|International Computer Science and Engineering Society (ICSES)|
ICSES Transactions on Image Processing and Pattern Recognition
Vol. 4, No. 4, Dec. 2018
A Novel Graph-based Image Mining Technique Using Weighted Substructure | Book Chapter of Image Segmentation: A Guide to Image Mining
Vikramsingh Parihar a,, Roshani Nage b, Atul Dahane c
Highlights and Novelties
This paper presents a novel image mining approach based on weighted substructure. The problem is modeled in terms of creating a dataset of images and extracting the features of each image. Then graphs are generated for each image based on these features. There are many possible ways to obtain features of images from a graph but one of the most natural ways is to represent a graph is by a set of its substructure. The weight factor is used to measure the actual importance of each different substructure in a given graph dataset. On the basis of weighted substructure graphs the image mining process is done. For image mining, an external query image is provided by user. Its features are extracted and graph is generated. Later the substructure of query image is matched with the substructure of the dataset. The most closely matched substructures of images from the dataset are identified and it can be concluded that the identified images are close to the query image. The experiments are carried out on a dataset of 1000 natural as well as synthetic images from online resources and it is found that the mined images are most closely related to the query image.
© Copyright was transferred to International Computer Science and Engineering Society (ICSES) by all the Authors.
Cite this manuscript as
Vikramsingh Parihar, Roshani Nage, Atul Dahane, "A Novel Graph-based Image Mining Technique Using Weighted Substructure," in Image Segmentation: A Guide to Image Mining, 1st ed., ITIPPR: ICSES, 2018, pp. 16-25.
For External Scientific Databeses
Written by: Admin | Link ...