A5082285724- Cryptography and Data Security
- Hormonal Regulation and Hypertension
- Medical Imaging Techniques and Applications
- Diabetes Treatment and Management
- Advanced MRI Techniques and Applications
- Privacy-Preserving Technologies in Data
- Diet and metabolism studies
- Cloud Data Security Solutions
- Big Data and Business Intelligence
- ECG Monitoring and Analysis
- Advanced Neural Network Applications
- AI in cancer detection
- Human Mobility and Location-Based Analysis
- Cell Image Analysis Techniques
- Cancer Genomics and Diagnostics
- Liver Disease Diagnosis and Treatment
- Privacy, Security, and Data Protection
- Artificial Intelligence in Healthcare
- Family Caregiving in Mental Illness
- Cardiovascular Function and Risk Factors
- Data Quality and Management
- Traffic Prediction and Management Techniques
- IoT and Edge/Fog Computing
- Brain Tumor Detection and Classification
- Data Mining Algorithms and Applications
University of Minho
2017-2025
INESC TEC
2019-2025
Finerenone reduced the risk of kidney and cardiovascular events in people with chronic disease (CKD) type 2 diabetes FIDELIO-DKD FIGARO-DKD phase 3 studies. Effects finerenone on outcomes patients taking sodium-glucose cotransporter inhibitors (SGLT2is) were evaluated a prespecified pooled analysis these studies.Patients urine albumin-to-creatinine ratio (UACR) ≥30 to ≤5,000 mg/g estimated glomerular filtration rate (eGFR) ≥25 mL/min/1.73 m2 randomly assigned or placebo; SGLT2is permitted at...
Alzheimer’s disease (AD) places a profound global challenge, driven by its escalating prevalence and the multifaceted strain it on individuals, families, societies. Family caregivers (FCs), who are pivotal in supporting family members with AD, frequently endure substantial emotional, physical, psychological demands. To better understand determinants of caregiving strain, this study employed machine learning (ML) to develop predictive models identifying factors that contribute caregiver...
Breakthroughs in sequencing technologies led to an exponential growth of genomic data, providing novel biological insights and therapeutic applications. However, analyzing large amounts sensitive data raises key privacy concerns, specifically when the information is outsourced untrusted third-party infrastructures for storage processing (e.g., cloud computing). We introduce Gyosa, a secure privacy-preserving distributed analysis solution. By leveraging trusted execution environments (TEEs),...
When dealing with electrocardiography (ECG) the main focus relies on classification of heart's electric activity and deep learning has been proving its value over years classifying heartbeats, exhibiting great performance when doing so. Following these assumptions, we propose a model based ResNet architecture convolutional 1D layers to classify beats into one 4 classes: normal, atrial premature contraction, ventricular contraction others. Experimental results MIT-BIH Arrhythmia Database...
We propose Soteria, a system for distributed privacy-preserving Machine Learning (ML) that leverages Trusted Execution Environments (e.g. Intel SGX) to run code in isolated containers (enclaves). Unlike previous work, where all ML-related computation is performed at trusted enclaves, we introduce hybrid scheme, combining done inside and outside these enclaves. The conducted experimental evaluation validates our approach reduces the runtime of ML algorithms by up 41%, when compared related...
Medical imaging, mainly Magnetic Resonance Imaging (MRI), plays a predominant role in healthcare diagnosis. Nevertheless, the diagnostic process is prone to errors and conditioned by available medical data, which might be insufficient. A novel solution resorting image generation algorithms address these challenges. Thus, this paper presents Deep Learning model based on Convolutional Generative Adversarial Network (DCGAN) architecture. Our generates 2D MRI images of size <tex...
Health facilities produce an increasing and vast amount of data that must be efficiently analyzed. New approaches for healthcare monitoring are being developed every day the Internet Things (IoT) came to fill still existing void on real-time monitoring. A new generation mechanisms techniques used facilitate practice medicine, promoting faster diagnosis prevention diseases. We proposed a system relies IoT storing medical sensors with analytic capabilities. To this end, we chose two which were...
Cities worldwide have agreed on ambitious goals regarding carbon neutrality; thus, smart cities face challenges active and shared mobility due to public transportation's low attractiveness lack of real-time multimodal information. These issues led a data the community's choices, traffic commuters' footprint corresponding motivation change habits. Besides, many consumers are reluctant use some software tools privacy guarantee. This paper presents methodology developed in FranchetAI project...
Deep Learning (DL) training requires efficient access to large collections of data, leading DL frameworks implement individual I/O optimizations take full advantage storage performance. However, these are intrinsic each framework, limiting their applicability and portability across solutions, while making them inefficient for scenarios where multiple applications compete shared resources.We argue that should be decoupled from moved a dedicated layer. To achieve this, we propose new...
The adoption of third-party machine learning (ML) cloud services is highly dependent on the security guarantees and performance penalty they incur workloads for model training inference. This paper explores security/performance trade-offs distributed Apache Spark framework its ML library. Concretely, we build upon a key insight: in specific deployment settings, one can reveal carefully chosen non-sensitive operations (e.g. statistical calculations). allows us to considerably improve...
Breakthroughs in sequencing technologies led to an exponential growth of genomic data, providing unprecedented biological insights and new therapeutic applications. However, analyzing such large amounts sensitive data raises key concerns regarding privacy, specifically when the information is outsourced third-party infrastructures for storage processing (e.g., cloud computing). Current solutions privacy protection resort centralized designs or cryptographic primitives that impose...
Abstract The correction of artifacts in Magnetic Resonance Imaging (MRI) is increasingly relevant as voluntary and involuntary can hinder data acquisition. Reverting from corrupted to artifact-free images a complex task. Deep Learning (DL) models have been employed preserve characteristics identify correct those artifacts. We propose MOANA , novel DL-based solution multi-contrast brain MRI scans. offers two models: the simulation models. model introduces perturbations similar occurring an...
The correction of artifacts in Magnetic Resonance Imaging (MRI) is increasingly relevant as voluntary and involuntary can hinder data acquisition. Reverting from corrupted to artifact-free images a complex task. Deep Learning (DL) models have been employed preserve characteristics identify correct those artifacts. We propose MOANA, novel DL-based solution multi-contrast brain MRI scans. MOANA offers two models: the simulation models. model introduces perturbations similar occurring an exam...
The correction of artifacts in Magnetic Resonance Imaging (MRI) is increasingly relevant as voluntary and involuntary can hinder data acquisition. Reverting from corrupted to artifact-free images a complex task. Deep Learning (DL) models have been employed preserve characteristics identify correct those artifacts. We propose MOANA, novel DL-based solution multi-contrast brain MRI scans. MOANA offers two models: the simulation models. model introduces perturbations similar occurring an exam...
With the rapid growth of Deep Learning models and neural networks, medical data available for training - which is already significantly less than other types becoming scarce. For that purpose, Generative Adversarial Networks (GANs) have received increased attention due to their ability synthesize new realistic images. Our preliminary work shows promising results brain MRI images; however, there a need distribute workload, can be supported by High-Performance Computing (HPC) environments. In...
Continuous Ambulatory Peritoneal Dialysis (CAPD) is one of the many treatments for patients with advanced kidney disease. It a treatment that needs regular monitoring and understanding all factors blood urine samples each patient to understand if going well. This article will explore data information from undergoing CAPD procedure. helps comprehend how interoperability acts in Health Information System since this contains patients' personal but also samples' results, meaning services must be...