Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 3
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Artificial Intelligence Pipeline for Mammography-Based Breast Cancer Detection: An Integrated Systematic Review and Large-Scale Experimental Validation

  • Añez, Daniel
  • Conti, Giuseppe
  • Uriarte, Juan José
  • Serrano-Olmedo, J. J.
  • Martínez-Murillo, Ricardo
  • Casanova-Carvajal, O.
This article belongs to the Special Issue Application of Artificial Intelligence in Disease Diagnosis and Treatment, [Background and Objectives] Breast cancer remains a leading cause of cancer-related morbidity and mortality worldwide, and robust, interpretable artificial intelligence (AI) pipelines are increasingly being explored to support mammography-based detection. This study combines a PRISMA 2020-compliant systematic review with an original experimental validation to characterize current evidence and address identified gaps in reproducibility and interpretability., [Materials and Methods] A PRISMA 2020-guided systematic review and an original experimental study were conducted. The review searched PubMed and Scopus/ScienceDirect for studies using convolutional neural networks (CNNs), support vector machines (SVMs) or eXtreme Gradient Boosting (XGBoost) for breast cancer detection in mammography and related imaging modalities, and identified 45 eligible articles. In parallel, we implemented and evaluated representative CNN (ResNet-50, EfficientNetB0 and MobileNetV3-Small) and classical machine learning (SVM and XGBoost) pipelines on the CBIS-DDSM dataset, following a CRISP-DM-inspired workflow and using Grad-CAM and SHAP to provide image- and feature-level explanations within a reproducible machine-learning-operations (MLOps)-oriented framework., [Results] The systematic review revealed substantial heterogeneity in datasets, preprocessing pipelines, and validation strategies, with a predominant reliance on internal validation and limited use of explainable AI methods. In our experimental evaluation, ResNet-50 achieved the best performance (AUC-ROC 0.95; sensitivity 89%), followed by XGBoost (AUC-ROC 0.90; sensitivity 74%) and SVM (AUC-ROC 0.84; sensitivity 66%), while EfficientNetB0 and MobileNetV3-Small showed lower discrimination. Grad-CAM produced qualitatively plausible heatmaps centered on annotated lesions, and SHAP analyses indicated that simple global image-intensity and size descriptors dominated the predictions of the classical models., [Conclusions] By integrating systematic evidence and large-scale experiments on CBIS-DDSM, this study highlights both the potential and the limitations of current AI pipelines for mammography-based breast cancer detection, underscoring the need for more standardized preprocessing, rigorous external validation, and routine use of explainable AI before clinical deployment., This research was partially funded by the Ministerio de Ciencia, Innovación y Universidades of Spain, ref.: PGC2018-097531-B-I00, PDC2022-133028-I00 and PDC2023-145812-I00, funded by the European Union-NextGenerationEU., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Modeling the Dynamics of Electric Field-Assisted Local Functionalization in Two-Dimensional Materials

  • Borrás, Fernando
  • Ramiro-Bargueño, J.
  • Casanova-Carvajal, O.
  • Andrés, Alicia de
  • Quesada, Sergio J.
  • Álvarez, Ángel Luis
This article belongs to the Section Materials Simulation and Design., Electric field-assisted local functionalization of materials is a resist-free technique generally applied at the nanoscale, which has been understood within the paradigm of the water meniscus. Using a home-made prototype the authors applied this technique at scales compatible with the biosensor industry (tens of microns). However, interpreting these results requires a different paradigm. The expansion of the oxidized region over time in two-dimensional materials under a localized electric field is modeled from first physical principles. Boltzmann statistics is applied to the oxyanion incorporation at the perimeter of the oxidized zone, and a new general relation between oxide radius and time is formulated. It includes the reduction in the energy barrier due to the field effect and its dependence on the oxide radius. To gain insight into this dependence whatever the layers structure, 2D material involved, or electrical operating conditions, simple structures based on multilayer stacks representing the main constituents are proposed, where the Poisson equation is solved using finite element calculations. This enables to derive energy barriers for oxyanion incorporation at varying spot radii which are consistent with those resulting from fitting experimental data. The reasonable agreement obtained provides researchers with a new tool to predict the evolution of local functionalization of 2D layers as a function of the following fabrication parameters: time, applied voltage, and relative humidity, solely based on materials properties., This work is supported by grants PDC2023–145812-I00 (Project SAMPL2D), PID2020–115514RB-I00 (Project HyMSODAE), and PID2023–148746OB-I00 (Project SCIENTIST), which are funded by MICIU/AEI/10.13039/501100011033 and by “Next Generation EU/PRTR”., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Upper-Limb Robotic Rehabilitation: Online Sliding Mode Controller Gain Tuning Using Particle Swarm Optimization

  • Sosa Méndez, Deira
  • Bedolla Martínez, David
  • Saad, Maarouf
  • Kali, Yassine
  • García Cena, Cecilia E
  • Álvarez, Ángel L
Two primary challenges in controlling robotic rehabilitation devices are the uncertainties in dynamic models and, more importantly, the need for controllers capable of adapting to external disturbances due to human–robot interaction. To address these issues, this paper proposes the particle swarm optimization (PSO) algorithm for the real-time gain tuning in the sliding mode controller (SMC) based on the exponential reaching law (ERL). The proposed approach was designed for a seven-degrees-of-freedom (DOF) robotic exoskeleton used in upper-limb physical rehabilitation. The optimization algorithm aims to minimize tracking errors in rehabilitation exercises through the robust ERL controller applied to nonlinear systems with external disturbances. The proposed method was validated through experimental tests conducted on two healthy subjects, and the outcomes indicated a reduction of over 20% in tracking errors compared to heuristically tuned gains. Mathematical analyses of dynamic modeling and algorithm convergence are shown., This research is partially supported by grant PDC2023-145812-I00 (Project SAMPL2D), which is funded by MICIU/AEI/10.13039/501100011033 and by “Next Generation EU/PRTR, Peer reviewed