Image reversible visual transformation based on MSB replacement and histogram bin mapping

Abstract

An image reversible transformation method is proposed for secret image transmission and privacy protection. With the proposed method, the visual content of a secret image is changed to mimic a target image while the secret image can be recovered from the transformed image. Different from the previous methods that divide the secret image into blocks, our method processes the secret image as a whole so that no block artifact will be caused. Firstly, the most significant bit (MSB) plane of the secret image is replaced and then a process called histogram bin mapping is conducted. The replaced MSB plane and histogram bin mapping information are further hidden into the transformed image. Via reversible data hiding, the secret image can be exactly recovered by extracting the hidden data and reversing the transformation process. The experimental results show that the visual content of secret images can be semantically changed, while almost all of the test images were successfully recovered from the corresponding transformed images.