.png){kind=icon}
.png){kind=icon}
Return
Graph Learning via Edge Constrained Sparse Representation for image Analysis
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2019.2907301.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The construction of a graph essentially determines the performance of the graph-based image analysis methods. Particularly, the sparse graph is vital in image analysis because of its sparse and adaptive properties for the enormous scale of data. However, most existing graph-based algorithms ignore some valuable natural information, such as edge information of image data. In this paper, we propose a novel graph learning method, called edge constrained sparse representation (ECSR), which makes full use of edge information to refine the similarity among image samples. We believe that it is beneficial to include edge constraints into the construction stage of a graph for image analyses. Compared with conventional graph construction methods, ECSR has not only automatic sparsity and adaptive neighborhood size but also more accurate similarity measurements among natural images. The experimental results on four natural image datasets demonstrate the validity and effectiveness of ECSR for semi-supervised classification and clustering tasks.
Keywords:
Graph learning
edge constraint
sparse representation
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3.6
Papers:
9.8W
Citations:
29.4W
Organization
No organization information available