Welcome to the InfoSci Platform
Graph-Based Spam Image Detection for Mobile Phone Spam Image Filtering

Graph-Based Spam Image Detection for Mobile Phone Spam Image Filtering

So Yeon Kim, Kyung-Ah Sohn
Copyright: © 2015 |Volume: 3 |Issue: 4 |Pages: 15
ISSN: 2166-7160|EISSN: 2166-7179|EISBN13: 9781466680623|DOI: 10.4018/IJSI.2015100106
Cite Article Cite Article

MLA

Kim, So Yeon, and Kyung-Ah Sohn. "Graph-Based Spam Image Detection for Mobile Phone Spam Image Filtering." IJSI vol.3, no.4 2015: pp.72-86. http://doi.org/10.4018/IJSI.2015100106

APA

Kim, S. Y. & Sohn, K. (2015). Graph-Based Spam Image Detection for Mobile Phone Spam Image Filtering. International Journal of Software Innovation (IJSI), 3(4), 72-86. http://doi.org/10.4018/IJSI.2015100106

Chicago

Kim, So Yeon, and Kyung-Ah Sohn. "Graph-Based Spam Image Detection for Mobile Phone Spam Image Filtering," International Journal of Software Innovation (IJSI) 3, no.4: 72-86. http://doi.org/10.4018/IJSI.2015100106

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

Spam images in mobile phones have increasingly appeared these days. As the spam filtering systems become more sophisticated, spams are being more intelligent. Although detection of email-spams has been quite successful, there have not been effective solutions for detecting mobile phone spams yet, especially, spam images. In addition to the expensive image processing time, insufficient spam image data in mobile phones makes it challenging to train a general model. To address this issue, the authors propose a graph-based approach that utilizes graph structure in abundant e-mail spam dataset. The authors employ different clustering algorithms to find a subset of e-mail spam images similar to phone spam images. Furthermore, the performance behavior with respect to different image descriptors of Pyramid Histogram of Visual Words (PHOW) and RGB histogram is extensively investigated. The authors' results highlight that the proposed idea is fairly meaningful in increasing training data size, thus effectively improving image spam detection performance.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.