Pattern Recognition Applications and Methods: 6th International Conference, ICPRAM 2017, Porto, Portugal, February 24–26, 2017, Revised Selected Papers (Lecture Notes in Computer Science, 10857)

✍ Scribed by Maria De Marsico (editor), Gabriella Sanniti di Baja (editor), Ana Fred (editor)

• Publisher: Springer
• Tongue: English
• Leaves: 250
• Category: Library

✦ Synopsis

This book contains revised and extended versions of selected papers from the 6th International Conference on Pattern Recognition, ICPRAM 2017, held in Porto, Portugal, in February 2017.

The 13 full papers presented were carefully reviewed and selected from 139 initial submissions. They aim at making visible and understandable the relevant trends of current research on pattern recognition.

✦ Table of Contents

Preface
Organization
Contents
Control Variates as a Variance Reduction Technique for Random Projections
1 Introduction
1.1 Notation and Intuition
1.2 Probability Bounds on Random Projection Estimates
1.3 Control Variates
1.4 Related Work
1.5 Our Contributions
2 Process of Using Control Variates
2.1 Control Variate for the Euclidean Distance
2.2 Control Variate for the Inner Product
2.3 The Optimal Control Variate Correction c
2.4 Overall Computational Time
3 Our Experiments
3.1 Generating Vectors from Synthetic Data
3.2 Experiments with Real Data
4 Conclusion and Future Work
References
Graph Classification with Mapping Distance Graph Kernels
1 Introduction
2 Preliminaries
3 Mapping Distance Graph Kernels
3.1 Graph Kernels Based on the Graph Edit Distance
3.2 Graph Kernels Based on Relabeling
4 Evaluation Experiments
4.1 Evaluation Using Synthetic Datasets
4.2 Classification Accuracy
5 Discussion on Related Work
5.1 Graph Edit Distance Based on the Mapping Distance
5.2 Support Vector Machines with Graph Kernels
6 Conclusion
References
Domain Adaptation Transfer Learning by Kernel Representation Adaptation
1 Introduction
2 Related Work
3 Presentation of the MMD Constrained SVM Method
3.1 Review of Basic Theoretical Foundations
3.2 MMD-Like Constrained SVM Transfer Learning
4 Dual Form of the Optimization Problem
5 Extension to KPCA Alignment
5.1 A Brief Review of PCA and KPCA
5.2 KPCA Transfer Learning via Alignment of Data Representations
6 Experiments
6.1 Data Sets
6.2 Experimental Results and Analysis
7 Conclusion and Future Directions
References
Optimal Linear Imputation with a Convergence Guarantee
1 Introduction
2 The Optimized Linear Imputation Method
2.1 Notation
2.2 Optimization Problem
2.3 Block Coordinate Descent Solution
2.4 Discussion
3 Non-linear Imputation Methods
4 Evaluation
4.1 Synthetic Data
4.2 UCI Datasets
4.3 Storks Behavioral Modes Dataset
5 Conclusion
References
Condensing Deep Fisher Vectors: To Choose or to Compress?
1 Introduction
2 The Fisher Kernel Framework
3 The Fisher Vector Normalisation
4 Compression Techniques
4.1 Principal Component Analysis (PCA)
4.2 Spectral Hashing (SH)
4.3 Autoencoder
4.4 Parametric t-SNE
5 Feature Selection Techniques
5.1 Conditional Mutual Information (MI)
5.2 Minimum Redundancy and Maximum Relevance (MRMR)
5.3 SVM-Recursive Feature Elimination (SVM-RFE)
5.4 Random Forest (RF)
6 Multi-collinearity Assessment Diagnostics
6.1 Variance Inflation Factor (VIF)
6.2 Condition Indices
7 Experiments
7.1 Experimental Setup
8 Discussion
9 Conclusion
References
Emotion Recognition Using Neighborhood Components Analysis and ECG/HRV-Based Features
Abstract
1 Introduction
2 Literature Studies
2.1 Supervised Dimensionality Reduction
