Papers
9
Total Citations
207
H-Index
6
About
Dimitrios Kosmopoulos is a leading researcher at the intersection of computer vision, robotics, and assistive technologies. His work spans three core areas: robust visual perception for robotic manipulation, statistical machine learning for sensor data, and socially assistive systems. Kosmopoulos made foundational contributions to superquadric-based segmentation, developing methods that fuse region and boundary information to recover 3D object models from range images—critical for bin-picking and depalletizing tasks (cited over 60 times collectively). His 2010 variational Bayesian methodology for hidden Markov models using Student’s-t mixtures (62 citations) advanced robust statistical modeling of sensor data. In visual servoing, he proposed a robust Jacobian estimation technique (59 citations) that improved robot control under uncertainty. Beyond industrial robotics, Kosmopoulos pioneered rehabilitation gaming systems for cerebral palsy patients and developed visual competencies for socially assistive robots in the HOBBIT project. His work on real-time depalletizing systems demonstrated practical deployment of active vision in manufacturing. With over 200 total citations across these domains, Kosmopoulos continues to bridge theoretical advances in perception and learning with tangible robotic applications that enhance both industrial automation and human quality of life.
Research Focus
Key Achievements
Top Papers
- 1
- 2Robust Jacobian matrix estimation for image-based visual servoing59 citations · 2010
- 3
- 4An efficient depalletizing system based on 2D range imagery25 citations · 2002
- 5A game system for remote rehabilitation of cerebral palsy patients11 citations · 2012
- 6Developing visual competencies for socially assistive robots7 citations · 2013
- 7
- 8MD-SIR: a methodology for developing sensor—guided industry robots4 citations · 2002
- 9