DMCA
Fast Image Search for Learned Metrics
Cached
Download Links
| Citations: | 103 - 11 self |
BibTeX
@MISC{Jain_fastimage,
author = {Prateek Jain and et al.},
title = {Fast Image Search for Learned Metrics},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We introduce a method that enables scalable image search for learned metrics. Given pairwise similarity and dissimilarity constraints between some images, we learn a Mahalanobis distance function that captures the images’ underlying relationships well. To allow sub-linear time similarity search under the learned metric, we show how to encode the learned metric parameterization into randomized locality-sensitive hash functions. We further formulate an indirect solution that enables metric learning and hashing for vector spaces whose high dimensionality make it infeasible to learn an explicit weighting over the feature dimensions. We demonstrate the approach applied to a variety of image datasets. Our learned metrics improve accuracy relative to commonly-used metric baselines, while our hashing construction enables efficient indexing with learned distances and very large databases.
Keyphrases
learned metric fast image search locality-sensitive hash function vector space commonly-used metric baseline learned distance image datasets sub-linear time similarity search feature dimension mahalanobis distance function indirect solution explicit weighting large database scalable image search learned metric parameterization dissimilarity constraint pairwise similarity high dimensionality metric learning
