We won the First place in the Semantic Robot Vision Challenge by AAAI!

Selected objects we found are shown in the results below. We used object models with good generalization, which allowed us to be the only team to find a "general" object (an unbranded object) without the use of template matching techniques.

Team Members:


Fei-Fei Li Princeton/UIUC	Jia Li UIUC	Juan Carlos Niebles UIUC/U. del Norte	Rahul Mehta UIUC	Richard Socher Princeton	Brendan Collins Princeton
Assistant Professor	ECE Grad Student	ECE Grad Student	ECE Undergrad Student	Visiting Grad Student	CS Undergrad Student

Some Example Results:

Ritter Sport Marzipan	DVD Gladiator	DVD Shrek
CD Hey Eugene by Pink Martini	Tide	Lindt Madagascar
	Scientific Calculator

References:

Li-Jia Li, Gang Wang and Li Fei-Fei. OPTIMOL: automatic Object Picture collecTion via Incremental MOdel Learning. CVPR 2007. [PDF] [Project page]

Li-Jia Li, Juan Carlos Niebles and Li Fei-Fei. OPTIMOL: a framework for Online Picture collecTion via Incremental MOdel Learning. AAAI 2007. [Conference Link] [Slides]

In the news:

Illinois-Princeton team takes first place in robot vision competition in UIUC ECE Headline News. August 22, 2007.

Robots surf the web to learn about the world in New Scientist. August 17, 2007.

Princeton Wins Robot Vision Competition in Princeton CS News. July 31, 2007.

UIUC-Princeton Team

Team Members:

Some Example Results:

References:

In the news: