Doğukan Bağcı
© Doğukan Bağcı
- Master’s Student
Hi everyone!
My name is Doğukan, and I’m a robotics student at the Technical University of Darmstadt. Currently, I work as a Student Researcher at the Visual Inference Lab at TU Darmstadt.
**Research Interests:**
My research focuses on understanding how design choices in computer vision models affect their performance across multiple quality dimensions [1]. Recently, I’ve expanded my scope to include Explainable AI for
computer vision. Specifically, I’m interested in explaining the behavior of geometric models, such as keypoint detection and depth estimation.
**Publications:**
[1] Hesse et al. (2025). Beyond Accuracy: What Matters in Designing Well-Behaved Models? In this paper, we analyzed 326 vision models to introduce the QUBA score – a metric that evaluates model quality across
multiple dimensions. Our work sheds light on how training methods and model architectures influence these aspects.
**Awards:**
I’m proud to have received the Deutschlandstipendium three times (2021-2023) for my academic achievements.