2 Proposed Method
- 2.1 Depth Map Skipping
- 2.2 Depth Map Projection
- 3 Experimental Method and Results
- 3.1 Data Acquisition and Prototypes
- 3.2 Experiments and Results
- 4 Conclusions
- References
- Author Index
System requirements
File format: PDF
Copy protection: Watermark-DRM (Digital Rights Management)
System requirements:
- Computer (Windows; MacOS X; Linux): Use the free software Adobe Reader, Adobe Digital Editions, or any other PDF viewer of your choice (see eBook Help).
- Tablet/Smartphone (Android; iOS): Install the free app Adobe Digital Editions or another reading app for eBooks, e.g., PocketBook (see eBook Help).
- E-reader: Bookeen, Kobo, Pocketbook, Sony, Tolino and many more (only limited: Kindle).
The file format PDF always displays a book page identically on any hardware. This makes PDF suitable for complex layouts such as those used in textbooks and reference books (images, tables, columns, footnotes). Unfortunately, on the small screens of e-readers or smartphones, PDFs are rather annoying, requiring too much scrolling.
This eBook uses Watermark-DRM, a „soft” copy protection. This means that there are no technical restrictions to prevent illegal distribution. However, there is a personalised watermark embedded in the eBook that can be used to identify the purchaser of the eBook in the event of misuse and to provide evidence for legal purposes.
For more information, see our eBook Help page.