Novel 3-D hyperchaotic map with hidden attractor and its application in meaningful image encryption

To enhance the performance limitations of existing image encryption schemes using compressive sensing (CS) technique in withstanding plaintext analysis-based security attack models, a new feasible solution is proposed in this paper. First, on the premise that the discrete wavelet packet transform (WPT) matrix which is incoherent with the measurement matrix constructed by the Xorshift algorithm is selected as the sparse basis matrix, an asymmetric compressive sensing model is designed by combining semi-tensor product operation and matrix transform. Then, the plaintext sparse coefficients are compacted and encrypted by this novel compression model and bidirectional hybrid diffusion algorithm. Next, the secret data without semantic features are hidden into the publicly available carrier image by the Haar transform (HT) embedding in the novel orthogonal YCbCr color space. Wherein, to relieve chaotic dynamics degradation phenomenon caused by the finite word-length (FWL) effect, a novel 3-D bounded non-linear map (3-BNM) with hyperchaotic properties is introduced to construct the highly unpredictable secret code streams with the assistance of the perceptual hash algorithm (PHA). Eventually, the performance analysis and comparative experiments demonstrate that the proposed encryption scheme is capable of passing multiple security tests and outperforming the recently proposed schemes.