Webcast: A Search Engine for the Real World, or, A Top-Down Approach to Vision

An August 2006 presentation sponsored by Microsoft Research at the University of British Columbia. Kevin Murphy is the speaker. The talk runs about 1 hour. From the description:

We consider the problem of finding instances of visual object categories (such as a cup or a pen) in cluttered, real-world environments. We propose a hierarchical approach, whereby we first categorize the scene (outdoors or indoors? kitchen or office?), then we use global image statistics (the “gist” of the image) to predict where to look within the image, and finally we run an object detector (based on boosted random fields) to localize the object, exploiting spatial constraints with other, easier-to-detect objects. We argue that this top-down approach is not only faster than standard “brute force” approaches, but also reduces the error-rate, since all decisions are made in context.

.

This entry was posted on Wednesday, October 4th, 2006 at 12:57 am and is filed under Info Management and Retrieval. You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.