Pattern Recognition: 15th Mexican Conference, MCPR 2023, Tepic, Mexico, June 21–24, 2023, Proceedings (Lecture Notes in Computer Science)

Preface
Organization
Contents
Pattern Recognition and Machine Learning Techniques
Feature Analysis and Selection for Water Stream Modeling*-12pt
	1 Introduction
	2 Data and Methods
		2.1 Dataset
		2.2 Data Preparation
		2.3 Machine Learning Models
	3 Experiments and Results
		3.1 Attribute Selection
	4 Conclusions and Future Work
	References
A Cloud-Based (AWS) Machine Learning Solution to Predict Account Receivables in a Financial Institution
	1 Introduction
		1.1 Problem Definition
	2 Literature Review
	3 Materials and Methods
		3.1 Account Receivables Variables Description
	4 Experiments and Results
		4.1 Architecture of the Solution
	5 Conclusions
	References
A New Approach for Road Type Classification Using Multi-stage Graph Embedding Method*-12pt
	1 Introduction
	2 Background and Related Work
	3 Materials and Methods
		3.1 Input Dataset and Line Graph Transformation
		3.2 Road Type Class Labels
		3.3 Feature Engineering
		3.4 Multi-stage Graph Embedding
		3.5 Classification with MLP Classifier
	4 Experimental Results
		4.1 Stage 1: Graph Embedding with Deep AutoEncoder
		4.2 Stage 2. Embedding with Graph Convolution Neural Network
		4.3 Comparison to Other Methods
	5 Conclusion
	References
Removing the Black-Box from Machine Learning
	1 Introduction
	2 Universal Approximation Theorem for Polynomials
	3 Approximation Using Genetic Algorithms
		3.1 The Ascent Algorithm
		3.2 Implementation of the Ascent Algorithm
	4 Determining the Best Number of Terms in the Approximant
	5 Interpretation of a Model from Experimental Data
	References
Using Machine Learning to Identify Patterns in Learner-Submitted Code for the Purpose of Assessment
	1 Introduction
	2 Related Work
	3 Binary Classification
	4 What Did the Binary Classifiers Learn?
		4.1 Methodology
		4.2 Results
	5 Multiclass Prediction
		5.1 Methodology
		5.2 Results
	6 Conclusion and Further Work
	References
Fitness Function Comparison for Unsupervised Feature Selection with Permutational-Based Differential Evolution
	1 Introduction
	2 Permutational-Based Differential Evolution
	3 Proposal Implementation
	4 Internal Metrics as Fitness Functions
		4.1 Silhouette Coefficient
		4.2 Calinski-Harabazs Index
		4.3 Davies-Bouldin Score
	5 Experiments and Results
	6 Conclusions and Future Work
	References
A Method for Counting Models on Cubic Boolean Formulas
	1 Introduction
	2 Preliminaries
		2.1 Cubic-Hamiltonian Graphs
		2.2 Conjunctive Normal Form
		2.3 The Restricted Graph of a 2-CNF
		2.4 Methods Already Reported to Compute #2SAT
	3 Counting Models on Cubic Formulas
	4 Planar and Non-planar Cubic Graphs
	5 Complexity Analysis
	6 Conclusions
	References
Automatic Identification of Learning Styles Through Behavioral Patterns
	1 Introduction
	2 Related Work
	3 Proposed Method
		3.1 Detection of Learning Style Using Genetic Algorithm
		3.2 Fitness Function
	4 Experimentation and Results
	5 Conclusions
	References
Comparison of Classifiers in Challenge Scheme*-12pt
	1 Introduction
	2 Collaborative Competitions
	3 Dataset - MeOffendEs@IberLEF 2021
	4 Proposed Approaches and Results
		4.1 Bootstrap
		4.2 Comparison of Classifiers
		4.3 Comparison of Classifiers Through Independents Samples
		4.4 Comparison of Classifiers Through Paired Samples
		4.5 Statistical Significance Testing
	5 Conclusions
	References
Deep Learning and Neural Networks
Robust Zero-Watermarking for Medical Images Based on Deep Learning Feature Extraction
	1 Introduction
	2 Methodology
		2.1 Context Encoder
		2.2 Watermark Sequence
	3 Experiment Results
	4 Conclusions
	References
