Australian Digital Forensics Conference

Document Type

Conference Proceeding

Abstract

Blind steganalysis performance is influenced by several factors including the features used for classification. This paper investigates the suitability of using lacunarity measure as a potential feature vectorfor blind steganalysis. Differential Box Counting (DBC) based lacunarity measure has been employed using the traditional sequential grid (SG) and a new radial strip (RS) approach. The performance of the multi-class SVM based classifier was unfortunately not what was expected. However, the findings show that both the SG and RS lacunarity produce enough discriminating features that warrant further research.

Comments

Originally published in the Proceedings of the 12th Australian Digital Forensics Conference. Held on the 1-3 December, 2014 at Edith Cowan University, Joondalup Campus, Perth, Western Australia.

Share

 
COinS