AI for Health
Medical imaging, neuroimaging, trustworthy clinical AI and data harmonisation for digital healthcare.
Healthcare is a demanding test bed for trustworthy AI: models must be accurate, robust to data shifts, explainable to clinicians and compliant with regulation. This line develops deep learning methods for medical imaging and clinical data, together with information fusion and design frameworks that make medical AI reliable and auditable. Much of it is done with long-standing partners such as Imperial College London and BOKU University Vienna.
My contributions work at three levels:
- Models: deep learning for medical image segmentation, reconstruction and classification (airways, cardiac and vascular imaging, fast MRI, brain tumours, capsule endoscopy, COVID-19 lesions and chest X-rays), and for physiological signals such as ECG-based pulse and arrhythmia detection during cardiac arrest and EEG and M/EEG connectomics.
- Populations: neuroimaging-based characterization of Alzheimer’s disease and brain age.
- Systems: information fusion, data harmonisation, synthetic data, federated learning and design frameworks for trustworthy clinical AI.
Earlier work applied bio-inspired optimization to medical image border detection, the planning of emergency medical units, the distribution of medical goods and molecular docking. Mikel Mendibe Abarrategi’s ongoing thesis studies uncertainty-aware deep learning for modelling tumour growth.
Main contributions
- Medical imaging: airway segmentation with fuzzy attention networks (Nan et al., 2024), human-AI collaborative active learning for pulmonary airway segmentation (Wang et al., 2024), fast MRI reconstruction (Huang et al., 2022; Huang et al., 2022), 3D and cardiac segmentation (Li et al., 2024; Yu et al., 2023), vascular and cardiac analysis (Liu et al., 2024; Zhang et al., 2021; Wu et al., 2022; Xing et al., 2023), brain tumor classification (Muhammad et al., 2021), hydrocephalus assessment (Zhou et al., 2023), COVID-19 lesions and chest X-rays (Li et al., 2023; Liz et al., 2023; Xiong et al., 2025) and wireless capsule endoscopy (Muhammad et al., 2020).
- Neuroimaging and neurodegeneration: ensemble and fuzzy deep learning for Alzheimer’s disease (Tanveer et al., 2024; Tanveer et al., 2024), multimodal prediction of Alzheimer’s progression (Rahim et al., 2023), brain age estimation (Tanveer et al., 2023), brain connectomics (Del Ser et al., 2018) and cognitive workload from EEG (Lobo et al., 2016).
- Physiological signals: ECG-based pulse detection and rhythm classification during cardiac arrest (Elola et al., 2019; Manibardo et al., 2019), arrhythmia detection (Kumar et al., 2023) and online heart monitoring on the Internet of Health Things (Santos et al., 2020).
- Trustworthy medical AI: information fusion as an enabler of robust and explainable medical AI (Holzinger et al., 2022), a design framework for operationalizing trustworthy AI in healthcare (Moreno-Sánchez et al., 2026) and non-imaging medical data synthesis (Xing et al., 2024).
- Data and collaboration: data harmonisation for digital healthcare (Nan et al., 2022) and federated deep learning across clinical centres (Wahab et al., 2024).
- Bio-inspired optimization in medicine: border detection of skin lesions (Galvez et al., 2019; Galvez et al., 2019), emergency medical units (Landa-Torres et al., 2013), distribution of medical goods (Osaba et al., 2019) and molecular docking (Garcia-Godoy et al., 2019).
Publications in this area (40)
All my publications related to this research line, in reverse chronological order, including those mentioned above.
2026
2025
2024
2023
2022
2021
2020
2019
2018
2016
- Congreso Anual de la Sociedad Española de Ingeniería Biomédica (CASEIB)Detección del Retorno de la Circulación Espontánea en Base al ElectrocardiogramaIn Congreso Anual de la Sociedad Española de Ingeniería Biomédica (CASEIB), 2016