Comparative study on different reversible image data hiding techniques

Abstract

Reversible steganography, also called reversible data hiding in digital images has been studied extensively in recent years. Reversible Image Data Hiding (RIDH) is a category of data hiding technique that ensures perfect reconstruction of cover image upon the extraction of the embedded message. The property of reversibility means that the original image can be recovered completely after the embedded bits are extracted. The main focus is to apply reversible data hiding algorithms on encrypted images by wishing to remove the embedded data before the image decryption. This paper compare and analyses the different methods which has been used for reversible image data hiding over several years. Also, proposes a novel Reversible Image Data Hiding (RIDH) scheme over encrypted domain. A public key modulation mechanism is applied to achieve data embedding, in which access to the secret encryption key is not needed. A powerful two-class SVM classifier is designed at the decoder side to distinguish encrypted and non-encrypted image blocks, which allows to decode the embedded message and the original image signal jointly. Compared with the state-of-the-arts, the proposed technique provides higher embedding capacity, and is able to reconstruct the original image as well as the embedded message perfectly.