Pattern Recognition and Image Analysis 9th Iberian Conference, IbPRIA 2019, Madrid, Spain, July 1-4, 2019, Proceedings, Part I

• Main Authors: Morales, Aythami, Fierrez, Julian, Sánchez, José Salvador, Ribeiro, Bernardete
• Format: eBook
• Language: English
• Published: Cham Springer International Publishing AG 2019
• Edition: 1
• Subjects: Optical pattern recognition-Congresses

Table of Contents:

• Intro -- Preface -- Organization -- Abstracts of Invited Tutorials -- Machine Learning with Scikit-Learn -- Computer Vision for Affective Computing -- Bayesian Optimization -- Abstracts of Invited Talks -- Building Computer Vision Systems that Really Work -- Face Analysis for Multimodal Emotional Interfaces -- Fun with Human-Machine Collaboration for Computer Vision -- Towards Human Behavior Modeling from (Big) Mobile Data -- Contents -- Part I -- Contents -- Part II -- Best Ranked Papers -- Towards a Joint Approach to Produce Decisions and Explanations Using CNNs -- 1 Introduction -- 2 Methodology -- 2.1 Proposed Architecture -- 2.2 Synthetic Dataset -- 2.3 Experiments -- 2.4 Experiments on Real Datasets -- 3 Results and Discussion -- 4 Conclusion -- References -- Interactive-Predictive Neural Multimodal Systems -- 1 Introduction -- 2 Interactive-Predictive Multimodal Pattern Recognition -- 2.1 Neural Architectures for Multimodal Sequence to Sequence Learning -- 2.2 Interactive-Predictive Pattern Recognition -- 2.3 Evaluation Metrics -- 2.4 Usage of the System and User Simulation -- 2.5 Description of the Systems -- 3 Results and Discussion -- 3.1 Quantitative Evaluation -- 3.2 Qualitative Analysis and Discussion -- 4 Related Work -- 5 Conclusions and Future Work -- References -- Uncertainty Estimation for Black-Box Classification Models: A Use Case for Sentiment Analysis -- 1 Introduction -- 2 Related Work -- 2.1 Uncertainty in Deep Learning -- 2.2 Rejection Methods -- 3 Uncertainty Measures from Black-Box Models -- 3.1 A Wrapper for Computing Aleatoric Heteroscedastic Uncertainty -- 3.2 Uncertainty Heuristics -- 4 Use Case and Results -- 5 Conclusions -- References -- Impact of Ultrasound Image Reconstruction Method on Breast Lesion Classification with Deep Learning -- 1 Introduction -- 2 Materials and Methods -- 2.1 Dataset
• 5.2 Performance Evaluation -- 6 Conclusion -- References -- Pattern Recognition -- Description and Recognition of Activity Patterns Using Sparse Vector Fields -- 1 Introduction -- 2 Estimation of Multiple Vector Fields -- 3 Activity Pattern Labeling -- 4 Experimental Results -- 4.1 Synthetic Data -- 4.2 Real Data -- 5 Conclusion -- References -- Instance Selection for the Nearest Neighbor Classifier: Connecting the Performance to the Underlying Data Structure -- 1 Introduction -- 2 Categorization of Sample Types -- 3 Databases and Experimental Setting -- 4 Results and Discussion -- 5 Concluding Remarks -- References -- Modified DBSCAN Algorithm for Microscopic Image Analysis of Wood -- 1 Introduction -- 2 Materials and Methods -- 2.1 Wood Anatomy Database -- 2.2 Segmentation of Ray Cells -- 2.3 DBSCAN -- 2.4 Ray Width -- 2.5 Evaluation -- 3 Results -- 4 Conclusion -- References -- Automatic Detection of Tuberculosis Bacilli from Microscopic Sputum Smear Images Using