Pattern Recognition and Artificial Intelligence

Name: Pattern Recognition and Artificial Intelligence | Third International Conference, ICPRAI 2022, Paris, France, June 1-3, 2022, Proceedings, Part II
Brand: Springer
Price: 149.79 EUR
Availability: OnlineOnly

Third International Conference, ICPRAI 2022, Paris, France, June 1-3, 2022, Proceedings, Part II

Mounîm El Yacoubi Eric Granger Pong Chi Yuen Umapada Pal Nicole Vincent(Editor)

Springer (Publisher)

Published on 28. May 2022

XXII, 520 pages

E-Book

PDF with digital watermarking

System requirements

978-3-031-09282-4 (ISBN)

€149.79incl. 7% vat

System requirements

for PDF with digital watermarking

E-Book Single Licence

Available for download

Description

More details

Other editions

Content

Intro
Preface
Organization
Contents - Part II
Contents - Part I
Pattern Recognition
PE-former: Pose Estimation Transformer
1 Introduction
2 Related Work
2.1 Vision Transformers
2.2 Transformers for Pose Estimation
3 Method
3.1 Transformer Encoder
3.2 Transformer Decoder
3.3 Training
4 Experiments
4.1 Evaluation Methodology
4.2 Comparison with SOTA
4.3 Transformer vs CNN vs VAB Encoders
4.4 Unsupervised Pre-training
5 Conclusions
References
ConDense: Multiple Additional Dense Layers with Fine-Grained Fully-Connected Layer Optimisation for Fingerprint Recognition
1 Problem Background
2 Related Work
3 Methods
4 Experiment Data
5 Experiment Setup
6 Experiment Results
6.1 Ablation Study
6.2 Architecture Performance
6.3 Comparison to Related Work
7 Conclusion
References
A Sensor-Independent Multimodal Fusion Scheme for Human Activity Recognition
1 Introduction
2 Related Work
3 Methods
3.1 Sensor Independent Fusion Model
3.2 Data Augmentation Method
4 Evaluation
5 Conclusions and Future Work
References
Comparative Study of Activation Functions and Their Impact on the YOLOv5 Object Detection Model
1 Introduction
2 Related Work
3 Methodology
3.1 YOLOv5 Network Architecture
3.2 Activation Functions
4 Experiments
4.1 Datasets
4.2 YOLOv5 Model Selection
4.3 Experimental Setup
4.4 Activation Function Comparative Analysis
4.5 Model Verification Using Transfer Learning
5 Conclusion
References
Application of A* to the Generalized Constrained Longest Common Subsequence Problem with Many Pattern Strings
1 Introduction
2 The State Graph for GCLCS Problem
3 A* Search for the GCLCS Problem
4 Experimental Evaluation
5 Conclusions and Future Work
References
Robot Path Planning Method Based on Improved Grey Wolf Optimizer
1 Introduction
2 Grey Wolf Optimizer
3 An Improved GWO Based on Random Walk and Dynamic Programming (RWDPGWO) to Solve the Robot Path Planning Problem
3.1 The Two Algorithms are Mixed to Quickly Find the Initial Path
3.2 Random Walk Nonlinear Adjustment Step Size Factor to Improve Search Efficiency
3.3 Improve Search Strategy Based on DP Algorithm Idea
3.4 Use Spline Curve B-spline Curve to Optimize the Path
3.5 Algorithm Flow
4 Simulation
4.1 Comparison of RWDPGWO and GWO Simulation
4.2 Comparison of RWDPGWO with GA and PSO
4.3 Comparison of RWDPGWO and Various Improved ACO
5 Conclusion
References
Space-Time Memory Networks for Multi-person Skeleton Body Part Detection
1 Introduction
2 STM Multi-person Skeleton Body Part Detection
2.1 STM Architecture
2.2 STM-skeletons Architecture
2.3 Training
3 Results
3.1 Video Skeleton Segmentation
