An approach to reversible information hiding for images

Information hiding (also known as data hiding), is a process of embedding secret message into a cover media for the purpose of covert communication, identification, security and copyright. Information hiding is used to conceal the content of messages. This is achieved by hiding secret message into other digital media like audio, images, video etc. For hiding secret message information in images, there exists a large variety of techniques, of which most techniques cause distortions in the original image even after secret message recovery, either due to bit-replacement or quantization error or truncation. It is especially important for medical and military applications that the image obtained after data extraction should be distortion free. Reversible data-hiding is the technique which embeds data in to a digital media such that the original media can be recovered without any distortion after the hidden message has been extracted. In this paper we propose a lossless method which embeds and extracts the data in the spatial domain. This method uses only one statistic parameter which controls the embedding and extraction of data. Two methods with different block orientations have been proposed. The novel method allows for an increase of 41.57% in average embedding efficiency when compared to existing methods, while maintaining the cover image degradation (PSNR - peak signal to noise ratio) at a comparable level. We also suggest a multi-layered embedding for increasing embedding capacity further.