Prices and Awards
Best Paper Award – (IV 2021)
We received the Best Paper Award – Third Place at IV 2021 for our paper Panoramic Panoptic Segmentation: Towards Complete Surrounding Understanding via Unsupervised Contrastive Learning by Alexander Jaus, Kailun Yang, and Rainer Stiefelhagen.
The paper introduces the task of panoramic panoptic segmentation to provide holistic 360° scene understanding for intelligent vehicles. We propose the Panoramic Robust Feature (PRF) framework using unsupervised contrastive learning to improve generalization on unseen panoramic domains.
HIDSS4Health Scholarship
I am a recipient of the Helmholtz Information & Data Science School for Health (HIDSS4Health) scholarship, a structured doctoral program jointly operated by the Karlsruhe Institute of Technology (KIT), the German Cancer Research Center (DKFZ), and Heidelberg University. The program brings together over 40 interdisciplinary research groups at the intersection of data science and health.
As a HIDSS4Health scholar, I am engaged in research projects spanning imaging, diagnostics, surgical technologies, and personalized medicine. The curriculum combines data science and life sciences, and graduates are awarded a certificate as "Data Scientist" with specialization in health-related applications.
AutoPET III – Data-Centric Track Award
Our approach (zero sugar) Tracer-Aware data filtering appraoch was awarded the second place in the Data-Centric Track of the AutoPET III Challenge at MICCAI 2024.
AutoPET is a well-regarded and challenging international competition in medical image analysis, with participation from over 490 teams worldwide. It focuses on robust lesion segmentation in whole-body PET/CT scans across different tracers and clinical centers.
CVPR 2024 – Segment Anything In Medical Images On Laptop
Our work on Revisiting classical segmentation methods achieved 6th place in the Segment Anything in Medical Images on Laptop Challenge at CVPR 2024.
Our paper Filters, Thresholds, and Geodesic Distances won 1st place in the Scribble-based Track of the CVPR 2024 MedSAM Challenge.
The challenge focused on efficient interactive segmentation across 11 medical imaging modalities. We showed that classical, low-resource methods—such as thresholding, k-means, and interpolation—can match or outperform vision foundation models like MedSAM in both performance and efficiency.
