Sabina B. van Rooij

Netherlands Organisation for Applied Scientific Research

Papers

Total Citations

H-Index

About

Sabina B. van Rooij is a rising researcher at the forefront of computer vision, with a particular focus on label-efficient segmentation and interactive machine learning. Her work addresses a critical bottleneck in modern AI: the need for large, manually annotated datasets. In her notable 2024 paper, "Guided SAM: Label-Efficient Part Segmentation," van Rooij introduces a novel framework that adapts the Segment Anything Model (SAM) to perform fine-grained part segmentation using only a handful of user-provided examples. This contribution significantly reduces the annotation burden for tasks ranging from medical imaging to autonomous driving, enabling models to learn detailed object components with minimal human input. While her citation count is still building—a natural stage for early-career impact—her work has already been recognized for its practical utility and technical elegance. By bridging the gap between powerful foundation models and real-world data scarcity, van Rooij is helping to democratize advanced segmentation, making it more accessible for researchers and practitioners alike. Her research promises to accelerate progress in areas where precise, part-level understanding is essential.