activity
20242026
collaborators

6 papers

cs.CV2026

Unsupervised Semantic Segmentation Facilitates Model Understanding

Xiaoyan Yu, Lisa Mais, Jannik Franzen +4

Self-supervised learning (SSL) has produced a diverse landscape of vision transformers (ViTs) whose pretrained representations support a wide range of downstream tasks. Towards a b…

cs.CV2026

Better than Average: Spatially-Aware Aggregation of Segmentation Uncertainty Improves Downstream Performance

Vanessa Emanuela Guarino, Claudia Winklmayr, Jannik Franzen +7

Uncertainty Quantification (UQ) is crucial for ensuring the reliability of automated image segmentations in safety-critical domains like biomedical image analysis or autonomous dri…

cs.CV2025

SelfAdapt: Unsupervised Domain Adaptation of Cell Segmentation Models

Fabian H. Reith, Jannik Franzen, Dinesh R. Palli +2

Deep neural networks have become the go-to method for biomedical instance segmentation. Generalist models like Cellpose demonstrate state-of-the-art performance across diverse cell…

cs.CV2025

PhenoBench: A Comprehensive Benchmark for Cell Phenotyping

Claudia Winklmayr, Jerome Luescher, Nora Koreuber +6

Digital pathology has seen the advent of a wealth of foundational models (FM), yet to date their performance on cell phenotyping has not been benchmarked in a unified manner. We th…

cs.CV2024

Arctique: An artificial histopathological dataset unifying realism and controllability for uncertainty quantification

Jannik Franzen, Claudia Winklmayr, Vanessa E. Guarino +6

Uncertainty Quantification (UQ) is crucial for reliable image segmentation. Yet, while the field sees continual development of novel methods, a lack of agreed-upon benchmarks limit…

cs.CV2024

Model Guidance via Explanations Turns Image Classifiers into Segmentation Models

Xiaoyan Yu, Jannik Franzen, Wojciech Samek +2

Heatmaps generated on inputs of image classification networks via explainable AI methods like Grad-CAM and LRP have been observed to resemble segmentations of input images in many…