An Image Authentication Technique in Frequency Domain using Genetic Algorithm (IAFDGA)

J.K Mandal

doi:10.5121/ijsea.2012.3504

Abstract

In this paper Genetic Algorithm based image authentication technique in frequency domain using Z transformation (IAFDGA) has been proposed.A 2×2 mask is taken from the source image in row major order. Z transformation is applied to transform it into freque ncy domain. Six bits are embedded in each sub mask into the second and third transformed coefficients. The sub mask is then transformed into spatial domain using inverse Z transform. Embedded image mask of size 32 bits are taken as initial population. New Generation followed by Crossover and Mutation are applied on it.To obtain New Generation , minimum coefficient of the mask is chosen, if the minimum coefficient is negative then subtract the minimum coefficient from each coefficient of the mask so that ex traction of hidden bits are ensured. For the Crossover operation consecutive bit -wise XOR is performed on the rightmost three bits of each byte in three steps. It will form a triangular form and the first bit of each step is taken as the output. Right most two LSBs of two consecutive pixels are swapped with each other as a part of Mutation operation. Genetic algorithm is applied to enhance a layer of security level. At the time of embedding dimension of the authenticating image followed by the content are e mbedded. Reverse process is followed at the time of extraction. High PSNR obtained for various images compared to existing Chin -Chen Chang et al.[2] conform the quality of invisible watermark IAFDGA. Large capacity as compared to existing algorithm [1] ensured the high payload of the scheme.

Highlights

Data security is the most important thing days
A 2×2 sub mask is taken in row major order from the host Gray scale image in recursive manner and Z transformation is applied on it to transform it into frequency domain
Analysis of result has been made on various images taken from image database [16] using IAFDGA technique in terms of visual interpretation, peak signal to noise ratio (PSNR), MSE and Image Fidelity (IF)

Summary

INTRODUCTION

Data security is the most important thing days. Data security can be achieved by hiding data into various natural cover data like sound, images, logos etc. Embedded data is referred to as stego-data and it must be embedded in such a way that the fidelity of the cover image is kept intact [4], [8], [13], [14], [15]. Information hiding [4] is a way to authenticate image. Image authentication and information security is very important thing to protect digital document from unauthorized access [3]. An example of steganography is the prisoner communicating with the outside world under the supervision of prisoner warden. To ensure another layer of security and make the extraction impossible for the intruder the Genetic Algorithm is used. In this paper an image authentication and data hiding technique has been proposed.

THE TECHNIQUE

Insertion Algorithm

Extraction Algorithm

Example

RESULTS

CONCLUSIONS