Adaptive Pixel Swapping Based Steganography Reducing Embedding Noise

A Block Permutational Steganographic Algorithm for Scanned Documents and other Images

“…We compare our own algorithm with a prior permutational algorithm, of which the best performing is the pixel swapbased algorithm of [35]. We implement the algorithm of [35] in order to directly compare performance using the same sample images. Certain parameters of [35] were open to specification.…”

“…Table 1 lists the average computation times, both of our algorithm and of [35], for the MSRCv2 [39] and NoisyOffice [36][37][38] datasets. Tests were run on an Intel 7th Generation Core i7 notebook computer.…”

“…For each image-with or without an encoded message-the Wavelet Motion Analyzer [34] is used to create a feature. As in [35], a Fischer linear classifier is used to discriminate between features resulting from images containing, or not containing, a message. 10-fold cross-validation is used to train and evaluate said classifier.…”

“…Their algorithm resists the Sample Pair Analysis [25], which estimates the embedding rate of a hypothetical message in any given image, but is detected by the Wavelet Motion Analyzer [34]. The approach of [33] is largely superseded by [35], which describes a variant block-based pixel-swapping algorithm based partially on comparisons with the block's mean value. Said algorithm was implemented, and evaluated based on detection by the Wavelet Motion Analyzer, for the purpose of comparison.…”

“…In addition to the Wavelet Motion Analyzer, we also use additional methods to compare the algorithms: the Sample Pair Analysis [25] and Peak Signal Noise Ratio (PSNR) [3]. For both the NoisyOffice dataset [36][37][38] of scanned documents, and the MSRCv2 dataset [39] of more general images, Wavelet Motion Analyzer results indicate that the algorithm of [35] is outperformed by our own, while the other measures are either roughly comparable or improved.…”