Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Online invariance selection for local feature descriptors

Pautrat, Rémi ; Larsson, Viktor LU ; Oswald, Martin R and Pollefeys, Marc (2020) 16th European Conference on Computer Vision, ECCV 2020 In Lecture Notes in Computer Science 12347. p.707-724
Abstract
To be invariant, or not to be invariant: that is the question formulated in this work about local descriptors. A limitation of current feature descriptors is the trade-off between generalization and discriminative power: more invariance means less informative descriptors. We propose to overcome this limitation with a disentanglement of invariance in local descriptors and with an online selection of the most appropriate invariance given the context. Our framework (https://github.com/rpautrat/LISRD) consists in a joint learning of multiple local descriptors with different levels of invariance and of meta descriptors encoding the regional variations of an image. The similarity of these meta descriptors across images is used to select the... (More)
To be invariant, or not to be invariant: that is the question formulated in this work about local descriptors. A limitation of current feature descriptors is the trade-off between generalization and discriminative power: more invariance means less informative descriptors. We propose to overcome this limitation with a disentanglement of invariance in local descriptors and with an online selection of the most appropriate invariance given the context. Our framework (https://github.com/rpautrat/LISRD) consists in a joint learning of multiple local descriptors with different levels of invariance and of meta descriptors encoding the regional variations of an image. The similarity of these meta descriptors across images is used to select the right invariance when matching the local descriptors. Our approach, named Local Invariance Selection at Runtime for Descriptors (LISRD), enables descriptors to adapt to adverse changes in images, while remaining discriminative when invariance is not required. We demonstrate that our method can boost the performance of current descriptors and outperforms state-of-the-art descriptors in several matching tasks, when evaluated on challenging datasets with day-night illumination as well as viewpoint changes. (Less)
Please use this url to cite or link to this publication:
author
; ; and
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
host publication
Computer Vision – ECCV 2020 : 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part II - 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part II
series title
Lecture Notes in Computer Science
volume
12347
pages
18 pages
publisher
Springer
conference name
16th European Conference on Computer Vision, ECCV 2020
conference location
Glasgow, United Kingdom
conference dates
2020-08-23 - 2020-08-28
external identifiers
  • scopus:85097255043
ISSN
1611-3349
0302-9743
ISBN
978-3-030-58535-8
978-3-030-58536-5
DOI
10.1007/978-3-030-58536-5_42
language
English
LU publication?
no
id
81e48a1c-de5f-4c76-89d5-626032c26f8d
date added to LUP
2022-09-06 11:57:49
date last changed
2024-04-19 19:48:06
@inproceedings{81e48a1c-de5f-4c76-89d5-626032c26f8d,
  abstract     = {{To be invariant, or not to be invariant: that is the question formulated in this work about local descriptors. A limitation of current feature descriptors is the trade-off between generalization and discriminative power: more invariance means less informative descriptors. We propose to overcome this limitation with a disentanglement of invariance in local descriptors and with an online selection of the most appropriate invariance given the context. Our framework (https://github.com/rpautrat/LISRD) consists in a joint learning of multiple local descriptors with different levels of invariance and of meta descriptors encoding the regional variations of an image. The similarity of these meta descriptors across images is used to select the right invariance when matching the local descriptors. Our approach, named Local Invariance Selection at Runtime for Descriptors (LISRD), enables descriptors to adapt to adverse changes in images, while remaining discriminative when invariance is not required. We demonstrate that our method can boost the performance of current descriptors and outperforms state-of-the-art descriptors in several matching tasks, when evaluated on challenging datasets with day-night illumination as well as viewpoint changes.}},
  author       = {{Pautrat, Rémi and Larsson, Viktor and Oswald, Martin R and Pollefeys, Marc}},
  booktitle    = {{Computer Vision – ECCV 2020 : 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part II}},
  isbn         = {{978-3-030-58535-8}},
  issn         = {{1611-3349}},
  language     = {{eng}},
  pages        = {{707--724}},
  publisher    = {{Springer}},
  series       = {{Lecture Notes in Computer Science}},
  title        = {{Online invariance selection for local feature descriptors}},
  url          = {{http://dx.doi.org/10.1007/978-3-030-58536-5_42}},
  doi          = {{10.1007/978-3-030-58536-5_42}},
  volume       = {{12347}},
  year         = {{2020}},
}