Moya-Albor, Ernesto
Main Affiliation
Preferred name
Moya-Albor, Ernesto
Official Name
Moya Albor, Ernesto
ORCID
0000-0002-9637-786X
Researcher ID
J-1891-2019
Scopus Author ID
55523223700
87 results
Now showing 1 - 10 of 87
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Metaheuristic Strategy in Automatic Robotics Navigation for Patient Care in Hospitals(Springer Nature Switzerland, 2025); ; This chapter presents a novel approach for developing a global automatic navigation strategy for a mobile robot for patient care into health centers and hospitals. The proposal uses a metaheuristic strategy through the Grey Wolf Optimizer algorithm to find the shortest route between the robot’s starting and patient destination points, while also integrating knowledge of the environment acquired through vision. The global navigation tracking is achieved through the use of Fuzzy Logic Rules to control the robot’s wheel velocity according to its current position and orientation. The proposed strategy was implemented in a virtual robotics environment demonstrating that the approach successfully generated optimal paths for various environments with a minimal number of control points and in a relatively short amount of time. In addition, a simulated hospital room with common furniture and tasks was used to evaluate the performance of the global navigation strategy, demonstrating that the robot could successfully generate a global route and navigate within this environment. ©The authors ©Springer. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Identifying and Mitigating Label Noise in Deep Learning for Image ClassificationLabeling errors in datasets are a persistent challenge in machine learning because they introduce noise and bias and reduce the model’s generalization. This study proposes a novel methodology for detecting and correcting mislabeled samples in image datasets by using the Cumulative Spectral Gradient (CSG) metric to assess the intrinsic complexity of the data. This methodology is applied to the noisy CIFAR-10/100 and CIFAR-10n/100n datasets, where mislabeled samples in CIFAR-10n/100n are identified and relabeled using CIFAR-10/100 as a reference. The DenseNet and Xception models pre-trained on ImageNet are fine-tuned to evaluate the impact of label correction on the model performance. Evaluation metrics based on the confusion matrix are used to compare the model performance on the original and noisy datasets and on the label-corrected datasets. The results show that correcting the mislabeled samples significantly improves the accuracy and robustness of the model, highlighting the importance of dataset quality in machine learning. ©The authors ©MDPI. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Design of an Underwater Mechatronics System for Detecting Dissolved Solids in Water(IEEE, 2024) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Critical Regimes of Systemic Risk: Flow Network Cascades in the U.S. Banking System(MDPI AG, 2026); Fernández Anaya, GuillermoSystemic risk in banking systems arises from losses transmitted through networks of contractual exposures. Yet, most widely used measures rely on market-implied volatility and equity prices rather than structural balance sheet fragilities. This paper develops a flow network framework that models systemic risk as a capacity-constrained loss-diffusion process governed by flow conservation, contractual seniority, and interbank topology. Using regulatory balance sheet data for four major U.S. banks across six quarters of the 2007–2008 financial crisis, we simulate millions of unit-consistent cascade scenarios to characterize the distribution of bank failures and aggregate losses. Despite severe macro-financial stress, the system remains in a subcritical contagion regime, exhibiting frequent single-bank failures, virtually no multi-bank cascades, and quasi-stationary aggregate losses concentrated around USD 420–430B.We extend the model to a stochastic setting in which the initial shock magnitude is randomized while propagation mechanics remain deterministic. The resulting loss distribution remains tightly concentrated and scales approximately linearly with shock size, suggesting that uncertainty in shock realizations does not induce nonlinear cascade amplification. Applying an efficient network benchmark, we estimate that 10–23% of expected systemic loss is attributable to suboptimal network architecture, implying potential gains from structural policy intervention. A comparison with SRISK reveals early divergence and convergence only at peak stress, highlighting the complementary roles of structural and market-based systemic risk measures. Finally, a graph neural network trained on synthetic flow network data fails to reproduce threshold-driven cascade dynamics, underscoring the importance of considering network structures vis-à-vis data-driven approaches. © The authors © MDPI. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Preface : Machine Learning Methods in Biomedical Field Computer-Aided Diagnostics, Healthcare and Biology Applications(Springer Science and Business Media Deutschland GmbH, 2026); ; ; - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Orthosis Control Based on Electromyographic Signals and Machine Learning(Springer Nature Switzerland, 2025) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Non-Contact Oxygen Saturation Estimation Using Deep Learning Ensemble Models and Bayesian OptimizationMonitoring Peripheral Oxygen Saturation (SpO2) is an important vital sign both in Intensive Care Units (ICUs), during surgery and convalescence, and as part of remote medical consultations after of the COVID-19 pandemic. This has made the development of new SpO2-measurement tools an area of active research and opportunity. In this paper, we present a new Deep Learning (DL) combined strategy to estimate SpO2 without contact, using pre-magnified facial videos to reveal subtle color changes related to blood flow and with no calibration per subject required. We applied the Eulerian Video Magnification technique using the Hermite Transform (EVM-HT) as a feature detector to feed a Three-Dimensional Convolutional Neural Network (3D-CNN). Additionally, parameters and hyperparameter Bayesian optimization and an ensemble technique over the dataset magnified were applied. We tested the method on 18 healthy subjects, where facial videos of the subjects, including the automatic detection of the reference from a contact pulse oximeter device, were acquired. As performance metrics for the SpO2-estimation proposal, we calculated the Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and other parameters from the Bland–Altman (BA) analysis with respect to the reference. Therefore, a significant improvement was observed by adding the ensemble technique with respect to the only optimization, obtaining 14.32% in RMSE (reduction from 0.6204 to 0.5315) and 13.23% in MAE (reduction from 0.4323 to 0.3751). On the other hand, regarding Bland–Altman analysis, the upper and lower limits of agreement for the Mean of Differences (MOD) between the estimation and the ground truth were 1.04 and −1.05, with an MOD (bias) of −0.00175; therefore, MOD ±1.96𝜎 = −0.00175 ± 1.04. Thus, by leveraging Bayesian optimization for hyperparameter tuning and integrating a Bagging Ensemble, we achieved a significant reduction in the training error (bias), achieving a better generalization over the test set, and reducing the variance in comparison with the baseline model for SpO2 estimation. ©The authors ©Technologies ©MDPI. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Automatic Robotics Medication Delivery System: The ANDIS Case Study(Springer Nature Switzerland, 2025) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Machine Learning Methods in Biomedical Field : Computer-Aided Diagnostics, Healthcare and Biology Applications(Springer Nature Switzerland, 2026); ; ; - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Towards the Distributed Wound Treatment Optimization Method for Training CNN Models: Analysis on the MNIST DatasetConvolutional neural network (CNN) is a prominent algorithm in Deep Learning methods. CNN architectures have been used successfully to solve various problems in image processing, for example, segmentation, classification, and enhancement task. However, automatic search for suitable architectures and training parameters remain an open area of research, where metaheuristic algorithms have been used to fine-tuning the hyperparameters and learning parameters. This work presents a bio-inspired distributed strategy based on Wound Treatment optimization (WTO) for training the learning parameters of a LenNet CNN model fast and accurate. The proposed method was evaluated over the popular benchmark dataset MNIST for handwritten digit recognition. Experimental results showed an improvement of 36.87% in training time using the distributed WTO method compared to the baseline with a single learning agent, and the accuracy increases 4.69% more using the proposed method in contrast with the baseline. As this is a preliminary study towards the distributed WTO method for training CNN models, we anticipate this approach can be used in robotics, multi-agent systems, federated learning, complex optimization problems, and many others, where an optimization task is required to be solved fast and accurate. © 2023 IEEE.Scopus© Citations 2 9 1