Article ID Journal Published Year Pages File Type
528506 Journal of Visual Communication and Image Representation 2016 12 Pages PDF
Abstract

•Reduce the information for recording block indexes by non-uniform clustering algorithm and thus improve the quality of camouflage image by setting small block size.•Modify the previous color transformation to be reversible. So the secret image can be transformed to a freely-selected target image, getting a camouflage image used as the camouflage of secret image with good visual quality, and the secret image can be restored without any loss.•Because of the reversibility, we can further realize multi-round transformation, that is, transform the camouflage image to another target image.

A new reversible image transformation technique is proposed, which not only improves the visual quality of the camouflage image created by transforming a secret image to a freely-selected target image, but also can restore the secret image without any loss. Effective clustering algorithm is utilized to reduce the information for recording block indexes that is vital for restoring the secret image. Therefore, the transformation can be made between the blocks with relatively small size, thus greatly improving the visual quality of the camouflage image. The root mean square error of camouflage image was reduced by 6 in most cases compared with the previous method. Since the proposed technique is reversible, we can further realize two-round transformation by transforming the camouflage image to another target image and thus hide two images into only one.

Related Topics
Physical Sciences and Engineering Computer Science Computer Vision and Pattern Recognition
Authors
, , ,