Superpixel based robust reversible data hiding scheme exploiting Arnold transform with DCT and CA

The emergence of data hiding schemes was to provide multimedia security besides the sharing of information in a confidential manner. Since then, various techniques have cropped up to meet the objective of secured data communication using multimedia documents. In this paper, a novel data hiding approach has been designed to exploit the superpixel of the image to identify the nature of the blocks for embedding data using spatial and transform domain techniques such as Cellular Automata (CA) and Discrete Cosine Transform (DCT) respectively. In addition, an Arnold transform is used to provide a blanket of security through a random selection of blocks via sharable keys. The proposed technique obtains YCbCr color model of an RGB image to use superpixel on only the Y color component. Thereafter, the acquired labeled image of superpixel is partitioned into ( 8 × 8 ) blocks which are to be categorised as homogeneous and heterogeneous blocks. Based on the categorization, DCT and CA are applied for the embedding of the secret data in the Cb and Cr color components. The experimental results and its comparison with state-of-the-art schemes uphold the effectiveness of the demonstrated scheme. Various analysis have been conducted to showcase the degree of robustness, imperceptibility, and visual quality attained by the proposed scheme.