Spatially Localized Visual Dictionary Learning

Valentin Leveau 1, 2 Alexis Joly 1 Olivier Buisson 2 Patrick Valduriez 1
1 ZENITH - Scientific Data Management
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier, CRISAM - Inria Sophia Antipolis - Méditerranée
Abstract : This paper addresses the challenge of devising new representation learning algorithms that overcome the lack of interpretability of classical visual models. Therefore, it introduces a new recursive visual patch selection technique built on top of a Shared Nearest Neighbors embedding method. The main contribution of the paper is to drastically reduce the high-dimensionality of such over-complete representation thanks to a recursive feature elimination method. We show that the number of spatial atoms of the representation can be reduced by up to two orders of magnitude without much degrading the encoded information. The resulting representations are shown to provide competitive image classification performance with the state-of-the-art while enabling to learn highly interpretable visual models.
Document type :
Conference papers
Complete list of metadatas

Cited literature [13 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01373778
Contributor : Alexis Joly <>
Submitted on : Wednesday, October 5, 2016 - 12:47:07 PM
Last modification on : Wednesday, October 9, 2019 - 11:44:04 AM
Long-term archiving on: Friday, January 6, 2017 - 1:34:47 PM

File

LeveauICMR2016.pdf
Files produced by the author(s)

Identifiers

Collections

Citation

Valentin Leveau, Alexis Joly, Olivier Buisson, Patrick Valduriez. Spatially Localized Visual Dictionary Learning. ICMR: International Conference on Multimedia Retrieval, Jun 2016, New York, United States. pp.367-370, ⟨10.1145/2911996.2912070⟩. ⟨hal-01373778⟩

Share

Metrics

Record views

542

Files downloads

295