3.2 Video Skeleton Edge Prediction
3.3 Video Pose Estimation
4 Conclusion
References
Lip-Based Identification Using YOLOR
1 Introduction
2 Literature Study
2.1 Problem Background
2.2 Similar Work
3 Experiment Setup
3.1 Dataset Selection
3.2 Proposed Architecture
4 Results
5 Discussion
6 Conclusion
References
Parallel O(log(n)) Computation of the Adjacency of Connected Components
1 Introduction
1.1 Motivations and Notations
2 The RtC Algorithm for Pyramid Construction
2.1 Edge Classification
2.2 Selecting the Contraction Kernel
2.3 Redundant Edges
2.4 Parallel Pyramidal Connected Component (//ACC)
3 Parallel Complexity
4 Comparisons and Results
5 Conclusions
References
Encoding Sensors' Data into Images to Improve the Activity Recognition in Post Stroke Rehabilitation Assessment
1 Introduction
2 Dataset Description and Preparation
3 Encoding Pipeline
3.1 Encoding Techniques
3.2 Image Fusion and Interpolation
3.3 Experimental Results
4 Conclusion
References
Momentum Residual Embedding with Angular Marginal Loss for Plant Pathogen Biometrics
1 Introduction
2 Related Work
3 Research Methodology
3.1 Backbone Network
3.2 Momentum
3.3 Loss Function
3.4 Training and Implementation
4 Experimental Results
4.1 Datasets Description
4.2 Results and Comparison
4.3 Qualitative Analysis
4.4 Ablation Study
4.5 Discussion
5 Conclusion and Future Works
References
Classification
Hierarchical Approach for the Classification of Multi-class Skin Lesions Based on Deep Convolutional Neural Networks
1 Introduction
2 Methodology
2.1 Convolutional Neural Networks
3 Experimental Results and Discussion
3.1 Dataset
3.2 Evaluation Metrics
3.3 Results and Discussion
4 Conclusion
References
The FreshPRINCE: A Simple Transformation Based Pipeline Time Series Classifier
1 Introduction
2 Background
2.1 State of the Art for TSC
2.2 Unsupervised Time Series Transformations
3 Experimental Structure
4 Results
4.1 Implementation and Reproduction of Results
5 Conclusion
References
Robust Detection of Conversational Groups Using a Voting Scheme and a Memory Process
1 Introduction
2 Related Work
3 Proposed Approach: Multiple Votes and Exploiting Temporal Information
3.1 Voting in Each Frame
3.2 Increasing Robustness by Taking into Account Temporal Memory
4 Experiments and Results
4.1 MatchNMingle Data Set
4.2 Evaluation Criteria
4.3 Parameter Setting
4.4 Results
5 Conclusion
References
Fusing AutoML Models: A Case Study in Medical Image Classification
1 Introduction
2 Background
2.1 Pattern Recognition with AutoML
2.2 Fusion
3 Datasets
3.1 MSU Stem Cell Dataset
3.2 Brain Tumor Classification
3.3 Prostate Cancer Classification
4 Proposed Approach
4.1 Data Extraction
4.2 Data Descriptors Through Feature-Sets
4.3 Model Generation
4.4 Model Scores
4.5 Model Selection for Fusion
5 Results and Analysis
6 Summary
References
Ordinal Classification and Regression Techniques for Distinguishing Neutrophilic Cell Maturity Stages in Human Bone Marrow
1 Introduction
2 Materials and Methods
2.1 Image Data
2.2 Classification Techniques
2.3 Regression Techniques
2.4 Experimental Setup
3 Results
4 Discussion
5 Conclusion
References
Towards Automated Monitoring of Parkinson's Disease Following Drug Treatment
1 Introduction
1.1 Learning Algorithms
1.2 rs-fMRI Data
2 Methods
2.1 Overview
2.2 Participants
2.3 Procedure
2.4 rs-fMRI Acquisition