Faster R-CNN, Transfer Learning and Augmentation -- Abstract -- 1 Introduction -- 2 Proposed Approach -- 3 Experimental Results -- 4 Conclusions -- Acknowledgement -- References -- Detection of Stone Circles in Periglacial Regions of Antarctica in UAV Datasets -- 1 Introduction -- 2 Data Acquisition -- 3 Detection Methods -- 3.1 Template Matching -- 3.2 Watershed -- 3.3 Sliding Band Filter -- 4 Results -- 4.1 Performance Measure -- 4.2 Statistical Evaluation -- 5 Conclusions -- References -- Lesion Detection in Breast Ultrasound Images Using a Machine Learning Approach and Genetic Optimization -- Abstract -- 1 Introduction -- 1.1 Modeling Lesions in Breast Ultrasound Images -- 1.2 Lesion Detection in Breast Ultrasound Images -- 1.3 Optimization of Machine Learning Methods with Genetic Algorithms -- 2 Materials and Methods -- 2.1 Dataset -- 2.2 Random Forest -- 2.3 Genetic Algorithms
• Model Based Recursive Partitioning for Customized Price Optimization Analytics -- 1 Introduction -- 2 The Decision Tree Modeling Approach -- 2.1 CART Algorithm -- 2.2 CTREE Algorithm -- 2.3 Model Based Recursive Partitioning -- 3 The MOB Method for Pricing Analytics -- 4 Business Case Application -- 4.1 Data Description -- 4.2 MOB Modeling -- 4.3 Logistic Regression Modeling -- 4.4 Optimization and Revenue Results -- 5 Summary and Concluding Remarks -- References -- 3D Reconstruction of Archaeological Pottery from Its Point Cloud -- Abstract -- 1 Introduction -- 2 Related Work -- 3 Data Acquisition -- 3.1 Calibration -- 3.2 Scanning the Pottery -- 4 The Registration Process -- 4.1 Data Pre-processing -- 4.2 Features Extraction -- 4.3 Correspondence Matching -- 4.4 Global Registration -- 5 Discussion -- 6 Conclusion -- Acknowledgement -- References -- Geometric Interpretation of CNNs' Last Layer -- 1 Introduction -- 2 Background -- 2.1 Geometric Interpretation -- 3 Experiments and Results -- 3.1 Balanced Dataset -- 3.2 Unbalanced Dataset -- 4 Conclusions and Future Work -- References -- Re-Weighted 1 Algorithms within the Lagrange Duality Framework -- 1 Introduction -- 2 Methodology with Oracle -- 3 Solutions of the Dual Problem -- 4 RW1 with Projected Subgradient Algorithm -- 5 Methodology and Algorithm Without Oracle -- 6 Problem with Noise -- 7 Experimental Results -- 7.1 Results for the Noise-Free Setting -- 7.2 Results for the Noisy Setting -- 8 Conclusions -- References -- A Note on Gradient-Based Intensity Normalization -- Abstract -- 1 Introduction -- 2 Review -- 3 Material -- 4 Methodology -- 5 Results -- 6 Conclusions -- Acknowledgements -- References -- Blind Robust 3-D Mesh Watermarking Based on Mesh Saliency and QIM Quantization for Copyright Protection -- 1 Introduction -- 2 Background -- 2.1 3-D Mesh Saliency
• 2.4 Proposed Method