2.2 Literature Review
3 Methods
3.1 Feature Extraction
3.2 Dimensionality Reduction
3.3 Classifier and Validation Method
3.4 Post-processing for the Final Feature Dimension
4 Results and Discussions
4.1 HRV-Based Features
4.2 HHT-Based Features
4.3 Summary of the Experiments
5 Conclusions
References
A Conversive Hidden Non-Markovian Model Approach for 2D and 3D Online Movement Trajectory Verification
1 Introduction
2 Related Work
2.1 Previous Work
2.2 Related Work
3 The CHnMM Verification System
3.1 CHnMM - Formal Definition
3.2 Trajectory Model Structure
3.3 Creating the StrokeMap
3.4 Creating the CHnMM
3.5 Trajectory Verification
4 Experiments
4.1 Databases
4.2 Experiment Protocol
4.3 Results
5 Conclusions
References
Prediction of User Interest by Predicting Product Text Reviews
1 Introduction
1.1 Recommendation Systems Review
1.2 Linguistic Processing
1.3 Contributions
2 User Modeling Based on Opinions
2.1 Notation
2.2 Predicting the Opinion Using Alternating Least Squares (ALS)
3 Experiments
4 Results
5 Conclusions
References
Blood Vessel Delineation in Endoscopic Images with Deep Learning Based Scene Classification
1 Introduction
2 Proposed Method
2.1 Scene-Based Classification
2.2 Blood Vessel Extraction
2.3 Background Removal
2.4 Removing Non-blood Vessel Edges
3 Experiments and Results
3.1 Scene Classification
3.2 Background Removal
3.3 Removal of Non-blood Vessel Edges
4 Conclusion and Future Work
References
Semi-automated Testing of an Architectural Floor Plan Retrieval Framework: Quantitative and Qualitative Comparison of Semantic Pattern-Based Matching Approaches
1 Introduction
2 Related Work
2.1 Case-Based Reasoning
2.2 Graph and Subgraph Matching
2.3 Sketch-Based Interfaces
3 Floor Plans Retrieval Techniques
3.1 Semantic Fingerprints Concept
3.2 Query Structure
3.3 Matching Techniques
3.4 Augmentation of Retrieved Floor Plans
4 Evaluation of Our System
4.1 Computational Limitations (Boundary Test)
4.2 Qualitative Analysis
4.3 Query-Result Mapping Case Study
5 Conclusion and Future Work
References
Characterization of a Virtual Glove for Hand Rehabilitation Based on Orthogonal LEAP Controllers
1 Introduction
2 System Set Up
3 Calibration
4 Spatial Characterization and Tracking
4.1 Spatial Characterization
4.2 Preliminary Hand Tracking
5 Conclusions
References
Congestion Analysis Across Locations Based on Wi-Fi Signal Sensing
1 Introduction
2 Related Work
2.1 People Flow Analysis Based on Wi-Fi Packet Sensing
2.2 Prediction and Recommendation Based on Tensor Factorization
3 Congestion Estimate and Collecting User Reports
3.1 System Overview
3.2 Probe Request Capturing and Filtering
3.3 Congestion Degree Based on Probe Requests
3.4 Visualizing Congestion and User Report
4 Analysis of Congestion and User Reports
4.1 Operation of Our System
4.2 Time Series of Congestion
4.3 Correlation Analysis of User Reports
5 Spatio-Temporal Feature Analysis Across Locations
5.1 Analysis Overview
5.2 Non-negative Tensor Factorization (NTF) for Extracting Understandable Patterns
6 Experimental Results
6.1 Dataset
6.2 Latent Patterns
6.3 Subjective Evaluation of Latent Patterns
7 Conclusion
References
Text Line Segmentation in Handwritten Documents Based on Connected Components Trajectory Generation
1 Introduction
2 Our Approach
2.1 Preprocessing
2.2 Connected Components Tracking
2.3 Merging Nearby Trajectories
2.4 Small Regions Label Propagation
3 Experimental Results
3.1 Dataset Used
3.2 Metrics Used
3.3 Parameters Tuning
3.4 Final Results
4 Conclusion
References
Author Index