Structural, Syntactic, and Statistical Pattern Recognition

Name: Structural, Syntactic, and Statistical Pattern Recognition | Joint IAPR International Workshops, S+SSPR 2022, Montreal, QC, Canada, August 26-27, 2022, Proceedings
Brand: Springer
Price: 106.99 EUR
Availability: OnlineOnly

Joint IAPR International Workshops, S+SSPR 2022, Montreal, QC, Canada, August 26-27, 2022, Proceedings

Adam Krzyzak Ching Y. Suen Andrea Torsello Nicola Nobile(Herausgeber*in)

Springer (Verlag)

Erschienen am 1. Januar 2023

XIII, 324 Seiten

E-Book

PDF mit Wasserzeichen-DRM

Systemvoraussetzungen

978-3-031-23028-8 (ISBN)

106,99 €inkl. 7% MwSt.

Systemvoraussetzungen

für PDF mit Wasserzeichen-DRM

E-Book Einzellizenz

Als Download verfügbar

Beschreibung

Weitere Details

Weitere Ausgaben

Inhalt

Intro
Preface
Organization
Contents
Realization of Autoencoders by Kernel Methods
1 Introduction
2 Related Work
3 Autoencoders by Kernel Methods
3.1 Encoder and Decoder
3.2 Fundamental Mapping Without Loss
3.3 Kernelized Autoencoder
4 Comparison with Neural Networks
5 Applications
5.1 Denoising Autoencoders
5.2 Generative Autoencoders
6 Discussion
7 Conclusion
References
Maximal Independent Vertex Set Applied to Graph Pooling
1 Introduction
2 Related Work
2.1 Graph Pooling
3 Proposed Method
3.1 Maximal Independent Vertex Set (MIVS)
3.2 Adaptation of MIVS to Deep Learning
4 Experiments
4.1 Datasets
4.2 Model Architecture and Training Procedure
4.3 Ablation Studies
4.4 Comparison of MIVSPool According to Other Methods
5 Conclusion
References
Annotation-Free Keyword Spotting in Historical Vietnamese Manuscripts Using Graph Matching
1 Introduction
2 Kieu Database
3 Annotation-Free Keyword Spotting (KWS)
3.1 Synthetic Dataset Creation
3.2 Character Detection
3.3 Graph Extraction
3.4 Graph Matching
3.5 Keyword Spotting (KWS)
4 Experimental Evaluation
4.1 Task Setup and Parameter Optimization
4.2 Results
4.3 Ablation Study
5 Conclusions
References
Interactive Generalized Dirichlet Mixture Allocation Model
1 Introduction
2 Model Description
3 Variational Inference
4 Interactive Learning Algorithm
5 Experimental Results
6 Conclusion
References
Classifying Me Softly: A Novel Graph Neural Network Based on Features Soft-Alignment
1 Introduction
2 Related Work
3 Features Soft-Alignment Graph Neural Networks
4 Experiments
4.1 Experimental Setup
4.2 Ablation Study
4.3 Graph Classification Results
4.4 Graph Regression Results
5 Conclusion
References
Review of Handwriting Analysis for Predicting Personality Traits
1 Introduction
1.1 History
1.2 Applications
1.3 Requirements
2 Research Progress
2.1 Advantages
2.2 Disadvantages
3 Research Steps
3.1 Database
3.2 Pre-processing
3.3 Feature Extraction
3.4 Personality Trait
3.5 Prediction Model
3.6 Performance Measurement
4 Experiment and Future Work
4.1 Experiment
4.2 Future Work
References
Graph Reduction Neural Networks for Structural Pattern Recognition
1 Introduction and Related Work
2 Graph Matching on GNN Reduced Graphs
2.1 Graph Reduction Neural Network (GReNN)
2.2 Classification of GReNN Reduced Graphs
3 Empirical Evaluations
3.1 Datasets and Experimental Setup
3.2 Analysis of the Structure of the Reduced Graphs
3.3 Classification Results
3.4 Ablation Study
4 Conclusions and Future Work
References
Sentiment Analysis from User Reviews Using a Hybrid Generative-Discriminative HMM-SVM Approach
1 Introduction
2 Related Work
3 Hybrid Generative-Discriminative Approach with Fisher Kernels
3.1 Hidden Markov Models
3.2 Inference on Hidden States: Forward-Backward Algorithm
3.3 Fisher Kernels
4 Experiments
4.1 Problem Modeling
4.2 Datasets
4.3 Results
5 Conclusion
References
Spatio-Temporal United Memory for Video Anomaly Detection
