Abstract:n this paper, a novel image steganography approach is proposed to enhance the visual quality of stego image. The cover image is decomposed using Discrete Wavelet Transform (DWT) to produce wavelet subbands and threshold value is calculated for each higher frequency wavelet subbands. Wavelet coefficients having magnitude larger than the threshold of its subband are selected to embed the secret data. Semi Hexadecimal Code (SHC) is proposed to convert pixel value of secret image into smaller equivalent value so that it distorts stego image as less as possible. Experimental results shows that maximum PSNR between cover image and stego image is more than 75 dB .Proposed approach is also compared with the existing approaches and this comparison shows that the proposed approach is better than the existing approaches.Keywords: DWT, PSNR, EF, SHC, Threshold.
I. INTRODUCTIONSteganography is a process of hiding a secret message in an ordinary message and the extraction of secret message, whenever required. Image steganography techniques are divided into two categories: spatial domain and frequency domain steganography. In spatial domain techniques, secret message is embedded in the image pixels, while in frequency domain techniques, image is first transformed and then secret message is embedded in the transformed image. Frequency domain techniques hide messages in significant areas of cover image which make them more robust than spatial domain techniques. Several approaches have been proposed for frequency domain steganography using different transforms [1][2].Market et al.[3] proposed a watermarking technique in which the secret image is decomposed into wavelet subbands using DWT. The watermark is embedded into largest coefficients of high and middle frequency wavelet subbands of an image. Maximum PSNR proposed by their technique is 32.7 dB which is 0.14 dB more than Hsu et al. method [4].Toaba et al.[5] proposed a technique in which secret message is embedded into least significant bits of wavelet coefficient of an image. Some pre-processing steps are applied on cover image to adjust saturated pixel components in order to recover the embedding message without losing cover image data. Cover image is adjusted before applying Integer Wavelet Transform followed by two levels DWT. Permutation of stego key is used to embed the bit stream of secret image after inverse IWT after that stego image is collected. Extraction process is just reverse of embedding process. Maximum PSNR offered by their technique is 73.91 dB.Chen et al. [6] proposed a data hiding technique to hide data in high frequency wavelet sub-bands. The proposed technique is divided into two modes and three cases. The modes are fixed and varying. The cases are low embedding capacity, medium capacity and high capacity. Sequence mapping tables are used in raster scan manner to embed the data. For fix mode, 46.83 dB is the highest PSNR value and 39.00 dB is lowest. For varying mode, highest PSNR is 50.85 dB and minimum PSNR is 44.76 dB.Ataby et al. [7] propo...