Locality-Constrained Multi-Task Joint Sparse Representation for Image Classification

Lihua GUO

doi:10.1587/transinf.E96.D.2177

Locality-Constrained Multi-Task Joint Sparse Representation for Image Classification

Lihua GUO

Full Text Views

0

Share
Cite this

Summary :

In the image classification applications, the test sample with multiple man-handcrafted descriptions can be sparsely represented by a few training subjects. Our paper is motivated by the success of multi-task joint sparse representation (MTJSR), and considers that the different modalities of features not only have the constraint of joint sparsity across different tasks, but also have the constraint of local manifold structure across different features. We introduce the constraint of local manifold structure into the MTJSR framework, and propose the Locality-constrained multi-task joint sparse representation method (LC-MTJSR). During the optimization of the formulated objective, the stochastic gradient descent method is used to guarantee fast convergence rate, which is essential for large-scale image categorization. Experiments on several challenging object classification datasets show that our proposed algorithm is better than the MTJSR, and is competitive with the state-of-the-art multiple kernel learning methods.

Publication: IEICE TRANSACTIONS on Information Vol.E96-D No.9 pp.2177-2181

Publication Date: 2013/09/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.E96.D.2177

Type of Manuscript: LETTER

Category: Image Recognition, Computer Vision

Authors

Lihua GUO
South China University of Technology, Guangzhou

Keyword

image classification, multi-task learning, sparse representation, manifold learning

Cite this

Copy

Lihua GUO, "Locality-Constrained Multi-Task Joint Sparse Representation for Image Classification" in IEICE TRANSACTIONS on Information, vol. E96-D, no. 9, pp. 2177-2181, September 2013, doi: 10.1587/transinf.E96.D.2177.
Abstract: In the image classification applications, the test sample with multiple man-handcrafted descriptions can be sparsely represented by a few training subjects. Our paper is motivated by the success of multi-task joint sparse representation (MTJSR), and considers that the different modalities of features not only have the constraint of joint sparsity across different tasks, but also have the constraint of local manifold structure across different features. We introduce the constraint of local manifold structure into the MTJSR framework, and propose the Locality-constrained multi-task joint sparse representation method (LC-MTJSR). During the optimization of the formulated objective, the stochastic gradient descent method is used to guarantee fast convergence rate, which is essential for large-scale image categorization. Experiments on several challenging object classification datasets show that our proposed algorithm is better than the MTJSR, and is competitive with the state-of-the-art multiple kernel learning methods.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.E96.D.2177/_p

Copy

@ARTICLE{e96-d_9_2177,
author={Lihua GUO, },
journal={IEICE TRANSACTIONS on Information},
title={Locality-Constrained Multi-Task Joint Sparse Representation for Image Classification},
year={2013},
volume={E96-D},
number={9},
pages={2177-2181},
abstract={In the image classification applications, the test sample with multiple man-handcrafted descriptions can be sparsely represented by a few training subjects. Our paper is motivated by the success of multi-task joint sparse representation (MTJSR), and considers that the different modalities of features not only have the constraint of joint sparsity across different tasks, but also have the constraint of local manifold structure across different features. We introduce the constraint of local manifold structure into the MTJSR framework, and propose the Locality-constrained multi-task joint sparse representation method (LC-MTJSR). During the optimization of the formulated objective, the stochastic gradient descent method is used to guarantee fast convergence rate, which is essential for large-scale image categorization. Experiments on several challenging object classification datasets show that our proposed algorithm is better than the MTJSR, and is competitive with the state-of-the-art multiple kernel learning methods.},
keywords={},
doi={10.1587/transinf.E96.D.2177},
ISSN={1745-1361},
month={September},}

Copy

TY - JOUR
TI - Locality-Constrained Multi-Task Joint Sparse Representation for Image Classification
T2 - IEICE TRANSACTIONS on Information
SP - 2177
EP - 2181
AU - Lihua GUO
PY - 2013
DO - 10.1587/transinf.E96.D.2177
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E96-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2013
AB - In the image classification applications, the test sample with multiple man-handcrafted descriptions can be sparsely represented by a few training subjects. Our paper is motivated by the success of multi-task joint sparse representation (MTJSR), and considers that the different modalities of features not only have the constraint of joint sparsity across different tasks, but also have the constraint of local manifold structure across different features. We introduce the constraint of local manifold structure into the MTJSR framework, and propose the Locality-constrained multi-task joint sparse representation method (LC-MTJSR). During the optimization of the formulated objective, the stochastic gradient descent method is used to guarantee fast convergence rate, which is essential for large-scale image categorization. Experiments on several challenging object classification datasets show that our proposed algorithm is better than the MTJSR, and is competitive with the state-of-the-art multiple kernel learning methods.
ER -