1 Introduction
2 Related Work
2.1 Dual-Flow Structure Based on Autoencoder
2.2 Memory
3 Methodology
4 Experiments
4.1 Datasets and Evaluation Metrics
4.2 Comparison with Existing Methods
4.3 Ablation Experiments
4.4 Running Time
5 Conclusion
References
A New Preprocessing Method for Measuring Image Visual Quality Robust to Rotation and Spatial Shifts
1 Introduction
2 Proposed Preprocessing Method
3 Experimental Results
4 Conclusions
References
Learning Distances Between Graph Nodes and Edges
1 Introduction
2 Related Work
2.1 Graph Edit Distance
2.2 Learning the Edit Costs
3 Method
3.1 The Learning Method
3.2 The Algorithm
4 Practical Experiments
5 Conclusions
References
Self-supervised Out-of-Distribution Detection with Dynamic Latent Scale GAN
1 Introduction
2 Out-of-Distribution Detection with DLSGAN
3 Experiments
3.1 Experiments Settings
3.2 Experiments Results
4 Conclusion
References
A Novel Graph Kernel Based on the Wasserstein Distance and Spectral Signatures
1 Introduction
2 Related Work
3 Background
3.1 Spectral Signatures: HKS and WKS
3.2 Wasserstein Distance
4 From Spectral Signatures to Graph Kernels
5 Experiments
5.1 Experimental Setup
5.2 Sensitivity Study
5.3 Graph Classification Results
6 Conclusion
References
Discovering Respects for Visual Similarity
1 Introduction
2 Method
2.1 Dataset Selection
2.2 Image Representation
2.3 What
2.4 Why
3 Evaluations
3.1 Automatic Interpretability and Human Validation
3.2 Human Assessment of Cluster Quality
4 Conclusion
References
Graph Regression Based on Graph Autoencoders
1 Introduction
2 Related Work
2.1 Graph Embedding and Graph Regression
2.2 Autoencoders and Graph Autoencoders
2.3 Prediction of Chemical Compound Properties
3 The Method
4 Motive and Practical Application
4.1 Database
4.2 Architecture Configuration
4.3 Energy Prediction
4.4 Runtime Analysis
5 Conclusions and Future Work
References
Distributed Decision Trees
1 Introduction
2 Different Tree Architectures
2.1 Hard Decision Trees
2.2 Soft Decision Trees
2.3 Budding Trees
3 Distributed Budding Trees
4 Experiments
5 Visualization
6 Conclusions
References
A Capsule Network for Hierarchical Multi-label Image Classification
1 Introduction
2 Hierarchical Multi-label Capsules
3 Experiments
3.1 Implementation Details and Datasets
3.2 Experimental Setup
3.3 Results
4 Conclusions
References
Monte Carlo Dropout for Uncertainty Analysis and ECG Trace Image Classification
1 Introduction
2 Literature Review
2.1 ECG Classification
2.2 Uncertainty Estimation in Medical Image Analysis
3 Proposed Methodology
3.1 Dataset
3.2 Data Preprocessing
3.3 CNN Architecture
3.4 Monte Carlo Dropout
4 Experimental Results
4.1 Experimentation
4.2 Result Analysis and Discussion
5 Conclusion and Future Work
References
One-Against-All Halfplane Dichotomies
1 Introduction
2 Prior Work
3 One-Against-All Halfplane Dichotomies
4 A Geometric Perspective
5 Pairwise Attribute Difference Vectors
6 Linear Programming and Neural Networks
7 Ranking
8 Summary
References
Fast Distance Transforms in Graphs and in Gmaps
1 Introduction
1.1 Notations and Definitions
2 Distance Transform in a Graph
2.1 Geodesic Distance Transform
3 Distance Transforms in n-Gmaps
4 Results
5 Conclusions
References
Retargeted Regression Methods for Multi-label Learning
1 Introduction
2 Proposal: Retargeted Multi-label Least Square Regression
2.1 Notations
2.2 Brief Review of ReLSR
2.3 Problem Formulation
2.4 Optimization
2.5 Computational Complexity
2.6 Learning Threshold
3 Experiments
3.1 Dataset and Evaluation Measurement
3.2 Settings
3.3 Results
4 Conclusion
References
Transformer with Spatio-Temporal Representation for Video Anomaly Detection
1 Introduction
