Heavy-Tailed Distances for Gradient Based Image Descriptors
Title | Heavy-Tailed Distances for Gradient Based Image Descriptors |
Publication Type | Conference Paper |
Year of Publication | 2011 |
Authors | Jia, Y., & Darrell T. |
Other Numbers | 3236 |
Abstract | Many applications in computer vision measure the similarity between images orimage patches based on some statistics such as oriented gradients. These are oftenmodeled implicitly or explicitly with a Gaussian noise assumption, leading tothe use of the Euclidean distance when comparing image descriptors. In this paper,we show that the statistics of gradient based image descriptors often followa heavy-tailed distribution, which undermines any principled motivation for theuse of Euclidean distances. We advocate for the use of a distance measure basedon the likelihood ratio test with appropriate probabilistic models that fit the empiricaldata distribution. We instantiate this similarity measure with the Gammacompound-Laplace distribution, and show significant improvement over existingdistance measures in the application of SIFT feature matching, at relatively lowcomputational cost. |
URL | https://www.icsi.berkeley.edu/pubs/vision/heavytailed11.pdf |
Bibliographic Notes | Proceedings of the 25th Annual Conference on Neural Information Processing Systems (NIPS 2011), Granada, Spain |
Abbreviated Authors | Y. Jia and T. Darrell |
ICSI Research Group | Vision |
ICSI Publication Type | Article in conference proceedings |