Discriminative sparse coding on multi-manifolds

Jim Jing Yan Wang, Halima Bensmail, Nan Yao, Xin Gao

Research output: Contribution to journalArticle

30 Citations (Scopus)

Abstract

Sparse coding has been popularly used as an effective data representation method in various applications, such as computer vision, medical imaging and bioinformatics. However, the conventional sparse coding algorithms and their manifold-regularized variants (graph sparse coding and Laplacian sparse coding), learn codebooks and codes in an unsupervised manner and neglect class information that is available in the training set. To address this problem, we propose a novel discriminative sparse coding method based on multi-manifolds, that learns discriminative class-conditioned codebooks and sparse codes from both data feature spaces and class labels. First, the entire training set is partitioned into multiple manifolds according to the class labels. Then, we formulate the sparse coding as a manifold-manifold matching problem and learn class-conditioned codebooks and codes to maximize the manifold margins of different classes. Lastly, we present a data sample-manifold matching-based strategy to classify the unlabeled data samples. Experimental results on somatic mutations identification and breast tumor classification based on ultrasonic images demonstrate the efficacy of the proposed data representation and classification approach.

Original languageEnglish
Pages (from-to)199-206
Number of pages8
JournalKnowledge-Based Systems
Volume54
DOIs
Publication statusPublished - 1 Jan 2013

Fingerprint

Labels
Medical imaging
Bioinformatics
Computer vision
Tumors
Ultrasonics

Keywords

  • Data representation
  • Large margins
  • Multi-manifolds
  • Sparse coding

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence
  • Management Information Systems
  • Information Systems and Management

Cite this

Discriminative sparse coding on multi-manifolds. / Wang, Jim Jing Yan; Bensmail, Halima; Yao, Nan; Gao, Xin.

In: Knowledge-Based Systems, Vol. 54, 01.01.2013, p. 199-206.

Research output: Contribution to journalArticle

Wang, Jim Jing Yan ; Bensmail, Halima ; Yao, Nan ; Gao, Xin. / Discriminative sparse coding on multi-manifolds. In: Knowledge-Based Systems. 2013 ; Vol. 54. pp. 199-206.
@article{087849fb21094d8d87c7bc54685a96e5,
title = "Discriminative sparse coding on multi-manifolds",
abstract = "Sparse coding has been popularly used as an effective data representation method in various applications, such as computer vision, medical imaging and bioinformatics. However, the conventional sparse coding algorithms and their manifold-regularized variants (graph sparse coding and Laplacian sparse coding), learn codebooks and codes in an unsupervised manner and neglect class information that is available in the training set. To address this problem, we propose a novel discriminative sparse coding method based on multi-manifolds, that learns discriminative class-conditioned codebooks and sparse codes from both data feature spaces and class labels. First, the entire training set is partitioned into multiple manifolds according to the class labels. Then, we formulate the sparse coding as a manifold-manifold matching problem and learn class-conditioned codebooks and codes to maximize the manifold margins of different classes. Lastly, we present a data sample-manifold matching-based strategy to classify the unlabeled data samples. Experimental results on somatic mutations identification and breast tumor classification based on ultrasonic images demonstrate the efficacy of the proposed data representation and classification approach.",
keywords = "Data representation, Large margins, Multi-manifolds, Sparse coding",
author = "Wang, {Jim Jing Yan} and Halima Bensmail and Nan Yao and Xin Gao",
year = "2013",
month = "1",
day = "1",
doi = "10.1016/j.knosys.2013.09.004",
language = "English",
volume = "54",
pages = "199--206",
journal = "Knowledge-Based Systems",
issn = "0950-7051",
publisher = "Elsevier",

}

TY - JOUR

T1 - Discriminative sparse coding on multi-manifolds

AU - Wang, Jim Jing Yan

AU - Bensmail, Halima

AU - Yao, Nan

AU - Gao, Xin

PY - 2013/1/1

Y1 - 2013/1/1

N2 - Sparse coding has been popularly used as an effective data representation method in various applications, such as computer vision, medical imaging and bioinformatics. However, the conventional sparse coding algorithms and their manifold-regularized variants (graph sparse coding and Laplacian sparse coding), learn codebooks and codes in an unsupervised manner and neglect class information that is available in the training set. To address this problem, we propose a novel discriminative sparse coding method based on multi-manifolds, that learns discriminative class-conditioned codebooks and sparse codes from both data feature spaces and class labels. First, the entire training set is partitioned into multiple manifolds according to the class labels. Then, we formulate the sparse coding as a manifold-manifold matching problem and learn class-conditioned codebooks and codes to maximize the manifold margins of different classes. Lastly, we present a data sample-manifold matching-based strategy to classify the unlabeled data samples. Experimental results on somatic mutations identification and breast tumor classification based on ultrasonic images demonstrate the efficacy of the proposed data representation and classification approach.

AB - Sparse coding has been popularly used as an effective data representation method in various applications, such as computer vision, medical imaging and bioinformatics. However, the conventional sparse coding algorithms and their manifold-regularized variants (graph sparse coding and Laplacian sparse coding), learn codebooks and codes in an unsupervised manner and neglect class information that is available in the training set. To address this problem, we propose a novel discriminative sparse coding method based on multi-manifolds, that learns discriminative class-conditioned codebooks and sparse codes from both data feature spaces and class labels. First, the entire training set is partitioned into multiple manifolds according to the class labels. Then, we formulate the sparse coding as a manifold-manifold matching problem and learn class-conditioned codebooks and codes to maximize the manifold margins of different classes. Lastly, we present a data sample-manifold matching-based strategy to classify the unlabeled data samples. Experimental results on somatic mutations identification and breast tumor classification based on ultrasonic images demonstrate the efficacy of the proposed data representation and classification approach.

KW - Data representation

KW - Large margins

KW - Multi-manifolds

KW - Sparse coding

UR - http://www.scopus.com/inward/record.url?scp=84901765974&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84901765974&partnerID=8YFLogxK

U2 - 10.1016/j.knosys.2013.09.004

DO - 10.1016/j.knosys.2013.09.004

M3 - Article

AN - SCOPUS:84901765974

VL - 54

SP - 199

EP - 206

JO - Knowledge-Based Systems

JF - Knowledge-Based Systems

SN - 0950-7051

ER -