Plant Stress Recognition Using Deep Learning and 3D Reconstruction
	1 Introduction
	2 The Proposed Methodology
		2.1 Plant Recognition
		2.2 Leaf Detection
		2.3 3D Model Analysis
	3 Discussion and Results
		3.1 Semantic Segmentation
		3.2 3D Model
		3.3 3D Model vs 2D Classification
	4 Conclusions
	References
Segmentation and Classification Networks for Corn/Weed Detection Under Excessive Field Variabilities
	1 Introduction
		1.1 Related Works
	2 Development of the Proposed Method
		2.1 Dataset Description and Image Pre-processing
		2.2 Training of the Architectures
		2.3 Evaluation Metrics
	3 Results
		3.1 Performance of the Segmentation Network
		3.2 Performance of the Classification Network
		3.3 Detection Approach Visualization
	4 Conclusions and Future Work
	References
Leukocyte Recognition Using a Modified AlexNet and Image to Image GAN Data Augmentation
	1 Introduction
	2 Dataset Description
		2.1 Detection Stage
		2.2 Crop Stage
		2.3 Data Augmentation
	3 Classification Model
		3.1 Deep Model Selection for Leukocytes Classification
		3.2 Implementation of the Selected Deep Models with Transfer Learning
		3.3 Hyperparameters of the Selected Models
	4 Results
		4.1 Classification Model Selection
		4.2 Results of GAN Data Augmentation
		4.3 Comparison of AlexNetv2 with the State-of-the-Art Models
	5 Conclusions
	References
Spoofing Detection for Speaker Verification with Glottal Flow and 1D Pure Convolutional Networks
	1 Introduction
	2 Related Work
	3 Implementation
		3.1 Segmenting Voiced Speech
		3.2 Discarding Frames with Unvoiced Speech
		3.3 Estimating Glottal Flow
		3.4 Classifying Frames as Spoof or Genuine
		3.5 Final Decision with a Voting Approach
	4 Experiments
	5 Conclusions and Future Work
	References
Estimation of Stokes Parameters Using Deep Neural Networks
	1 Introduction
	2 Dataset
	3 Methods
		3.1 Feedforward Neuronal Network
		3.2 Selection of Neural Network Parameters
		3.3 Selection of Neuronal Network Architecture
		3.4 General Model vs. Specialized Model
	4 Results
	5 Conclusions
	References
Experimental Study of the Performance of Convolutional Neural Networks Applied in Art Media Classification
	1 Introduction
	2 Related Work
	3 Materials and Methods
		3.1 Dataset
		3.2 CNN Architecture and Transfer Learning
		3.3 Improving Model Classification
	4 Experiments and Results
	5 Conclusions and Future Work
	References
Medical Applications of Pattern Recognition
Hadamard Layer to Improve Semantic Segmentation in Medical Images
	1 Introduction
	2 Related Work
		2.1 Semantic Segmentation
		2.2 Linear Error-Correction Codes
	3 Methodology
	4 Experiments and Results
	5 Conclusions and Future Work
	References
Patterns in Genesis of Breast Cancer Tumor
	1 Antecedents
	2 Material and Methods
		2.1 Gene Expressions and Mutual Information
		2.2 Genetic Networks from Matrix Correlation
	3 Results: GN of Breast Cancer Primary Tumor
		3.1 Single GN Topology
		3.2 Statistical Analysis on GN Topology
		3.3 Plausible Main Genes of Breast Cancer Primary Tumor
	4 Discussion
	5 Conclusions
	References
Realistic Simulation of Event-Related Potentials and Their Usual Noise and Interferences for Pattern Recognition
	1 Introduction
	2 Methods
		2.1 Scheme of Simulation
		2.2 Selection of the Parameters for Simulation
		2.3 Model Selection for Background Noise Simulation
	3 Results and Discussion
		3.1 Selected Model
		3.2 Simulated Noise
		3.3 Simulated EP Records
	4 Conclusions
	References
Chest X-Ray Imaging Severity Score of COVID-19 Pneumonia
	1 Introduction
	2 Proposed Method
		2.1 Semantic Genesis Classification and Severity Level Evaluation
	3 Experimental Results and Discussion
		3.1 Data Set
		3.2 Implementation and Training Details
