Feature Based Encryption Technique for Securing Digital Image Data Based on FCA-Image Attributes and Visual Cryptography

Abstract : Lossless pixel value encrypted images still maintains the some properties of their respective original plain images. Most of these cryptographic approaches consist of visual cryptographic techniques and pixel displacement approaches. These methods of cryptography are useful in cases as medical image security where pixel expansion is avoided in both the encryption and decryption processes. In this paper we propose a hybrid cryptographic encryption approach by using features generated from digital images based on Galois lattice theory and a visual cryptographic technique based on RGB pixel displacement. The features extracted from a plain image and a lattice was generated which was then used to generate a key used to encrypt the plain image. At the end of the process, there was no pixel expansion and the arithmetic mean, the entropy as well as the Galois lattice of both ciphered and plain image remained the same. The features extracted from the plain image were the same as that of the ciphered image irrespective of pixel displacement that occurred, this makes our approach a suit-able basis for image encryption and storage as well as encrypted image indexing and searching based on pixel values. The implementation was done using Galicia, Lattice Miner and MATLAB..