Double‐image asymmetric cryptosystem using cylindrical diffraction and spectrum fusion and compression

Abstract

A double-image asymmetric cryptosystem is proposed by using cylindrical diffraction and spectrum fusion and compression in the zigzag-scanned domain of discrete cosine transform (DCT). First, each plaintext of double image is converted into spectrum coefficients (SCs) by DCT, and the SCs are zigzag scanned. Second, the first 1/4 of SCs is resized to two low-frequency images, the second 1/4 of SCs is resized to two high-frequency images, while the last 1/2 of SCs is discarded. Third, the two low-frequency and two high-frequency images are fused and merged into three rounds of encryption process by using the complex combination, random phase encoding, and phase reserving-truncation in the cylindrical diffraction transform domain. Finally, the plaintexts of the double image are successfully encrypted into a ciphertext and three private keys with each only 1/8 and totally 1/2 of the size of the plaintext, which gains a high compression efficiency. In the proposed cryptosystem, the quality of the reconstructed image is improved due to the compression in the zigzag-scanned DCT domain, and the security is improved owing to the asymmetric cylindrical diffraction and spectrum-fusion algorithm. The effectiveness and flexibility of the proposed cryptosystem have been validated by the numerical simulation results. Proposed asymmetric cryptosystem PRT in CDT domainAccording to the cylindrical diffraction theory in [32,33], the object and the observation surfaces are concentric cylindrical surfaces. Figs. 1a and b show the two, inside-out propagation (IOP)