Rafael Sterzinger

PhD Student at TU Wien in Computer Vision

I am a PhD student in Computer Science at TU Wien, where I am part of the Computer Vision Lab, supervised by Robert Sablatnig. I am honored to be a recipient of the DOC Fellowship from the Austrian Academy of Sciences. My research interests revolve around developing reliable and efficient computer vision systems for low-resource domains, specifically leveraging vision foundation models for few-shot learning.

Previously, I earned my MSc in Computer Science from ETH Zurich, where I majored in Machine Intelligence and conducted my master's thesis in the Learning & Adaptive Systems Group under the guidance of Andreas Krause. I have also gained experience as a Research Engineer Intern at Hitachi Energy and as a Software Engineer Intern at iteratec.

Rafael Sterzinger

News

Undertook a two-week Marie Skłodowska-Curie Actions staff exchange to Bucharest as part of the iMARS project.

Presented our work DATTA: Domain-Adversarial Test-Time Adaptation for Cross-Domain WiFi-Based Human Activity Recognition at WACV 2026 in Tucson, Arizona, USA.

Received the DOC Fellowship from the Austrian Academy of Sciences (OeAW).

Successfully passed the PhD Proficiency Evaluation at TU Wien.

Attended Wintergraph 2026 in Kaprun, Austria.

Presented our work on Few-Shot Connectivity-Aware Text Line Segmentation in Historical Documents at ACPR 2025 in Gold Coast, Australia.

Presented our paper on Few-Shot Segmentation of Historical Maps via Linear Probing of Vision Foundation Models at ICDAR 2025 in Wuhan, China.

Started my PhD at TU Wien under the supervision of Professor Sablatnig.

Selected Publications

WACV POSTER

DATTA: Domain-Adversarial Test-Time Adaptation for Cross-Domain WiFi-Based Human Activity Recognition

J. Strohmayer , R. Sterzinger , M. Wödlinger , M. Kampel

IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
Code
ICDAR POSTER

Few-Shot Segmentation of Historical Maps via Linear Probing of Vision Foundation Models

R. Sterzinger , M. Peer , R. Sablatnig

International Conference on Document Analysis and Recognition (ICDAR)
Code
ACPR POSTER

Few-Shot Connectivity-Aware Text Line Segmentation in Historical Documents

R. Sterzinger , T. Lin , R. Sablatnig

Asian Conference on Pattern Recognition (ACPR)
Code
ICDAR POSTER

Drawing the Line: Deep Segmentation for Extracting Art from Ancient Etruscan Mirrors

R. Sterzinger , S. Brenner , R. Sablatnig

International Conference on Document Analysis and Recognition (ICDAR)
ICPR POSTER

Fusing Forces: Deep-Human-Guided Refinement of Segmentation Masks

R. Sterzinger , C. Stippel , R. Sablatnig

International Conference on Pattern Recognition (ICPR)
View all publications →