Perceptual image encryption via reversible histogram spreading

Perceptual encryption is commonly used in encoded images and videos to intentionally introduce visual quality degradation to the media data to protect the added value out of the content in a reversible way. We propose in this paper a perceptual image encryption algorithm using iterative operation of local pixel shuffling and reversible histogram spreading to an extended gray-level range, via which images' local details can be visually better encrypted in a finer scalability compared to ordinary shuffling schemes, while images' global content in terms of local blocks' mean brightness can be strictly preserved same as the original without decryption. Security in terms of local content leakage is also enhanced by smoothly spreading the original image's histogram to a flattened and less discriminative one.