| Speaker | Ian Simon |
| Date | October 13, 2006 |
| Time | 3:00PM to 4:00PM |
| Place | GRAIL (CSE 291) |
Object recognition and image segmentation are two of the central problems in computer vision. Recently, methods based on local features have greatly improved the state of the art in object recognition. In this talk, I will discuss our work on performing object recognition and image segmentation simultaneously, using local feature methods. More specifically, given a set of reference objects, we perform segmentation on a target image by labeling each pixel as belonging to one (or none) of the reference objects.