Progress in Artificial Intelligence and Pattern Recognition: 7th International Workshop on Artificial Intelligence and Pattern Recognition, IWAIPR ... Vision, Pattern Recognition, and Graphics)

✍ Scribed by Yanio Hernández Heredia (editor), Vladimir Milián Núñez (editor), José Ruiz Shulcloper (editor)

• Publisher: Springer
• Year: 2021
• Tongue: English
• Leaves: 454
• Category: Library

✦ Synopsis

This book constitutes the refereed proceedings of the 7th International Workshop on Artificial Intelligence and Pattern Recognition, IWAIPR 2021, held in Havana, Cuba, in October 2021. The 42 full papers presented were carefully reviewed and selected from 73 submissions.

The papers promote and disseminate ongoing research on mathematical methods and computing techniques for artificial intelligence and pattern recognition, in particular in bioinformatics, cognitive and humanoid vision, computer vision, image analysis and intelligent data analysis.

✦ Table of Contents

Preface
Organization
Contents
Artificial Intelligence, Data Mining and Applications
Layer-Wise Relevance Propagation in Multi-label Neural Networks to Identify COVID-19 Associated Coinfections
1 Introduction
2 LRP to Explain Multi-label Neural Networks
2.1 LRP
2.2 LRP in Multi-label Scenarios
3 Identifying COVID-19 Associated Coinfections
3.1 Dataset Under Consideration
3.2 Solution Based on Multi-label Neural Networks
3.3 Applying the LRPml Algorithm to the Proposed Neural Network
4 Concluding Remarks
References
A Machine Learning Based Approach for Estimation of the Lung Affectation Degree in CXR Images of COVID-19 Patients
1 Introduction
2 Related Works
3 Method for Automatic Estimation of Lung Affectation Degree by COVID-19 in CXR Images
3.1 Supervised Classification of the Lung Image and Calculation of the Index of Affectation
4 Experimental Results and Discussion
5 Conclusions and Future Work
References
Deep Neural Network to Detect Gender Violence on Mexican Tweets
1 Introduction
2 Related Work
3 Theoretical Framework
4 Methodology
5 Experimental Results and Discussion
6 Conclusions and Future Work
References
Consumer Price Index Forecasting Based on Univariate Time Series and a Deep Neural Network
1 Introduction
2 General Notation and State of the Art
2.1 LSTM in Time Series Forecasting
2.2 Conventional LSTM Architectures
3 Results and Discussion
4 Concluding Remarks and Further Work
References
Application of a Generalized Regression Neural Network to the Estimation of Average Gait Speed
1 Introduction
1.1 Inertial Measurement Unit (IMU)
2 Generalized Regression Neural Networks (GRNN)
2.1 Route Optimization Algorithm
3 Materials and Methods
4 Results and Discussion
5 Conclusions
References
Gait Speed Estimation Based on Artificial Neural Network: Comparison with the Application of the Inverted Pendulum Model
1 Introduction
2 Related Work
3 Methods and Materials
4 Results
5 Conclusions
References
Scalable Generalized Multitarget Linear Regression With Output Dependence Estimation
1 Introduction
2 Big Data: Concepts and Frameworks
3 Multitarget Regression
4 Scalable MTR with Output Dependence Estimation
5 Experiments and Results
5.1 Convergence Experiments
5.2 Scalability Experiments
6 Conclusions and Further Work
References
Covariance Matrix Adaptation Evolution Strategy for Convolutional Neural Network in Text Classification
1 Introduction
2 Computational Methodology
2.1 Design of Convolutional Neural Network for the Text Categorization
2.2 Formalization of the Problem and Optimization by Proposed Method
2.3 Dataset
3 Results and Discussion
4 Conclusions and Further Work
References
A Novel Approach for Detection and Location of Cyber-Attacks in Water Distribution Networks
