Reversible data hiding using bit flipping and histogram shifting

Medical, military, and cloud computing applications require high degrees of security and privacy, and thus they are potential candidates to benefit from hiding data in encrypted images. This paper introduces a new algorithm that combines cryptography and data hiding techniques to hide data into encrypted grayscale images in a reversible manner. Based on the proposed algorithm, the original cover image is encrypted before a data hiding procedure is applied to embed two watermarks in the encrypted image. One watermark is embedded by flipping the two middle bits of encrypted image pixels which are secretly selected using a pre-shared hiding key. The bits of the second watermark are embedded by applying the histogram shifting method on the watermarked image. At the receiving send, the bits of the second watermark are extracted using the inverse histogram shifting and the integrity and authenticity of the recovered image are verified. The experimental results obtained demonstrate that the proposed algorithm meets the requirements of effective reversible data hiding such as high visual image quality, high embedding capacity, and high entropy. The algorithm is also competitive with other recent algorithms with respect to these performance metrics.