2.5 Imaging Data Analysis
2.6 Cartesian Genetic Programming
2.7 Adaptive Synthetic Sampling
2.8 k-Fold Cross-Validation
3 Results
3.1 Classification of Timeseries
3.2 Classification of Dynamic Causal Modeling (DCM)
3.3 k-Fold Cross-Validation
4 Conclusion
References
A Hierarchical Prototypical Network for Few-Shot Remote Sensing Scene Classification
1 Introduction
2 Few-Shot Classification with Prototype Learning
2.1 Problem Formulation of the FSL
2.2 Prototypical Networks
3 A Hierarchical Prototypical Network for Few-Shot Image Classification
3.1 Overall Framework
3.2 Hierarchical Prototypical Network
4 Few-Shot Learning for Remote Sensing Scene Classification
4.1 Dataset Description
4.2 Implementation Details
4.3 Evaluation Metrics
4.4 Experimental Results
5 Conclusion and Future Works
References
TS-QUAD: A Smaller Elastic Ensemble for Time Series Classification with No Reduction in Accuracy
1 Introduction
2 Background and Related Work
2.1 Classification in the Time Domain
2.2 The Elastic Ensemble (EE) and Extensions
3 EE with Fewer Constituents: TS-QUAD
4 Experimental Procedure
5 Results
6 Conclusions, Future Work and Extensions
References
Machine Learning
Shop Signboards Detection Using the ShoS Dataset
1 Introduction
2 Related Work
2.1 Street View Imagery Object Detection
2.2 Storefront Dataset
3 The ShoS Dataset
3.1 Data Collection
3.2 Data Annotation
3.3 Challenges and Limitation
4 Experiments
5 Results and Discussion
6 Conclusion
References
One-Shot Decoupled Face Reenactment with Vision Transformer
1 Introduction
2 Methods
2.1 Facial Feature Extractor and Optical Flow Estimation
2.2 Landmark Reenactment Module
2.3 Face Reenactment Module
2.4 Loss Function
3 Experiments
3.1 Model Variants
3.2 Metrics
3.3 Analysis
4 Conclusions
References
ADG-Pose: Automated Dataset Generation for Real-World Human Pose Estimation
1 Introduction
2 Related Work
3 Real-World Pose Estimation Challenges
4 ADG-Pose
5 Results and Evaluation
6 Conclusion
References
Deep Reinforcement Learning for Autonomous Navigation in Robotic Wheelchairs
1 Introduction
2 Related Work
3 Methodology
3.1 MDP Formulation
3.2 Solving the MDP
4 Experiments and Results
4.1 Training
4.2 Evaluation
5 Conclusions
References
Understanding Individual Neurons of ResNet Through Improved Compositional Formulas
1 Introduction
2 Related Work
3 Algorithmic Compositional Explanations
3.1 Setup
3.2 Connecting Annotated Concepts
4 Locality and the ImRoU Metric
5 Results
5.1 Scene-Level Annotations
5.2 Scores
6 User Study
6.1 Study Design
6.2 Results
6.3 Discussion
References
Visual Transformer-Based Models: A Survey
1 Introduction
2 Transformer
2.1 Transformer Architecture
2.2 Attention
2.3 Other Components
3 ViT (Vision Transformer)
4 Transformer-Based Models
4.1 ViT Variants for Image Classification
4.2 Transformer-Based General Backbone
5 Discussions and Conclusions
References
Stochastic Pairing for Contrastive Anomaly Detection on Time Series
1 Introduction
2 Related Works
2.1 Reconstruction Based Time Series Anomaly Detection
2.2 Self Supervised Contrastive Learning
2.3 Time Series Augmentation
3 Our Approach
3.1 Preprocessing and Stochastic Pairing
3.2 Shape-Dynamic Time Warping
3.3 Augmentation Methods
3.4 Learning Architecture
3.5 Anomaly Score
3.6 Detection Method
4 Experiments on Public Datasets
5 Effects of Parameter Settings
