Frank Dittrich
Papers
4
Total Citations
20
H-Index
2
About
Frank Dittrich is a researcher at the forefront of safe human-robot collaboration, with a focus on computer vision and probabilistic risk modeling. His work centers on enabling robots to perceive and adapt to human presence in industrial settings, primarily through pixelwise object class segmentation and real-time optical flow estimation. Dittrich’s most cited paper, "Pixelwise object class segmentation based on synthetic data using an optimized training strategy" (2014, 14 citations), introduces a novel approach for low-level body part segmentation using RGB-D sensors mounted on ceilings, allowing robots to accurately identify human workers in shared workspaces. He further advances safety in human-robot cooperation by integrating high-performance optical flow with Bayesian filtering (2010), enabling real-time motion tracking on GPUs. Dittrich also developed a probabilistic risk modeling framework (2016) that allows robots to adapt their activities based on user-related information, using Bayesian Networks to reason about potential hazards. His contributions are critical for creating cognitive, adaptive systems that ensure safe collaboration between humans and robots in dynamic industrial environments, demonstrating a clear impact on the field of human-robot interaction.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4