Image encryption and authentication verification using fractional nonconventional joint transform correlator

Abstract

We propose an image encryption and authentication verification scheme based on fractional nonconventional joint transform correlator architecture. For image encryption, an input image bonded with a random phase mask (RPM) and a separate key phase mask (another RPM) are fractional Fourier transformed independently. The obtained spectra are multiplexed together to what is called as the encrypted fractional joint power spectrum (EFJPS). For decryption, the key phase mask is fractional transformed and multiplied with the EFJPS and its inverse FRT is calculated. To check the quality of the decrypted image, the mean square signal-to-noise ratio (SNR) and peak-signal-to-noise ratio (PSNR) have been calculated. For security authentication verification, the input image bonded with an RPM and a separate key phase mask (identical RPM) are fractional Fourier transformed with same orders independently. The EFJPS is now fractional transformed with optimum fractional order, which gives a sharp autocorrelation peak in the output, if identical RPMs are used during the encryption process. Parameters such as SNR and peak-to-sidelobe ratio have been calculated to check the effectiveness of the proposed verification scheme. Computer simulation results with a binary text image and phase-encoded text image support the proposed idea.