5.1 Augmentation Techniques
5.2 Batch Parameters
6 Conclusion
References
Visual Radial Basis Q-Network
1 Introduction
2 Related Works
3 Background
3.1 Reinforcement Learning
3.2 Radial Basis Function Network (RBFN)
4 Method
4.1 Selection of RBF Parameters
4.2 Vizdoom Scenarios
5 Analysis of Pattern Activations
6 Reinforcement Learning Application
7 Conclusion
References
GANs Based Conditional Aerial Images Generation for Imbalanced Learning
1 Introduction
2 Methods
2.1 Datasets
2.2 Generative Adversarial Networks
3 Evaluation
4 Results and Discussion
4.1 Data Preprocessing
4.2 Generation and Evaluation
5 Conclusion
References
Augment Small Training Sets Using Matching-Graphs
1 Introduction and Related Work
2 Basic Definitions
3 Augment Training Sets by Means of Matching-Graphs
4 Experimental Evaluation
4.1 Experimental Setup
4.2 Data Sets
4.3 Validation of Metaparameters
4.4 Test Results and Discussion
5 Conclusion and Future Work
References
Progressive Clustering: An Unsupervised Approach Towards Continual Knowledge Acquisition of Incremental Data
1 Introduction
2 Related Works
3 Progressive Clustering
3.1 Incremental Data Generation for Progressive Clustering
3.2 Latent Space Representation
3.3 Initial Clustering
3.4 Proximity Clustering
3.5 Detection of Concept Drift
3.6 Collateral Clustering
3.7 Deep Dynamic Incremental Classification
4 Experimental Setup
4.1 About the Datasets
4.2 Evaluation Metrics
5 Results and Discussions
5.1 Evaluation Methodology
5.2 Qualtitative Study
6 Conclusion
References
Malware Detection Using Pseudo Semi-Supervised Learning
1 Introduction
2 Related Work
2.1 Semi-Supervised Learning
2.2 Pseudo-Labeling
3 Proposed Framework
3.1 Model
3.2 Regularization
4 Experiments
4.1 Dataset
4.2 Comparison Methods
4.3 Training and Evaluation
5 Results and Discussion
6 Conclusion
References
Information Extraction
Temporal Disaggregation of the Cumulative Grass Growth
1 Introduction
2 Data
3 Methodology
3.1 Formalization and General Approach
3.2 Pre and Post Processings
3.3 Order of Models and Initialization
4 Experiments and Results
4.1 Comparison of the Prediction Models
4.2 Effect of Approximating the First Values
4.3 Improvements by Post-processings
4.4 Qualitative Evaluation of Reconstructions
5 Related Work
6 Conclusion
References
Extraction of Entities in Health Domain Documents Using Recurrent Neural Networks
1 Introduction
2 Related Work
3 Proposed Solution
3.1 Information Pre-processing
3.2 Identification of Entities
4 Results
4.1 Description of the Corpus
4.2 Pre-processing Results
4.3 Results: Word Embedding
4.4 System Results for Entity Identification
4.5 Evaluation of the Proposed Systems
4.6 Results of the Systems Proposed in eHealth-KD Challenge 2021
5 Conclusion
References
An Overview of Methods and Tools for Extraction of Knowledge for COVID-19 from Knowledge Graphs
1 Introduction
2 COVID-19 Knowledge Graphs and Search Engines
2.1 Search Engines in Internet for COVID-19 Information
2.2 Introduction of Leading Knowledge Graphs for COVID-19 and Data Sets
3 Definition of Constraints When Building COVID-19 Knowledge Graph
3.1 COVID-19 Data Constraints
3.2 Respect the Quality of the Data When Constructing Knowledge Graph
3.3 Scale Constraints in Knowledge Graph Construction
3.4 Constraints Defined from Interface Presentation Tools and Visualizations
