Improving quality of stego images through dithering techniques for pixel pair matching steganographic schemes

Abstract

Steganography plays a vital role in sending any covert data via public media without hiding the covert channel. Due to visually imperceptible embedding process of secret, it is able to escape visual attacks. But when it is suspected, steganography can be detected by any efficient steganalysis techniques. It is an important aspect to improve the security of stego image if the information hidden is highly confidential, by improving the stego image quality. This paper presents a comprehensive analysis of deploying dithering (diffusion) techniques for enhancing the quality of stego images created by various Pixel Pair matching (PPM) steganographic algorithms. The dithering process attempts to keep the difference between original cover image and secret embedded stego image to a very minimum value. The quality of stego images is evaluated on parameters such as Peak Signal to Noise Ratio (PSNR), Structural SIMilarity index (SSIM), Net Pixel Change Rate (NPCR) and Unified Averaged Changed Intensity (UACI). Experimental results show that, depending on the type of steganographic algorithm used, the behaviour of dithering techniques differ and the choice of dithering must be made considering computational complexity of dithering process also.