2 Related Work
2.1 Video Anomaly Detection
2.2 Transformer
3 Methodology
4 Experiments
4.1 Datasets and Evaluation Metrics
4.2 Implementation Details
4.3 Comparison with Existing State-of-the-Arts
4.4 Ablation Experiments
4.5 Running Time
5 Conclusion
References
Efficient Leave-One-Out Evaluation of Kernelized Implicit Mappings
1 Introduction
2 Kernelized Implicit Mapping
3 LOO Evaluation of KIM
4 Positioning of this Study
5 Kernels
6 Multiple Applications of LOO Matrix
6.1 Visualization and Model Selection
6.2 Nonlinear Classification
6.3 Recovery
7 Analysis of LOO Matrix
8 Discussion
9 Conclusion
References
Graph Similarity Using Tree Edit Distance
1 Introduction
2 Motivation and Basic Concepts
3 Algorithm
4 Experimental Results
5 Conclusion
References
Data Augmentation on Graphs for Table Type Classification
1 Introduction
2 Related Works
3 Method
3.1 Preprocessing
3.2 Data Augmentation
3.3 Model
4 Experiments
4.1 The Tab2Know Dataset
4.2 Using the Dataset
4.3 Results
5 Conclusions
References
Improved Training for 3D Point Cloud Classification
1 Introduction
2 Related Work
3 Method
3.1 Preliminaries
3.2 Proposed Method
3.3 Proposed Training Protocol
4 Experimental Results
4.1 Hyper-parameter Sensitivity
4.2 Results for Proposed Model Variants
4.3 Results for Explored Loss Functions
4.4 Explored Training Protocol
4.5 Effect of Augmentation
4.6 Transfer Learning
4.7 Confusion Matrix
5 Conclusions
References
On the Importance of Temporal Features in Domain Adaptation Methods for Action Recognition
1 Introduction
2 Related Works
3 Recalls Basics of an Architecture for Domain Adaptation
4 The New Designed Architecture
5 Experiments and Results
5.1 Datasets and Metrics
5.2 Parameters Setting Details
5.3 Results and Comments
6 Conclusions
References
Zero-Error Digitisation and Contextualisation of Piping and Instrumentation Diagrams Using Node Classification and Sub-graph Search
1 Introduction
2 Predicting Improperly Identified Components
2.1 Graph Representation and Data Embedding
2.2 Machine Learning and Verification
3 Searching Groups of Components
3.1 Input and Output Parameters of Our Method
3.2 Algorithm
4 Conclusions and Future Work
References
Refining AttnGAN Using Attention on Attention Network
1 Introduction
2 Background
3 Method
3.1 AttnGAN
3.2 Global Vector Attention
3.3 Attention on Attention
3.4 Mode-seeking Function
3.5 Refined-GAN Architecture
3.6 Objective Function
3.7 Implementation Details
4 Results
4.1 Datasets
4.2 Quantitative Results
4.3 Qualitative Results
5 Conclusion
References
An Autoencoding Method for Detecting Counterfeit Coins
1 Introduction
2 Selecting Similarity Criterion
2.1 Comparison of Images
2.2 Selecting a Similarity Function
3 Autoencoder Architecture
4 Training the Autoencoder
5 Custom Classifier
6 Experimental Results
6.1 Dataset
6.2 Results
7 Conclusion
References
Tarragona Graph Database for Machine Learning Based on Graphs
1 Introduction
2 Database Format
3 Database Description
3.1 Letters
3.2 Rotation Zoom
3.3 House-Hotel
3.4 Horse
3.5 Palmprint
3.6 Sagrada Familia 3D
3.7 Fingerprint
3.8 Proteins
3.9 Mutagenicity
3.10 Grec
4 Database Interaction
5 Conclusions
References
Human Description in the Wild: Description of the Scene with Ensembles of AI Models
1 Introduction
2 Related Work
3 Methods
3.1 Open Pose Estimation
3.2 Clothing Recognition by Object Detection
3.3 Clothing Color Recognition
3.4 Action Recognition
3.5 Gender and Age Classification from the Face
4 Dataset
5 Experimentation and Results
5.1 Clothing Recognition
5.2 Actions Classification
5.3 Gender and Age Classification
6 Conclusion
References
Author Index

Systemvoraussetzungen

Als PDF speichern Als Link merken