		3.3 Classifier Training
		3.4 Validation of the Obtained Classifier
		3.5 Comparison to State-of-Art
	4 Conclusions
	References
Leukocyte Detection with Novel Fully Convolutional Network and a New Dataset of Blood Smear Complete Samples*-12pt
	1 Introduction
	2 Datasets
	3 CUNet
		3.1 Input and Preprocessing
		3.2 Encoder and Bottleneck
		3.3 Decoder
		3.4 Final Convolution Layers
	4 Results
		4.1 Training and Performance of CUNet
		4.2 Comparative Analysis of CUNet
	5 Conclusion
	References
Comparison of Deep Learning Architectures in Classification of Microcalcifications Clusters in Digital Mammograms
	1 Introduction
	2 Materials and Methods
		2.1 Deep Learning Architectures
		2.2 Experimental Framework
		2.3 Data Preparation
	3 Experiments and Results
	4 Discussion
	5 Conclusion
	References
Retinal Artery and Vein Segmentation Using an Image-to-Image Conditional Adversarial Network
	1 Introduction
	2 Methodology
		2.1 Neatwork Architecture
		2.2 Dataset
	3 Experiments and Results
		3.1 Understanding cGAN Predictions and Metrics
		3.2 Using Data Augmentation
		3.3 Colored Images
		3.4 Performance
	4 Conclusions and Future Work
	References
Evaluation of Heatmaps as an Explicative Method for Classifying Acute Lymphoblastic Leukemia Cells
	1 Introduction
		1.1 Explanatory Methods
	2 Related Work
	3 Methodology
		3.1 iNNvestigate Library
		3.2 Evaluation of Heat Maps by Expert Hematologists
	4 Experiments and Results
		4.1 Retrained Architectures
		4.2 Heat Map Generation
		4.3 Results of Heat Map Evaluation by Expert Hematologists
	5 Discussion
	6 Conclusions
	7 Future Work
	References
Language Processing and Recognition
Machine Learning Models Applied in Sign Language Recognition
	1 Introduction
	2 Related Work
	3 Proposed Method
	4 Design of Experiments
	5 Results and Discussion
	6 Conclusions
	References
Urdu Semantic Parsing: An Improved SEMPRE Framework for Conversion of Urdu Language Web Queries to Logical Forms
	1 Introduction
	2 Related Work
	3 Semantic Parser for Urdu Web Queries
	4 Results and Discussion
	5 Conclusion and Future Work
	References
Improving the Identification of Abusive Language Through Careful Design of Pre-training Tasks
	1 Introduction
	2 Related Work
	3 Adaptation of Pretraining Tasks
		3.1 Data Collection Approach
		3.2 Adapted Pre-training Tasks
		3.3 Generation of Robust Language Models for the Detection of AL
	4 Experimental Settings
		4.1 Datasets for the Detection of Abusive Language
		4.2 Implementation Details
	5 Experimental Results
		5.1 Effectiveness of the Specially Suited Pre-training Tasks
		5.2 Qualitative Analysis
	6 Conclusions and Future Work
	References
Industrial Applications of Pattern Recognition
TOPSIS Method for Multiple-Criteria Decision-Making Applied to Trajectory Selection for Autonomous Driving
	1 Introduction
	2 Proposed Method
		2.1 Experimentation
		2.2 Results and Discussion
	3 Conclusions and Future Work
	References
Machine-Learning Based Estimation of the Bending Magnitude Sensed by a Fiber Optic Device
	1 Introduction
	2 Methods
		2.1 Setup
		2.2 Bending Classification
	3 Results and Discussion
	4 Conclusions
	References
Graph-Based Semi-supervised Learning Using Riemannian Geometry Distance for Motor Imagery Classification
	1 Introduction
	2 Methods
		2.1 Semi-supervised Methods
		2.2 Distances
		2.3 Performance Evaluation
		2.4 Database
	3 Results
	4 Discussion and Conclusion
	References
Correction to: Machine Learning Models Applied in Sign Language Recognition
	Correction to: Chapter “Machine Learning Models Applied in Sign Language Recognition” in: A. Y. Rodríguez-González et al. (Eds.): Pattern Recognition, LNCS 13902, https://doi.org/10.1007/978-3-031-33783-3_25
Author Index