| HAL-Inria | Publications, software ... of Inria's scientists |
Conference papers
Fast Spectral Ranking for Similarity Search
Abstract : Despite the success of deep learning on representing images for particular object retrieval, recent studies show that the learned representations still lie on manifolds in a high dimensional space. This makes the Euclidean nearest neighbor search biased for this task. Exploring the mani-folds online remains expensive even if a nearest neighbor graph has been computed offline. This work introduces an explicit embedding reducing manifold search to Euclidean search followed by dot product similarity search. This is equivalent to linear graph filtering of a sparse signal in the frequency domain. To speed up online search, we compute an approximate Fourier basis of the graph offline. We improve the state of art on particular object retrieval datasets including the challenging Instre dataset containing small objects. At a scale of 10 5 images, the offline cost is only a few hours, while query time is comparable to standard similarity search.
Cited literature [73 references]
https://hal.inria.fr/hal-01839744
Contributor : Teddy Furon <>
Submitted on : Monday, July 16, 2018 - 8:46:56 AM
Last modification on : Saturday, July 11, 2020 - 3:14:57 AM
Long-term archiving on: : Wednesday, October 17, 2018 - 12:41:07 PM
Contributor : Teddy Furon <>
Submitted on : Monday, July 16, 2018 - 8:46:56 AM
Last modification on : Saturday, July 11, 2020 - 3:14:57 AM
Long-term archiving on: : Wednesday, October 17, 2018 - 12:41:07 PM
File
Iscen_Fast_Spectral_Ranking_CV...
Files produced by the author(s)