Home | Search | Help  
Home Page Università di Genova

Seminar Details

Date 21-12-2009
Time 11:00
Room/Location Sala conferenze DISI, 3 piano
Title Reliable shape-based recognition of objects in images
Speaker John Oliensis
Affiliation Stevens Institute of Technology
Link https://www.disi.unige.it/index.php?eventsandseminars/seminars
Abstract People understand an unexpected new image amazingly fast; how can an artificial recognition system do the same? As a first step, it must find image attributes distinctive enough to reduce the image's possible interpretations to a relatively small, manageable number, and researchers agree that this requires perceptual organization (PO), the identification of mid-size image structures such as salient lines, regions or object shapes. Yet the intrinsic ambiguities of images make PO unreliable, and recognition systems based on PO have not worked well. I present a recognition method that exploits PO without succumbing to its unreliability. The method recognizes objects by shape. It compares the shapes in two images by comparing the two images' segmentations (i.e., their partitions into homogeneous regions) as computed by PO, using a new similarity measure based on mutual information. To overcome the segmentations' unreliability, we compute the average similarity between all possible segmentations of the two images weighted by probability, in a simple, accurate closed-form approximation. The result is a robust measure which directly compares images according to their contained shapes, without needing to find the shapes explicitly. Other applications include tracking moving objects in video, computing stable image segmentations, and edge-preserving smoothing.
Back to Seminars