J. Dunker

About

J. Dunker’s research lies at the intersection of computer vision and robotics, with a focus on 3D object recognition and pose estimation. In their seminal 1996 work, Dunker introduced a pioneering approach to single-view 3D recognition using sets of prototypical views and spatially tolerant contour representations. This method demonstrated that extending 2D vision systems to 3D was both straightforward and computationally tractable, enabling robots to recognize objects from arbitrary orientations. The paper, with 17 citations, laid foundational groundwork for view-based recognition strategies. Dunker further advanced applied robotics with the SENROB vision system (2002, 13 citations), which integrated a color TV-camera mounted on a MANUTEC r2 robot’s hand. This system could detect, foveate, and recognize previously learned objects in arbitrary positions, while also allowing unknown objects to be interactively added to the system’s database. By bridging theoretical recognition frameworks with practical robotic manipulation, Dunker’s work has influenced both academic research and industrial automation, demonstrating how robust contour-based representations can enable real-time, adaptive robot vision in unstructured environments.

Research Focus

Key Achievements

Top Papers

1. 1
   Single view recognition and pose estimation of 3D objects using sets of prototypical views and spatially tolerant contour representations

   17 citations · 1996
2. 2
   The SENROB vision-system and its philosophy

   13 citations · 2002

Key Collaborators

Contact & Links