1 Introduction
2 Materials and Methods
2.1 Adaptive Exponential Weighted Moving Average Chart
2.2 Autoencoders
2.3 Structural Analysis
3 Methodology for Detection and Location of Cyber-Attacks
3.1 Offline and Online Stages
4 Application of the Proposed Methodology to the C-Town Case Study
4.1 Case Study: C-Town WDN and BATADAL Datasets
4.2 Detection Module in the Offline Stage
4.3 Location Module in the Offline Stage
4.4 Performance Assessment
5 Results and Discussion
6 Conclusions
References
Identification of Non-linear Chemical Systems with Neural Networks
1 Introduction
2 TE Process
3 Methodology
3.1 Experiment Design
3.2 Data Collection and Pretreatment
3.3 Selection of the Type and Structure of the Model
3.4 Estimation of Parameters, Application of Algorithms
3.5 Validation
4 Conclusions and Future Work
References
Convolutional Neural Networks as Support Tools for Spinocerebellar Ataxia Detection from Magnetic Resonances
1 Introduction
1.1 Spinocerebellar Ataxias
1.2 Convolutional Neural Networks
2 Materials and Methods
2.1 Image Preparation
2.2 Analysis Description
2.3 Metrics
3 Results and Discussion
4 Conclusions
References
A Hybrid Brain-Computer Interface using Extreme Learning Machines for Motor Intention Detection
1 Introduction
2 Materials and Methods
2.1 Dataset
2.2 Proposed HBCI
2.3 Evaluation
3 Results and Discussion
4 Conclusions
References
Peripheral Nerve Segmentation in Ultrasound Images Using Conditioned U-Net
1 Introduction
2 Materials and Methods
2.1 Dataset and Data Augmentation
2.2 Conditioned U-Net for Nerve Segmentation
2.3 Training Procedure
2.4 Validation Approach
3 Results and Discussion
4 Concluding Remarks and Future Work
References
Assessing the Relationship Between Binarization and OCR in the Context of Deep Learning-Based ID Document Analysis
1 Introduction
2 Background
3 Algorithm Selection
4 Proposed Pipeline Analysis and Capture Conditions
5 Evaluation
6 Conclusions
References
Pattern Recognition and Applications
Pattern Recognition Strategies to Classify Traced Neurons
1 Introduction
2 Materials and Methods
2.1 Data Bases
2.2 Methods and Tools to Extract and Select Features
2.3 Pattern Recognition Strategies
2.4 Classification Process
3 Results
3.1 Results of Classification into Neuron Types Using the First Pattern Recognition Strategy
3.2 Results of Classification into Neuron Types Using the Second Pattern Recognition Strategy
4 Discussion
5 Conclusion
References
Evaluation of Hepatic Fibrosis Stages Using the Logical Combinatorial Approach
1 Introduction
2 Mathematical Model
2.1 Voting Algorithm
2.2 Metavir Scoring System
3 Mathematical Model Applied to Hepatic Fibrosis Diagnostic
3.1 Variables
3.2 Comparison Criteria
3.3 Similarity Function
3.4 Fibrosis Hepatic Evaluation Using the Metavir Score as Classification Criterion
3.5 Fibrosis Hepatic Evaluation Using Fibrosis Percentage as Classification Criterion
4 Conclusion
References
A Proposal of Condition Monitoring with Missing Data and Small-Magnitude Faults in Industrial Plants
1 Introduction
2 Materials and Methods
2.1 Description of the Proposed Methodology
2.2 Case of Study: Tennessee Eastman Process
3 Experimental Results
3.1 Classification of Incomplete Observations
3.2 Results of the Imputation and Detection Stages
4 Conclusions
References
A Proposal of Metric for Improving Remaining Useful Life Prediction in Industrial Systems
1 Introduction
2 Preprocessing
3 Performance Metrics
3.1 Proposed Performance Metric
4 Prognostic Model and Model Training
5 Dataset Description
5.1 Preprocessing
6 Results
7 Conclusions and Future Work
References
