Article Preview
TopIntroduction
Steganography is the art of concealing secret information in a multimedia object while hiding its existence; the secret data is then extracted at its destination. It strives for high security and capacity. Various steganography techniques on images have been studied upon in (Kadhim, Premaratne, Vial, & Halloran, 2019);(Ghosal & Mandal, 2019) and used in (Huang, Echizen, & Nishimura, 2010);(Wu, Hsu, & Lee, 2009). Reversible steganography is one of the techniques of concealing secret data in a multimedia object in a manner in which it can be completely recovered after the classified data is extracted (Zhicheng Ni, Yun-Qing Shi, Ansari, & Wei Su, 2006). This important technique has significance in several fields such as military imagery, medical imagery and law forensics where the original multimedia object is so important that its exposure can risk the major security issues and even the slightest damage to it is unaffordable. Reversible data hiding (RDH) can be applied to ensure the security of data over the cloud as well. The data owner encrypts the data and hides it in an encrypted cover multimedia object before sending it to the cloud. To avoid any kind of permanent distortion by several users, RDH technique can be used so that the assigned user possessing the valid keys can obtain the data in its original form and perform computations on it.
The use of reversible steganography can be combined with encryption as an extra step for protecting data and confidentiality. There are various techniques which are discovered from time to time to encrypt the images to make it more secure. (Saha & Geetha, 2017);(Saha, Geetha, Kumar, & Kim, 2018);(S Geetha, Punithavathi, A Magnus Infanteena, & Sindhu, 2018);(Gupta & Biswas, 2017) describe numerous image encryption techniques. The characteristics of the chaotic maps such as ergodicity, mixing property and sensitivity to initial condition/system parameters have attracted the attention of many cryptographers to develop new encryption algorithms (Sankpal & Vijaya, 2014);(Fu, Lin, Miao, Liu, & Chen, 2011);(Wang, Liu, & Zhang, 2015);(Patidar, Pareek, Purohit, & Sud, 2011). These properties can be considered analogous to some cryptographic properties of ideal ciphers such as confusion, diffusion, balance and avalanche property etc. It also provides a good combination of speed, security, complexity, and computational power. The paper uses encryption based on chaotic maps to encrypt the image in its first phase. Encryption using chaotic maps allows the user to maintain the spatial correlation among the pixels of the image The initial parameters that are used to perform chaotic mapping serve the purpose of encryption key. Due to the weakness of one dimensional chaotic maps of small key space, the image is doubly encrypted post embedding using homomorphism based encryption providing more resistance to attacks.
Apart from reversibility, separability and decrypted marked image fidelity are two other evaluation criterions that are used in RDH in encrypted images. This paper proposes a novel, improved reversible steganography scheme in encrypted grayscale images that fulfils both the criteria. The scheme is separable i.e. data extraction and image recovery can be done separably. The embedded secret data can be extracted from the marked encrypted image only if the receiver possesses both, the data hiding key (used to encrypt the secret data) and the encryption key used in post embedding encryption phase. The scheme fulfils decrypted marked image fidelity i.e. the encryption keys of both the phases are sufficient to obtain the decrypted marked image similar to the original one. Decrypted marked image fidelity is usually evaluated by comparing the Peak Signal to Noise Ratio (PSNR) values of the decrypted marked image with the original one. Apart from working on all these aspects, the proposed scheme successfully overcomes the drawback of previous work done in (Xiao, Xiang, Zheng, & Wang, 2017).