Reversible visual transformation via exploring the correlations within color images

Reversible visual transformation reversibly transforms a secret image to a freely-selected target image and gets a camouflage image similar to the target image, which has been proved to be very useful in such two applications: privacy protection of images and reversible data hiding in encrypted images. Now, a new reversible transformation technique for color images is proposed by exploring and utilizing the correlations among three color channels and inside each color channel. The amount of the accessorial information for recording transformation parameters is largely reduced. Therefore, the visual quality of the created camouflage image is much improved by dividing the secret image and the target image into even smaller blocks for transformation.