SegSALSA-STR

A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization

Filipe Condessa, Jose Bioucas-DIas, Jelena Kovacevic

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

Abstract

In this paper we present a supervised hyperspectral image segmentation algorithm based on a convex formulation of a marginal maximum a posteriori segmentation with hidden fields and structure tensor regularization: Segmentation via the Constraint Split Augmented Lagrangian Shrinkage by Structure Tensor Regularization (SegSALSA-STR). This formulation avoids the generally discrete nature of segmentation problems and the inherent NP-hardness of the integer optimization associated. We extend the Segmentation via the Constraint Split Augmented Lagrangian Shrinkage (SegSALSA) algorithm [1] by generalizing the vectorial total variation prior using a structure tensor prior constructed from a patch-based Jacobian [2]. The resulting algorithm is convex, time-efficient and highly parallelizable. This shows the potential of combining hidden fields with convex optimization through the inclusion of different regularizers. The SegSALSA-STR algorithm is validated in the segmentation of real hyperspectral images.

Original languageEnglish (US)
Title of host publication2015 7th Workshop on Hyperspectral Image and Signal Processing
Subtitle of host publicationEvolution in Remote Sensing, WHISPERS 2015
PublisherIEEE Computer Society
Volume2015-June
ISBN (Electronic)9781467390156
DOIs
StatePublished - Oct 19 2017
Event7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015 - Tokyo, Japan
Duration: Jun 2 2015Jun 5 2015

Other

Other7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015
CountryJapan
CityTokyo
Period6/2/156/5/15

Fingerprint

Image segmentation
Tensors
Convex optimization
Hardness

Keywords

  • Constrained Split Augmented Lagrangian Shrinkage Algorithm (SALSA)
  • hidden fields
  • Image segmentation
  • structure tensor regularization

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Condessa, F., Bioucas-DIas, J., & Kovacevic, J. (2017). SegSALSA-STR: A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization. In 2015 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015 (Vol. 2015-June). [8075464] IEEE Computer Society. https://doi.org/10.1109/WHISPERS.2015.8075464

SegSALSA-STR : A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization. / Condessa, Filipe; Bioucas-DIas, Jose; Kovacevic, Jelena.

2015 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015. Vol. 2015-June IEEE Computer Society, 2017. 8075464.

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

Condessa, F, Bioucas-DIas, J & Kovacevic, J 2017, SegSALSA-STR: A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization. in 2015 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015. vol. 2015-June, 8075464, IEEE Computer Society, 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015, Tokyo, Japan, 6/2/15. https://doi.org/10.1109/WHISPERS.2015.8075464
Condessa F, Bioucas-DIas J, Kovacevic J. SegSALSA-STR: A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization. In 2015 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015. Vol. 2015-June. IEEE Computer Society. 2017. 8075464 https://doi.org/10.1109/WHISPERS.2015.8075464
Condessa, Filipe ; Bioucas-DIas, Jose ; Kovacevic, Jelena. / SegSALSA-STR : A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization. 2015 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015. Vol. 2015-June IEEE Computer Society, 2017.
@inproceedings{8a9d88dd6623459492c0bd596dc7b363,
title = "SegSALSA-STR: A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization",
abstract = "In this paper we present a supervised hyperspectral image segmentation algorithm based on a convex formulation of a marginal maximum a posteriori segmentation with hidden fields and structure tensor regularization: Segmentation via the Constraint Split Augmented Lagrangian Shrinkage by Structure Tensor Regularization (SegSALSA-STR). This formulation avoids the generally discrete nature of segmentation problems and the inherent NP-hardness of the integer optimization associated. We extend the Segmentation via the Constraint Split Augmented Lagrangian Shrinkage (SegSALSA) algorithm [1] by generalizing the vectorial total variation prior using a structure tensor prior constructed from a patch-based Jacobian [2]. The resulting algorithm is convex, time-efficient and highly parallelizable. This shows the potential of combining hidden fields with convex optimization through the inclusion of different regularizers. The SegSALSA-STR algorithm is validated in the segmentation of real hyperspectral images.",
keywords = "Constrained Split Augmented Lagrangian Shrinkage Algorithm (SALSA), hidden fields, Image segmentation, structure tensor regularization",
author = "Filipe Condessa and Jose Bioucas-DIas and Jelena Kovacevic",
year = "2017",
month = "10",
day = "19",
doi = "10.1109/WHISPERS.2015.8075464",
language = "English (US)",
volume = "2015-June",
booktitle = "2015 7th Workshop on Hyperspectral Image and Signal Processing",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - SegSALSA-STR

T2 - A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization

AU - Condessa, Filipe

AU - Bioucas-DIas, Jose

AU - Kovacevic, Jelena

PY - 2017/10/19

Y1 - 2017/10/19

N2 - In this paper we present a supervised hyperspectral image segmentation algorithm based on a convex formulation of a marginal maximum a posteriori segmentation with hidden fields and structure tensor regularization: Segmentation via the Constraint Split Augmented Lagrangian Shrinkage by Structure Tensor Regularization (SegSALSA-STR). This formulation avoids the generally discrete nature of segmentation problems and the inherent NP-hardness of the integer optimization associated. We extend the Segmentation via the Constraint Split Augmented Lagrangian Shrinkage (SegSALSA) algorithm [1] by generalizing the vectorial total variation prior using a structure tensor prior constructed from a patch-based Jacobian [2]. The resulting algorithm is convex, time-efficient and highly parallelizable. This shows the potential of combining hidden fields with convex optimization through the inclusion of different regularizers. The SegSALSA-STR algorithm is validated in the segmentation of real hyperspectral images.

AB - In this paper we present a supervised hyperspectral image segmentation algorithm based on a convex formulation of a marginal maximum a posteriori segmentation with hidden fields and structure tensor regularization: Segmentation via the Constraint Split Augmented Lagrangian Shrinkage by Structure Tensor Regularization (SegSALSA-STR). This formulation avoids the generally discrete nature of segmentation problems and the inherent NP-hardness of the integer optimization associated. We extend the Segmentation via the Constraint Split Augmented Lagrangian Shrinkage (SegSALSA) algorithm [1] by generalizing the vectorial total variation prior using a structure tensor prior constructed from a patch-based Jacobian [2]. The resulting algorithm is convex, time-efficient and highly parallelizable. This shows the potential of combining hidden fields with convex optimization through the inclusion of different regularizers. The SegSALSA-STR algorithm is validated in the segmentation of real hyperspectral images.

KW - Constrained Split Augmented Lagrangian Shrinkage Algorithm (SALSA)

KW - hidden fields

KW - Image segmentation

KW - structure tensor regularization

UR - http://www.scopus.com/inward/record.url?scp=85039157831&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85039157831&partnerID=8YFLogxK

U2 - 10.1109/WHISPERS.2015.8075464

DO - 10.1109/WHISPERS.2015.8075464

M3 - Conference contribution

VL - 2015-June

BT - 2015 7th Workshop on Hyperspectral Image and Signal Processing

PB - IEEE Computer Society

ER -