My research focuses on the development of trustworthy and clinically meaningful AI systems for biomedical imaging, computational pathology and multimodal medicine. The central goal is to connect tissue morphology with molecular, spatial and clinical information in order to better understand disease mechanisms and support precision medicine.
AI → Medicine
Interdisciplinary Research
22
PhD Students Supervised
MICCAI
Board Member & General Chair
20+
Years of AI Research
Recent Highlights
Spatial Transcriptomics
New work linking tissue morphology and gene expression at MICCAI 2026 (COMPAYL workshop), MIDL 2026 and ECDP 2026.
Virtual Staining
Generative model producing eight IHC markers from a single H&E slide, published in PLOS Computational Biology (2025), with two international patent applications.
AI in Neuro-Ophthalmology
Best Scientific Platform Presentation Prize at EUNOS 2026, following work on early retinal artery occlusion detection in JAHA (2025).
Invited Talks
SIPAIM 2026 keynote in Lima, AAAI-26 SPARTA workshop, CVPR 2025 Computer Vision for Drug Discovery workshop and JOBIM 2025.
Research Vision
From Microscopy to Molecular Insight
Histopathology is being transformed by digital pathology, whole-slide imaging, spatial omics and modern artificial intelligence. My research develops computational methods that make tissue data more measurable, interpretable and clinically actionable.
We study how to analyze and model complex biomedical images across scales, from microscopy to tissue architecture, and how to bridge morphology with molecular information such as spatial transcriptomics and spatial proteomics. A central goal is to build robust, explainable and human-centered AI systems that can assist clinicians and biomedical researchers while remaining scientifically grounded and ethically responsible.
Core Research Areas
Computational Pathology
AI methods for detection, segmentation, quantification and interpretation in whole-slide images.
Omics Imaging
Integration of histology with spatial transcriptomics, spatial proteomics and multimodal tissue data.
Virtual Staining
Computational generation of molecular and immunohistochemical contrasts from standard H&E images.
Trustworthy AI
Human-centered models with interpretability, uncertainty estimation and responsible decision support.
