Abstract
A three-dimensional (3D) watermark integral imaging-based watermarking algorithm operating in the cellular automata (CA) domain is presented. Instead of using conventional two-dimensional (2D) watermarking, in this proposed method, an elemental image array (EIA) obtained by recording the rays emanating from a 3D watermark through a pinhole array is used as a watermark to be embedded into the coefficients of CA-transformed pattern. A series of depth-dependent 3D plane images can be reconstructed by the volumetric computational reconstruction (VCR) technique. The embedded 2D EIA is composed of many elemental images possessing their own perspectives of the 3D watermark. The hologram-like 3D properties of the 2D EIA can make a robust reconstruction of the watermark available even though the embedded watermark data may be damaged by image-processing attacks. To embed the watermark imperceptibly, a masking model is introduced to adapt the watermark embedding strength by exploiting the characteristics of the human vision system. In addition, a CA transform with various CA bases and rule numbers can provide many transform planes for watermark embedding. To demonstrate the usefulness of the proposed method, we carry out preliminary experiments and present the experimental results.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have