The Community for Technology Leaders
2009 IEEE Conference on Computer Vision and Pattern Recognition (2009)
Miami, FL, USA
June 20, 2009 to June 25, 2009
ISBN: 978-1-4244-3992-8
pp: 2254-2261
Yong Jae Lee , Univ. of Texas at Austin, Austin, TX, USA
K. Grauman , Univ. of Texas at Austin, Austin, TX, USA
ABSTRACT
Can we discover common object shapes within unlabeled multi-category collections of images? While often a critical cue at the category-level, contour matches can be difficult to isolate reliably from edge clutter-even within labeled images from a known class, let alone unlabeled examples. We propose a shape discovery method in which local appearance (patch) matches serve to anchor the surrounding edge fragments, yielding a more reliable affinity function for images that accounts for both shape and appearance. Spectral clustering from the initial affinities provides candidate object clusters. Then, we compute the within-cluster match patterns to discern foreground edges from clutter, attributing higher weight to edges more likely to belong to a common object. In addition to discovering the object contours in each image, we show how to summarize what is found with prototypical shapes. Our results on benchmark datasets demonstrate the approach can successfully discover shapes from unlabeled images.
INDEX TERMS
object contour, shape discovery, unlabeled image collection, spectral clustering
CITATION

Yong Jae Lee and K. Grauman, "Shape discovery from unlabeled image collections," 2009 IEEE Conference on Computer Vision and Pattern Recognition(CVPR), Miami, FL, USA, 2009, pp. 2254-2261.
doi:10.1109/CVPRW.2009.5206698
184 ms
(Ver 3.3 (11022016))