About Simon

I am Postdoctoral Researcher at the Karlsruhe Institute of Technology (KIT), where I work on computer vision and machine learning systems with a focus on data-efficient learning algorithms to make vision technologies more accessible, affordable and human-centric.

I am broadly interested in research collaborations on innovative learning paradigms like semi-weakly supervised and self-supervised learning, visual in‑context learning and data-centric AI. If you’re interested, feel free to reach out! At KIT I am also frequently supervising thesis projects and am always looking for bright, motivated students to work with.

Selected Publications I co-authored publications at top venues such as CVPR, ECCV, ICCV, NeurIPS, AAAI and MICCAI, where I contributed to research fields such as semi-weakly supervised learning, visual in-context learning and to applications in domains where data-scarcity is prevalent, such as semantic segmentation in medical imaging.

List of some works:

	Is Visual in-Context Learning for Compositional Medical Tasks within Reach? International Conference on Computer Vision (ICCV) 2025 Simon Reiß, Zdravko Marinov, Alexander Jaus, Constantin Seibold, Saquib Sarfraz, Erik Rodner, Rainer Stiefelhagen (Paper)
	Decoupled Semantic Prototypes Enable Learning From Diverse Annotation Types for Semi-Weakly Segmentation in Expert-Driven Domains Conference on Computer Vision and Pattern Recognition (CVPR) 2023 Simon Reiß, Constantin Seibold, Alexander Freytag, Erik Rodner, Rainer Stiefelhagen (Paper)
	Graph-constrained Contrastive Regularization for Semi-weakly Volumetric Segmentation* European Conference on Computer Vision (ECCV) 2022 Simon Reiß, Constantin Seibold, Alexander Freytag, Erik Rodner, Rainer Stiefelhagen (Paper)
	Every Annotation Counts: Multi-Label Deep Supervision for Medical Image Segmentation Conference on Computer Vision and Pattern Recognition (CVPR) 2021 Simon Reiß, Constantin Seibold, Alexander Freytag, Erik Rodner, Rainer Stiefelhagen (Paper)

Research Interests Machine learning in low-data scenarios, learning paradigms from semi- and self- to semi-weakly supervised learning, visual in-context learning and data-centric AI.