Learning dictionaries for local sparse coding in image classification

Jayaraman J. Thiagarajan, Andreas Spanias

Research output: Chapter in Book/Report/Conference proceedingConference contribution

11 Scopus citations

Abstract

Low dimensional embedding of data samples lying on a manifold can be performed using locally linear modeling. By incorporating suitable locality constraints, sparse coding can be adapted to modeling local regions of a manifold. This has been coupled with the spatial pyramid matching algorithm to achieve state-of-the-art performance in object recognition. In this paper, we propose an algorithm to learn dictionaries for computing local sparse codes of descriptors extracted from image patches. The algorithm iterates between a local sparse coding step and an update step that searches for a better dictionary. Evaluation of the local sparse code for a data sample is simplified by first estimating its neighbors using the proposed distance metric and then computing the minimum ℓ 1 solution using only the neighbors. The proposed dictionary update ensures that the neighborhood of a training sample is not changed from one iteration to the next. Simulation results demonstrate that the sparse codes computed using the proposed dictionary achieve improved classification accuracies when compared to using a K-means dictionary with standard image datasets.

Original languageEnglish (US)
Title of host publicationConference Record of the 45th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2011
Pages2014-2018
Number of pages5
DOIs
StatePublished - Dec 1 2011
Event45th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2011 - Pacific Grove, CA, United States
Duration: Nov 6 2011Nov 9 2011

Publication series

NameConference Record - Asilomar Conference on Signals, Systems and Computers
ISSN (Print)1058-6393

Other

Other45th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2011
CountryUnited States
CityPacific Grove, CA
Period11/6/1111/9/11

Keywords

  • Local sparse codes
  • dictionary learning
  • linear classifiers
  • sparse representations

ASJC Scopus subject areas

  • Signal Processing
  • Computer Networks and Communications

Fingerprint Dive into the research topics of 'Learning dictionaries for local sparse coding in image classification'. Together they form a unique fingerprint.

Cite this