The web contains vast collections of unstructured image databases that could be exploited to generated useful models for image recognition. Search engines currently only use keywords found in image filenames rather than the image data itself. However, the thousands of images retrieved in such a search could be used to generate models for recognizing specific objects. The goal of this research project is to develop the infrastructure for a new research competition designed to push the state of the art in image understanding and automatic acquisition of knowledge from large unstructured databases of images, such as those generally found on the web. This competition will be held at the AAAI (American Association for Artificial Intelligence) Mobile Robot Competition and Exhibition which has been co-located with the AAAI conference for the past ten years. By providing the necessary infrastructure for organizing and managing such a competition, there is a strong chance to attract the best and the brightest in the computer vision, robotics and AI research fields to push the envelope of the state of the art in image and video understanding technologies. Competitions such as these provide a standardized testbed on which researchers can compare the results of their research. At the end of the competition, the PIs will hold a workshop so that the specific technical aspects of each entry can be presented and discussed. In addition, the competitors will have to share their code under an open source license, so that newcomers of 2008 will stand on the shoulders of the winning teams of 2007.
Project URL: www.cs.cmu.edu/prybski/AAAI RVSC