3.5 Social Constraints
4 Introduction to COVID-19 Reasoning from KGs
4.1 Definition of Knowledge Reasoning Applied for Extraction of COVID-19 from KGs ch90refspsarticle14
4.2 General Model for COVID-19 Knowledge Reasoning from KGs
4.3 Classification of Knowledge Reasoning Oriented for COVID-19 Entity Extraction ch90refspsarticle14
5 Conclusions
References
Explaining Image Classifications with Near Misses, Near Hits and Prototypes
1 Introduction
2 Methodology
2.1 Data Sets
2.2 Models and Embeddings
2.3 Architecture Overview
3 Prototype Selection
3.1 Prototype Selection Using Maximum Mean Discrepancy
3.2 Parameter Selection
3.3 Evaluation
4 Near Miss and Hit Selection
4.1 Evaluation
5 Demonstrator
6 Conclusion and Future Work
References
Adaptive Threshold for Anomaly Detection in ATM Radar Data Streams
1 Introduction
2 Related Work
3 Method
3.1 VPOT Approach
3.2 Methodology
4 Experimental Assessment
4.1 Experimental Protocol
4.2 Data Set
4.3 Benchmarking
4.4 Observation
4.5 Discussion
5 Conclusion and Future Work
References
Covid-19 Vaccine Sentiment Analysis During Second Wave in India by Transfer Learning Using XLNet
1 Introduction
2 Methodology
3 Experimentation and Results
3.1 Dataset Preparation
3.2 Results
4 Conclusion
References
Improving Drift Detection by Monitoring Shapley Loss Values
1 Introduction
2 Related Work
2.1 Drift Detection
2.2 Shapley Values
2.3 Shapley Values for Machine Learning
2.4 Shapley Loss Values
3 Shap-ADWIN: Drift Detection on Shapley Values
3.1 Intuition
3.2 Mathematical Foundations
4 Experimental Results
4.1 Background Dataset
4.2 Influence of Noise on Drift Detection
4.3 Influence of Sensitivity of the Detector
5 Conclusion
References
Interpolation Kernel Machine and Indefinite Kernel Methods for Graph Classification
1 Introduction
2 Graph Kernels
3 Interpolation Kernel Machines
4 Indefinite Kernel Methods
4.1 Need of Indefinite Kernel Methods
4.2 Indefinite Interpolation Kernel Machines
4.3 Indefinite Support Vector Machines
5 Experimental Results
6 Extended Experimental Protocol
7 Conclusion
References
DRN: Detection and Removal of Noisy Instances with Self Organizing Map
1 Introduction
2 Related Work
3 DRN: Our Proposed Framework
3.1 Noise and Outliers in Learning
3.2 Self-organizing Map and Outlier
3.3 DRN
4 Experiments
4.1 Experimental Methodology
4.2 Results
5 Conclusions
References
Informativeness in Twitter Textual Contents for Farmer-centric Plant Health Monitoring
1 Introduction
2 Use Cases and Data Collection
3 Histogram by Mention of Keywords
4 Processing Tweets for Natural Hazard Detection
4.1 Topic Detection Based on Bag of Word Models
4.2 Text Classification Based on Pre-trained Language Models
5 Conclusion
References
A Deep Learning Approach to Detect Ventilatory Over-Assistance
1 Introduction
2 Related Work
3 Methodology
3.1 Data Acquisition and Filtering
3.2 Automatic Breath Segmentation
3.3 Automatic Breath Labeling
3.4 Breath Classification
4 Experimental Results
4.1 Experimental Setup
4.2 Breath Segmentation Evaluation
4.3 Classification Assessment
5 Discussion and Future Work
References
Author Index

System requirements

Save as PDF Copy link into clipboard

Schweitzer Fachinformationen

Pattern Recognition and Artificial Intelligence

Description

More details

Other editions

Additional editions

Content

System requirements