At the 22nd European Congress on Digital Pathology (ECDP 2026) in Graz, Carmen Colin delivered an oral presentation in the Bias in AI session of the scientific programme on 18 June 2026. Her talk, “Training of State‑of‑the‑Art Deep Learning Models With and Without Ambiguous Regions for Nuclei Instance Segmentation”, contributed to discussions on data quality, annotation uncertainty, and bias in computational pathology, highlighting methodological considerations for developing robust and transparent AI models.
In the days leading up to the conference, Carmen also attended the 1st ESDIP–MICCAI–ASDP Summer School in Computational Pathology, held on 15–16 June 2026 as part of the ECDP framework. The two‑day programme, “From Zero to Hero”, introduced foundational concepts in pathology and machine learning and included hands‑on sessions in digital image analysis and selected advanced topics. The summer school provided structured training and valuable opportunities to exchange with researchers and clinicians working in computational pathology.
.

