Manfredo Atzori

Overview

Manfredo Atzori serves as a Scientific Assistant (Adjoint-e scientifique HES A) at HES-SO Valais-Wallis's Haute Ecole de Gestion. His work focuses on advancing machine learning applications in biomedical engineering, digital pathology, and neuroimaging. He leads development of open-source tools like BIDSAlign for EEG standardization and SelfEEG for self-supervised learning in biomedical signals.

Research Interests

Atzori's research bridges AI and healthcare, emphasizing:

• Medical Imaging: Histopathology WSI analysis, artifact correction, and cross-modal fusion
• Prosthetics Control: sEMG-based gesture recognition and adaptive neural interfaces
• Deep Learning Methodology: Transfer learning optimization, self-supervised pretraining, and reproducibility frameworks

Key Contributions

2024 highlights include:

• Developed RegWSI - winning ACROBAT 2023 challenge for WSI registration
• Pioneered multimodal WSI-report integration to halve annotation needs
• Published benchmark studies on Gleason grading deep learning methods

Collaborations

Active in international projects with institutions like DeeperHistReg group and OpenNeuro. Tools developed are open-source (GitHub: @manfredoatzori) to advance reproducible research practices.