Pattern Recognition Applications and Methods : 10th International Conference, ICPRAM 2021, and 11th International Conference, ICPRAM 2022, Virtual Event, February 4–6, 2021 and February 3–5, 2022, Revised Selected Papers

This book contains revised and extended versions of selected papers from the 10th and 11 th International Conference on Pattern Recognition, ICPRAM 2021 and 2022, held in February 2021 and 2022. Due to COVID-19 pandemic the conferences were held virtually. Both conferences received in total 204 submissions from which 8 full papers were carefully reviewed and selected for presentation in this volume. The papers span a wide range of investigation as well as development lines, which of course always reflect the last trends of research in the pattern recognition community. Preface Organization Contents Theory and Methods Reduced Precision Research of a GAN Image Generation Use-case 1 Introduction 1.1 Reduced Precision Computing 1.2 High Energy Physics Simulations 1.3 Deep Convolutional Generative Adversarial Networks 2 Quantization Tool 3 Evaluation Quantization 3.1 Computational Evaluation 3.2 Physics Accuracy Evaluation 4 Conclusion References Similarity Constrained Conditional Generative Auto-encoder with Generalized Dilated Networks*-6pt 1 Introduction 2 Related Work 2.1 Generative Networks 2.2 Dilated Convolution Kernels 3 Fundamentals 3.1 SimAEGAN 3.2 Dilation 4 Experiments 4.1 Custom dSprites 4.2 MNIST 4.3 CelebA 5 Conclusion References Retinotopic Image Encoding by Samples of Counts 1 Introduction 2 Image Representation by Sample of Counts (Sampling Representation) 3 Retinotopic Encoding of the Sampling Representation as the Image Compression Method (Perceptual Coding) 4 Conclusions References Gesture Recognition and Multi-modal Fusion on a New Hand Gesture Dataset 1 Introduction 2 Related Work 2.1 Hand Gesture Datasets 2.2 Multi-modal Fusion 2.3 Contribution 3 Dataset 3.1 Gesture Classes 3.2 Modalities 3.3 Preprocessing 3.4 Acceleration 4 Unimodal Classification 4.1 LSTM 4.2 CNN 5 Multi-modal Fusion 6 Live Demonstrator 6.1 Implementation 6.2 Experiments 7 Conclusion References Adaptive Sampling for Weighted Log-Rank Survival Trees Boosting 1 Introduction 2 Related Work 2.1 Survival Tree 2.2 Random Survival Forest 2.3 Gradient Boosting Survival Analysis 3 Proposed Approach 3.1 Survival Tree with Weighted Log-Rank Criteria 3.2 Survival Bagging with Weighted Log-Rank Criteria 3.3 Weighted Boosting 3.4 Proposed Adaptive Boosting Ensemble 4 Experiments 4.1 Metrics 4.2 Datasets 4.3 Experimental Setup 4.4 Results 5 Conclusions References Applications Forecasting Overtime Budgets for Naval Fleet Maintenance Facilities Using Time-Series Analysis During Transient System States 1 Introduction 1.1 Background 1.2 Motivation 1.3 Related Work 2 Methodology 2.1 Prior Approach 2.2 Revised Approach 3 Results 3.1 Model Fitting 3.2 Model Testing 3.3 Model Forecasting 4 Conclusions 4.1 Prior Approach 4.2 Revised Approach 4.3 Future Work References Exploiting Temporal Coherence to Improve Person Re-identification 1 Introduction 2 Related Work 3 Dataset 4 Proposal Description 4.1 Runners in the Gallery 4.2 Fastest Elapsed Time 4.3 Adaptive Time Estimation 4.4 Average Time 4.5 Individual Time 5 ReID Experimental Evaluation 6 Conclusions References Perusal of Camera Trap Sequences Across Locations 1 Introduction 2 Related Works 3 Proposed Pipeline to Process Camera Trap Sequences 4 Experimental Studies for Processing Camera Trap Sequences 4.1 Experimental Studies on Empty Frame Removal 4.2 Experimental Studies for Wildlife Detection 4.3 Experimental Studies on Species Recognition from Camera Trap Sequences 5 Conclusion References Author Index