Computer Engineering, with a focus on computer vision applied to digestive endoscopy
[Open soon]
Work description
Brief Presentation of the Work and Traning: Title: Trustworthy structured deep learning for gastric endoscopy. The GATE project (GenerAlizable and Trustable computer assisted upper gastrointestinal Endoscopy) is a collaboration between INESC TEC and IPO Porto aiming to develop trustworthy artificial intelligence technologies for gastric cancer prevention. Work includes collaboration with hospitals in Rome, Bucharest and Rotterdam. Current AI models typically analyze individual images independently, ignoring the anatomical organization of the stomach and providing limited information regarding the reliability of their predictions. In contrast, endoscopists naturally integrate observations acquired across multiple anatomical regions to form a patient-level assessment. This PhD project will investigate novel deep learning methodologies that jointly model anatomical structure and prediction reliability. The research will explore graph-based representations of endoscopic examinations, anatomically structured learning, uncertainty estimation, and confidence-aware aggregation strategies, enabling the development of trustworthy AI systems capable of producing robust patient-level EGGIM estimations across multicentric clinical datasets.
Academic Qualifications
Degree in Computer Science, Informatics Engineering, Data Science, Bioengineering or Applied Mathematics.
Minimum profile required
Degree in Computer Science, Informatics Engineering, Data Science, Bioengineering or Applied Mathematics.Enrollment in a PhD program. This condition will be verified by the date of the signature of the contract and not on the date of application.
Preference factors
Previous experience in biomedical image and signal processing Previous experience in computer vision
Application Period
Since 16 Jul 2026 to 29 Jul 2026
[Open soon]
Centre
Biomedical Engineering Research