A Fuzzy Logic Proposal for Diagnosis Multiple Incipient Faults in a Power Transformer
1 Introduction
2 Materials and Methods
2.1 Analysis of the Main Electrical Interruptions that Occurred in the Steel Mill
2.2 Description of the Design Proposal
3 Results and Discussion
4 Conclusions
References
A Robust Fault Diagnosis Strategy in Mechanical Systems Using Pythagorean Fuzzy Sets
1 Introduction
2 Materials and Methods
2.1 Intuitionistic Fuzzy C-Means Algorithm
2.2 Pythagorean Fuzzy C-Means Algorithm (PyFCM)
2.3 Pythagorean Fuzzy C-Means Algorithm Based on a Kernel Approach
3 Description of the Proposal
4 Benchmark Case Study: DAMADICS
5 Analysis of Results
5.1 Statistical Tests
6 Conclusions
References
Prediction of Diabetes Mortality in Mexico City Applying Data Science
1 Introduction
2 Data Science Methodology
2.1 Source of Data
2.2 Data Warehouse Description
2.3 Data Model
3 Results and Discussion
4 Conclusions
References
Balancing of Samples in Class Hierarchy
1 Introduction
2 Related Works
2.1 Class Decision Trees (CDTs)
3 Sample Balanced Class Hierarchy (SBCH)
3.1 Evaluation of Class Set Pairs
3.2 Max-Min Algorithm
4 Experiments
4.1 Results
5 Discussion
6 Conclusion
References
Feature Selection with Class Hierarchy for Imbalance Problems
1 Introduction
2 Related Works
2.1 Algorithms Against Class Imbalance
2.2 Class Hierarchy
3 Construction of Class Hierarchy
3.1 Class Decision Tree (CDT)
3.2 Proposed Method: Feature Selection Based CDT
4 Experiments
4.1 Experimental Method
4.2 Data Set
4.3 Class Hierarchy
4.4 Over-All Accuracy and Balanced Accuracy
5 Discussion
6 Conclusion
References
SVM Based EVM for Open Space Problems
1 Introduction
2 Related Work
2.1 Algorithm for Open Space Recognition
2.2 EVT Based Algorithm
3 SVM Based EVM
3.1 Algorithm of SVM-EVM
3.2 An Achievement of SVM-EVM
3.3 Solving Three Problems
4 Experiments
4.1 Experimental Procedure
4.2 Result
5 Discussion
6 Conclusion
References
Data Drive Fuzzy Cognitive Map for Classification Problems
1 Introduction
2 Theoretical Background
2.1 Fuzzy Cognitive Maps
2.2 Causality and Statistical Association Analysis
3 Data Drive FCM Construction
3.1 Concepts Definition
3.2 Relation Value Definition
4 Experimental Results
5 Conclusions
References
Extracting Composite Summaries from Qualitative Data
1 Introduction
1.1 Some Preliminary Concepts
2 Extracting Composite Summaries from Qualitative Data
3 Evidence Composite Relation
3.1 Contrast Composite Relation
3.2 Emphasis Composite Relation
4 Use Case
5 Conclusions
References
A Novel Initial Population Construction Heuristic for the DINOS Subgroup Discovery Algorithm
1 Introduction
2 Background
3 Initial Population Construction Heuristic Based on Random Decision Forests
4 Experimental Results
5 Conclusions
References
A Survey on the Methods to Determine the Sensitivity of Textual Documents: Solutions and Problems to Solve
1 Introduction
2 Systems for Data Leakage Prevention (DLP)
3 Analysis of Methods for Determining the Sensitivity of Documents: Advantages and Limitations
3.1 Context Analysis
3.2 Content Analysis
4 Problems to Be Solved
5 Conclusions
References
Automatic Classification of Diabetic Foot Ulcers Using Computer Vision Techniques
1 Introduction
2 Materials and Methods
2.1 Image Sets Used
2.2 Image Set Curation
2.3 Feature Extraction Strategies
2.4 Classification Algorithms and Performance Measures
3 Results and Discussion
4 Conclusions
References
Biometrics, Image, and Video Analysis
Image Segmentation Applied to Line Separation and Determination of GPN2 Protein Overexpression for Its Detection in Polyacrylamide Gels
1 Introduction
2 Methodology
2.1 Obtaining Gels
2.2 Obtaining Images
2.3 Image Pre-processing
2.4 Line Detection
