Image Classification by Iterative Semi-Supervised Sparse Coding

Zheng, Haixia; Ip, Horace H. S.

doi:10.1007/978-3-319-03731-8_45

Cited by 1 publication

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It has been widely observed that better performance can be achieved with smooth solutions under this assumption [26]. The Graph-based SSL (GSSL) algorithms are typical SSL methods based on this assumption and attract much attention recently [19], [20], [21], [22].…”

Section: Related Workmentioning

confidence: 99%

“…Therefore, the method still belongs to the unsupervised method. [19] and [22] proposed similar methods that utilize the manifold structure of the labeled and unlabeled data and the label constraint provided by the labeled data to learn labels for the unlabeled data. Therefore, the manifold constraint in these methods is weak and no discriminative property is enforced for the learned labels.…”

Section: Related Workmentioning

confidence: 99%

“…Update the dictionary associated matrix B according to (22) Update the sparse coefficient S according to the Algorithm 2; 5.…”

Section: Optimization Of the Classifier Matrix Wmentioning

confidence: 99%

“…Therefore, developing methods to learn information from both labeled and unlabeled data is more desirable. Many SSL methods have been proposed to learn classifier or the labels of the unlabeled data by utilizing both the labeled and unlabeled data recently [19], [20], [21], [22]. Obviously, successful methods should be able to utilize the available labels of the data and feature vectors of the labeled and unlabeled data simultaneously to learn discriminative features.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Semi-supervised Sparse Representation with Graph Regularization for Image Classification

2020

Preprint

View full text Add to dashboard Cite

Image classification is a challenging problem for computer in reality. Large numbers of methods can achieve satisfying performances with sufficient labeled images. However, labeled images are still highly limited for certain image classification tasks. Instead, lots of unlabeled images are available and easy to be obtained. Therefore, making full use of the available unlabeled data can be a potential way to further improve the performance of current image classification methods. In this paper, we propose a discriminative semi-supervised sparse representation algorithm for image classification. In the algorithm, the classification process is combined with the sparse coding to learn a data-driven linear classifier. To obtain discriminative predictions, the predicted labels are regularized with three graphs, i.e., the global manifold structure graph, the within-class graph and the between-classes graph. The constructed graphs are able to extract structure information included in both the labeled and unlabeled data. Moreover, the proposed method is extended to a kernel version for dealing with data that cannot be linearly classified. Accordingly, efficient algorithms are developed to solve the corresponding optimization problems. Experimental results on several challenging databases demonstrate that the proposed algorithm achieves excellent performances compared with related popular methods.

show abstract