HIDING IMAGE IN IMAGE USING ITERATED FUNCTION SYSTEM (IFS)

摘要

In this paper, an image high hiding steganographic method for concealing digital image in a cover image is introduced, and implemented. The proposed method exhibits the following advantages over the existing methods, first, it shows high hiding rate such that the embedded secret image can be larger than the cover image. Second, the quality of stego-image is little degraded. Third, a small difference exists between the extracted secret image and the original secret. In embedding phase, the cover image is partitioned into overlapped blocks, but the secret image is partitioned into non-overlapped blocks, and converted from RGB space to YUV color space, then the chromatic bands U and V space are downsampled. Then, the linear affine mapping is applied to represent the blocks of bands of secret image in terms of cover image blocks using IFS-code. To speed up the mapping-search process a moment based classifier is used to classify the secret and cover blocks to reduced the encoding time without causing a reduction in image fidelity, where the encoding time of whole hiding process becomes within [6.94-9.67] second for a 24-bitmap color secret image its size is 256x256 pixels, the size of the color secret image (24-bitmap) is 1.48 times the size of cover image. The technique used to embed the generated IFS-code, is the Least-Significant-Bit (LSB) substitution method. The embedded secret image can be extracted from the resulting stego-image without need to referencing to the original secret image. The experimental results demonstrated that the values of the PSNR fidelity measure ranging between (31.04-31.41) decibels when the ration of secret image size to that of cover image is (100% and 123%), respectively, and also the difference between the stego-image and the cover image is almost imperceptible.