Open AccessImage Steganography by Modified Simple Linear Iterative Clustering
Author(s) -
Ismail Kich,
El Bachir Ameur,
Youssef Taouil
Publication year - 2020
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.c8903.029420
Subject(s) - computer science , steganography , pixel , artificial intelligence , cover (algebra) , cluster analysis , enhanced data rates for gsm evolution , grayscale , computer vision , pattern recognition (psychology) , human visual system model , segmentation , image (mathematics) , digital image , image processing , mechanical engineering , engineering
Steganography is an information security technique that consists of concealing secret data into digital medias including videos, texts, network protocols and images. In this paper, a steganography method to dissimulate the secret information in gray-scale images is proposed; the dissimulation is adapted to the cover image’s texture, data is hidden in the edge areas. The edge pixels are selected by over-segmentation using Modified Simple Linear Iterative Clustering (M-SLIC). This algorithm allows to decompose the cover image into K regions which we call superpixels. The image’s texture and the amount of the secret data are the factors that help to determine the value of the parameter K. Choosing the pixels of complex regions to conceal secret information is due to the fact that the human visual system is designed to notice changes in the pixels of smooth areas. Therefore, edge areas tolerate larger changes than smooth areas without causing detectable distortions. Experiment on a large set of images were carried out; results illustrate the good performance of the proposed work in terms of capacity, security and imperceptibility in comparison to recent works.