Versatile mobile robots need to find and interact with objects in their environment. I am interested in the search part. Mainly, what are the vision algorithms, spatial representations and search methods needed to visually localize objects in large environments?
In this regard, I have come up with four main contributions that each spawned a project on its own:
Learning 3D context from Kinect images Everyday ObjectsUnderstanding indoor environments
Active Visual Search in Unexplored Large Environments
Kinect@Home: Crowdsourcing in the wild Kinect images
News
- Feb. 2012: I am co-organising the semantic perception workshop at ICRA 2012!
- Feb. 2011: I am organising a workshop at IROS 2011, check it out!
Research Interests
- Active Visual Search
- Semantic Mapping
- Machine Learning
Collaborators
- Andrzej Pronobis, KTH, Sweden
- Kristoffer Sjöö, KTH, Sweden
- Dejan Pangercic , TUM, Germany
- Bhaskara Marthi, Willow Garage, USA
- Dirk Holz , University of Bonn, Germany