Sparse representations for image decomposition with occlusions

Mike Donahue, Davi Geiger, Robert Hummel, Tyng Luh Liu

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

Abstract

We study the problem of how to detect 'interesting objects' appeared in a given image, I. Our approach is to treat it as a function approximation problem based on an over-redundant basis, and also account for occlusions, where the basis superposition principle is no longer valid. Since the basis (a library of image templates) is over-redundant, there are infinitely many ways to decompose I. We are motivated to select a sparse/compact representation of I, and to account for occlusions and noise. We then study a greedy and iterative 'weighted Lp Matching Pursuit' strategy, with 0 < p < 1. We use an Lp result to compute a solution, select the best template, at each stage of the pursuit.

Original languageEnglish (US)
Title of host publicationProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
PublisherIEEE
Pages7-12
Number of pages6
StatePublished - 1996
EventProceedings of the 1996 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - San Francisco, CA, USA
Duration: Jun 18 1996Jun 20 1996

Other

OtherProceedings of the 1996 IEEE Computer Society Conference on Computer Vision and Pattern Recognition
CitySan Francisco, CA, USA
Period6/18/966/20/96

Fingerprint

Decomposition

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Software
  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

Donahue, M., Geiger, D., Hummel, R., & Liu, T. L. (1996). Sparse representations for image decomposition with occlusions. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (pp. 7-12). IEEE.

Sparse representations for image decomposition with occlusions. / Donahue, Mike; Geiger, Davi; Hummel, Robert; Liu, Tyng Luh.

Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE, 1996. p. 7-12.

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

Donahue, M, Geiger, D, Hummel, R & Liu, TL 1996, Sparse representations for image decomposition with occlusions. in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE, pp. 7-12, Proceedings of the 1996 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, CA, USA, 6/18/96.
Donahue M, Geiger D, Hummel R, Liu TL. Sparse representations for image decomposition with occlusions. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE. 1996. p. 7-12
Donahue, Mike ; Geiger, Davi ; Hummel, Robert ; Liu, Tyng Luh. / Sparse representations for image decomposition with occlusions. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE, 1996. pp. 7-12
@inproceedings{1c358d810b0e49cda6a08be67032fe5d,
title = "Sparse representations for image decomposition with occlusions",
abstract = "We study the problem of how to detect 'interesting objects' appeared in a given image, I. Our approach is to treat it as a function approximation problem based on an over-redundant basis, and also account for occlusions, where the basis superposition principle is no longer valid. Since the basis (a library of image templates) is over-redundant, there are infinitely many ways to decompose I. We are motivated to select a sparse/compact representation of I, and to account for occlusions and noise. We then study a greedy and iterative 'weighted Lp Matching Pursuit' strategy, with 0 < p < 1. We use an Lp result to compute a solution, select the best template, at each stage of the pursuit.",
author = "Mike Donahue and Davi Geiger and Robert Hummel and Liu, {Tyng Luh}",
year = "1996",
language = "English (US)",
pages = "7--12",
booktitle = "Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
publisher = "IEEE",

}

TY - GEN

T1 - Sparse representations for image decomposition with occlusions

AU - Donahue, Mike

AU - Geiger, Davi

AU - Hummel, Robert

AU - Liu, Tyng Luh

PY - 1996

Y1 - 1996

N2 - We study the problem of how to detect 'interesting objects' appeared in a given image, I. Our approach is to treat it as a function approximation problem based on an over-redundant basis, and also account for occlusions, where the basis superposition principle is no longer valid. Since the basis (a library of image templates) is over-redundant, there are infinitely many ways to decompose I. We are motivated to select a sparse/compact representation of I, and to account for occlusions and noise. We then study a greedy and iterative 'weighted Lp Matching Pursuit' strategy, with 0 < p < 1. We use an Lp result to compute a solution, select the best template, at each stage of the pursuit.

AB - We study the problem of how to detect 'interesting objects' appeared in a given image, I. Our approach is to treat it as a function approximation problem based on an over-redundant basis, and also account for occlusions, where the basis superposition principle is no longer valid. Since the basis (a library of image templates) is over-redundant, there are infinitely many ways to decompose I. We are motivated to select a sparse/compact representation of I, and to account for occlusions and noise. We then study a greedy and iterative 'weighted Lp Matching Pursuit' strategy, with 0 < p < 1. We use an Lp result to compute a solution, select the best template, at each stage of the pursuit.

UR - http://www.scopus.com/inward/record.url?scp=0029701007&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029701007&partnerID=8YFLogxK

M3 - Conference contribution

SP - 7

EP - 12

BT - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

PB - IEEE

ER -