Double random phase encoding based optical encryption systems using some linear canonical transforms: Weaknesses and countermeasures

During the last two decades or so, a large number of optical information security systems have been proposed by researchers, by exploiting various inherent advantages of optics, with a view to gaining superiority over the existing digital security systems. Among them, double random phase encoding (DRPE) scheme is the oldest and most exhaustively explored optical scheme. However, symmetric nature and linearity prove to be bane of the DRPE scheme, and give an open invitation to unauthorized users to crack the system by mounting different type of attacks. Due to this fact, security of the DRPE scheme has been compromised, making the system vulnerable to attacks. A number of schemes have also been introduced based on the use of fractional Fourier-, Fresnel-, gyrator-, Hartley-, and other transforms. However, vulnerabilities are not limited only to the Fourier domain based systems. Fresnel- and fractional Fourier domain based systems have also been found weak against the chosen-, and the known-plaintext attacks. Resistance of many of the security enhanced DRPE schemes in some other linear canonical transform domains has also been found to be weak against these attacks or their modified forms. When linearity is claimed to be broken with introduction of an amplitude mask at the Fourier plane, the DPRE scheme is able to nullify the high level known-plaintext attack but susceptibility is found against a simple impulse function attack. Recently, some strategies have been adopted to improve the resistance of the DRPE based schemes against the impulse function attacks. Some aspects of the problem have been discussed in the present chapter.