| Many digital images involve information related to personal privacy,corporate privacy,and military secret.which may be stolen or tampered during the transmission process.Therefore,secure and efficient image encryption algorithm is one of the vital measures to guarantee image secure transmission.Inspired by existing image encryption algorithms,two chaotic color image encryption algorithms were proposed by combining real-valued discrete Gabor transform with bit permutation.The research contents of this dissertation are as follows.A chaotic color image encryption algorithm based on real-valued discrete Gabor transform and bit permutation was designed.The image information was converted into the transform domain for encryption with the real-valued discrete Gabor transform.With plaintext image and initial key as input,the initial value and the control parameters of a 3D chaos,the number of cyclic shifts and the number of iterations of Arnold transform were obtained by the SHA-256 algorithm.Subsequently,the Arnold transform was performed on the transformed image to disorder bit values,and the coefficients of Arnold transform were selected from the chaotic sequence.Then the scrambled matrix was scanned by Zigzag operation.3D chaotic sequence and index matrix were utilized to perform nonlinear scrambling operation and cyclic shift diffusion operation on the scanned matrix to obtain ciphertext image.Key space,correlation and histogram were also analyzed.It is illustrated that the proposed color image encryption scheme is feasible and secure.By combining real-valued discrete Gabor transform with quaternion discrete Fourier transform,a hybrid chaotic color image encryption algorithm was investigated.A 3D chaotic sequence was generated with the initial value and the control parameters of the 3D Logistic map.The initial value and the control parameters of the 2D coupled chaotic map were generated by SHA-256 algorithm with the 3D chaotic sequence and the original plaintext image.After performing the real-valued discrete Gabor transform on the original image,the pixel positions and the pixel values in the transformed image were changed by means of bit permutation.Subsequently,the scrambled matrices were reconstructed into the imaginary parts of the quaternion,so as to achieve unified encryption of the three components of the color image and enhance the encryption efficiency.Phase mask was performed on the quaternion matrix with the 3D chaotic sequence.After phase mask,the real and the imaginary parts of quaternion matrix were obtained by the quaternion discrete Fourier transform,and the real and the imaginary parts were reconstructed into three matrices.The ciphertext image was obtained by quickly scrambling and diffusing the reconstructed three matrices with 2D coupled chaotic sequence and index matrix.It demonstrates that the color image encryption algorithm based on real-value discrete Gabor transform and quaternion discrete Fourier transform is secure. |