2.5 Band Detection
3 Results and Discussion
4 Conclusion
References
Underwater Image Enhancement Using Adaptive Algorithms
1 Introduction
2 Related Work
3 Methods and Techniques
3.1 Contrast Limited Adaptive Histogram Equalization
3.2 Calculate Dehaze Factor
3.3 Dehazing by Adaptive Dark Channel Prior
4 Results and Discussion
4.1 Analysis of Results
4.2 Quantitative Analysis of Results
5 Conclusion
References
Improving the Robustness of DCT-Based Handwritten Document Image Watermarking Against JPEG-Compression
1 Introduction
2 Preliminaries
2.1 Quantization Index Modulation (QIM)
3 Watermarking Techniques
3.1 Scheme Used
3.2 Quantification Formulas Optimization
4 Experiments and Results
5 Conclusions
References
Automated System for the Detection of Lung Nodules
1 Introduction
2 Material and Methods
3 Results and Discussion
4 Conclusions
References
Video and Image Complexity in Human Action Recognition
1 Introduction
2 Related Work
3 Proposed Method
3.1 Image Statistics
3.2 Mask-RCNN
3.3 Correlation Metric
3.4 Classification Algorithms
4 Experiments and Results
4.1 Dataset
4.2 Complexity from Image Statistics
4.3 Camera Movement Influence
4.4 Types of Activities
5 Conclusions and Future Work
References
A Wide 2D Median Filter forGPU Devices
1 Introduction
2 Methods and Materials
3 Results
4 Conclusions
References
Medical Image Watermarking with Separable Moments
1 Introduction
2 Orthogonal Image Moments
3 Proposed Method
3.1 Optimization Process
3.2 Watermarking Scheme
4 Experimental Study
4.1 Robustness Evaluation
4.2 Discusion
References
Signals Analysis and Processing
Intra-day Population Fluxes from Mobile Phone Data in Havana, Cuba
1 Introduction
2 Mobile Phone Data and Human Mobility
3 Trip Identification
3.1 Frequent Night and Day Locations as Proxies for Home and Work
3.2 Trips and Trips to Work
4 Understanding the Fluxes
4.1 Validation with Job Survey
5 Conclusions
References
Semantic Segmentation of Radio-Astronomical Images
1 Introduction
2 Related Work
3 Semantic Segmentation
4 Experiments
4.1 Dataset
4.2 Architecture and Training Details
4.3 Results
5 Conclusion
References
Content-Based Image Retrieval for Surface Defects of Hot Rolled Steel Strip Using Wavelet-Based LBP
1 Introduction
2 Content Based Image Retrieval
3 Local Binary Pattern
4 Image Retrieval Using Wavelet Based Local Binary Patterns
5 Experimental Results
5.1 Similarity Measurement
5.2 Performance Evaluation Metrics
5.3 Retrieval Results
6 Conclusion
References
Decision Rules for Radar Detection in the Moments Space with Arbitrary Correlation Degree and Constant False Alarm Rate
1 Introduction
2 The Moments Space for Radar Detection
3 Optimum Decision Rule for Moments with Arbitrary Correlation Degree
3.1 Mahalanobis Distance and Its Relation to Optimal Detection
3.2 Probability Density Function of the Sufficient Decision Statistic
4 CFAR Decision Rule in the Moments Space
5 Conclusions
References
Interpretable Diagnosis of ADHD Based on Wavelet Features and Logistic Regression
1 Introduction
2 Materials and Methods
2.1 Dataset and Preprocessing
2.2 DWT-Based Feature Extraction
2.3 Interpretable Classification Through Logistic Regression
3 Results
3.1 Parameter Tuning
3.2 Performance Analysis
4 Discussion and Concluding Remarks
References
Evaluation of GOI Detectors in EGG Signals Assuming Different Models for the Pulse Length Variability
1 Introduction
1.1 The Glottal Pulse
1.2 Objectives
2 Materials and Methods
2.1 The Closure and Opening Instants Detection on EGG Signals
2.2 GOI Detectors Evaluated
2.3 Datasets
2.4 Evaluation Procedure: Assumptions and Measures Considered
3 Results and Discussion
4 Conclusions
References
Author Index