Gulfize Coskun
Papers
3
Total Citations
248
H-Index
3
About
Gulfize Coskun is a leading researcher at the intersection of computer vision and medical robotics, whose work is fundamentally advancing the capabilities of surgical navigation systems. Her primary research focuses on developing robust Simultaneous Localization and Mapping (SLAM) and depth estimation methods specifically tailored for the challenging environment of endoscopic surgery. Coskun’s most significant contribution is the creation of the **EndoSLAM dataset**, a comprehensive benchmark that, for the first time, enabled effective quantitative evaluation of SLAM methods in endoscopy. This foundational work, detailed in her highly cited 2021 paper (240 citations), directly addressed a critical gap in the field by providing the standardized data necessary for reproducible research. Building on this, she pioneered **Endo-SfMLearner**, an unsupervised deep learning approach for monocular visual odometry and depth estimation. By removing the need for ground-truth depth data during training, her method offers a practical and scalable solution for dense 3D reconstruction from standard endoscopic video. Coskun’s work is not merely academic; it provides the essential tools for creating more accurate, real-time surgical guidance systems, directly impacting the future of minimally invasive surgery.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Quantitative Evaluation of Endoscopic SLAM Methods: EndoSLAM Dataset.3 citations · 2020