Strengthening steganoghraphy by using crow search algorithm of fingerprint image

Omar Younis Abdulhammed

doi:10.15587/1729-4061.2020.200282

Abstract

In image steganography, secret communication is implemented to hide secret information into the cover image (used as the carrier to embed secret information) and generate a stego-image (generated image carrying hidden secret information).Nature provides many ideas for computer scientists. One of these ideas is the orderly way in which the organisms work in nature when they are in groups. If the group itself is treated as an individual (the swarm), the swarm is more intelligent than any individual in the group. Crow Search Algorithm (CSA) is a meta-heuristic optimizer where individuals emulate the intelligent behavior in a group of crows. It is based on simulating the intelligent behavior of crow flocks and attempts to imitate the social intelligence of a crow flock in their food gathering process. This paper presents a novel meta-heuristic approach based on the Crow Search Algorithm (CSA), where at the beginning the color cover image is converted into three channels (RGB) and then those channels are converted into three spaces, which are Y, Cb, Cr. After applying Discrete wavelet transform (DWT) on each space separately, the CSA algorithm is used on each space (YCbCr) to find the best location that will be used to hide secret information, the CSA is used to increase the security force by finding the best locations that have high frequency and are invulnerable to attacks, the DWT is used to increase robustness against noise. The proposed system is implemented on three fingerprint cover images for experiments, for the quality of stego image the histogram, Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), Number of Pixel Change Rate Test (NPCR), Structural Similarity Index Metric (SSIM) and Correlation Coefficients (CC) are computed. The result demonstrated the strength of the CSA to hide data, also discovered that using CSA may lead to finding favorable results compared to the other algorithms

Highlights

In the last decade, with the advance of network technology, a large amount of valuable information has been transferred over the Internet
The aim of this work is to increase and improve the security of steganography and high capacity of concealment by using Crow algorithm and Discrete wavelet transformation (DWT), where CSA is used as a powerful swarm intelligence search technique for finding random positions that have high frequency in the cover image to hide a secret message, evaluate the behavior of Crow algorithm classically, quantitatively and humanly using image steganography, maintain image quality (Perceptual Transparency), difficult to alter secret information once it has been embedded into the stego image and reduce the time spent for embedding and extraction of secret data
The same case applies to the cover image (b) with stego image (b) and cover image (c) with stego image (c)

Summary

Introduction

With the advance of network technology, a large amount of valuable information has been transferred over the Internet. Encryption makes the message unreadable, but making message suspicious enough to attract eavesdropper’s attention. For overcoming this problem, we must utilize data hiding techniques which hide secret data behind a cover media. Steganography is a data hiding technique which embeds secret data into a cover media such as text, image, audio and video, so that the result is not noticed by any third’s attention. Application of the data hiding can be used in military, commercial and anti-criminal depended application, transmission of confidential documents between international governments and be anonymous in internet [2]. The algorithms that simulate natural operations are called swarm intelligence; they are considered as part of artificial intelligence [3]. There are many examples of swarm techniques or algorithms in real life such as fish schools, colonies of ants and termites, animal herds, bird flocks, crow and fireflies [4]

Objectives

Results

Conclusion