A Novel Image Encryption Scheme Based on Orthogonal Vectors
N. Ahmed, Yasir Saleem, Hasanuddin Al-Habib, Shazia Afzal, Sayyad Khurshid · The Nucleus · 2015
Image is an important utility of daily life. Use of internet and transmission of digital media over insecure channel such as broadcasting and unicasting through satellite pose a threat to security. Therefore, a novel encryption scheme is proposed for color images with the ability to tolerate noise and JPEG compressed. The algorithm operates in two phases. First phase is a transposition cipher that transposes the position of each pixel. The second phase uses orthogonal vectors for further processing in frequency domain to produce the cipher image. The tri-color image is separated into three channels and at the end, fused to produce the RGB cipher image. Due to the use of orthogonal vectors, it develops tolerance towards compression and channel noise. Experimental test are performed on a reasonable dataset of images to prove cryptographic security. The recovered image is tested based on a performance metric of image quality. The cryptographic analysis and performance evaluation has shown tolerance to noise and compression with adequate cryptographic security.