• 2.2 Ultrasound Image Reconstruction -- 2.3 Transfer Learning with Convolutional Neural Networks -- 2.4 Experiments and Evaluation -- 3 Results -- 4 Discussion -- 5 Conclusions -- References -- Segmentation of Cell Nuclei in Fluorescence Microscopy Images Using Deep Learning -- 1 Introduction -- 2 Proposed Approach -- 3 Experiments -- 3.1 Data -- 3.2 Training -- 3.3 Evaluation Criteria -- 3.4 Performance Comparison -- 4 Results -- 4.1 Nuclei Segmentation -- 4.2 F1 Score vs IoU Threshold -- 4.3 Computational Efficiency -- 5 Conclusions and Future Work -- References -- Food Recognition by Integrating Local and Flat Classifiers -- 1 Introduction -- 2 Proposed Method -- 2.1 Model Setup -- 2.2 Epistemic Uncertainty -- 2.3 Prediction by Integrating Local and Flat Classifiers -- 3 Experiments -- 3.1 Dataset -- 3.2 Metric -- 3.3 Experimental Setup -- 3.4 Results -- 4 Conclusions -- References -- Machine Learning -- Combining Online Clustering and Rank Pooling Dynamics for Action Proposals -- 1 Introduction -- 2 Related Work -- 3 Action Proposals Generation -- 3.1 Feature Extraction -- 3.2 Online Clustering for Action Proposals -- 3.3 Filtering Proposals -- 4 Experiments -- 4.1 Experimental Setup -- 4.2 Comparison with the State of the Art -- 4.3 Ablation Study -- 5 Conclusion -- References -- On the Direction Guidance in Structure Tensor Total Variation Based Denoising -- 1 Introduction -- 2 Background -- 3 Method -- 4 Experimental Results -- 5 Conclusion and Future Work -- References -- Impact of Fused Visible-Infrared Video Streams on Visual Tracking -- 1 Introduction -- 2 Visual Trackers and Fusion -- 2.1 Reference Trackers -- 2.2 Fusion Strategies -- 3 Evaluation -- 3.1 Data Set to Subsets -- 3.2 Evaluation Metrics -- 3.3 Evaluation on Subsets -- 3.4 Fused Subsets Versus Domain Subsets -- 3.5 Special Case Analysis -- 4 Conclusion -- References
• 2.2 Quantization Index Modulation -- 3 The Proposed Method -- 3.1 Watermark Embedding -- 3.2 Watermark Extraction -- 4 Experimental Results -- 4.1 Experimental Setup -- 4.2 Results Discussion -- 4.3 Comparison with Alternative Methods -- 5 Conclusion -- References -- Using Copies to Remove Sensitive Data: A Case Study on Fair Superhero Alignment Prediction -- 1 Introduction -- 2 Case Study -- 3 Methodological Proposal -- 3.1 Copying Machine Learning Classifiers -- 3.2 Using Copies to Remove Sensitive Data -- 4 Experiments -- 5 Discussion of Results -- 6 Conclusions and Future Work -- References -- Weighted Multisource Tradaboost -- 1 Introduction -- 2 Transfer Learning -- 3 State-of-the-Art -- 3.1 Multi-KT - Support Vector Machines svm -- 3.2 Transfer Learning Decision Forests (TLDF) tldf -- 3.3 TrAdaboost -- 4 Experimental Design -- 4.1 Method Hyperparameters -- 5 Results -- 6 Conclusion -- 6.1 Future Work -- References -- A Proposal of Neural Networks with Intermediate Outputs -- 1 Introduction -- 2 Standard Neural Network -- 3 Proposed Method -- 3.1 Backpropagation Based on the Proposed Variant of Cost Function -- 4 Experiments -- 5 Conclusions -- References -- Addressing the Big Data Multi-class Imbalance Problem with Oversampling and Deep Learning Neural Networks -- 1 Introduction -- 2 Theoretical Framework -- 2.1 Deep Learning Multilayer Perceptron -- 2.2 Classical Sampling Methods Used to Deal with the Class Imbalance Problem -- 3 Experimental Set-Up -- 4 Results and Discussion -- 5 Conclusion -- References -- Reinforcement Learning and Neuroevolution in Flappy Bird Game -- 1 Introduction -- 2 Related Work -- 3 Flappy Bird Game -- 4 Reinforcement Learning to Flappy Bird Game -- 4.1 Q-Learning to Flappy Bird Game -- 4.2 Performance Evaluation -- 5 Neuroevolution Applied to Flappy Bird Game -- 5.1 Genetic Optimization