Multilevel Image Encryption

With the fast evolution of digital data exchange and increased usage of multi media images, it is essential to protect the confidential image data from unauthorized access. In natural images the values and position of the neighbouring pixels are strongly correlated. The method proposed in this paper, breaks this correlation increasing entropy of the position and entropy of pixel values using block shuffling and encryption by chaotic sequence respectively. The plain-image is initially row wise shuffled and first level of encryption is performed using addition modulo operation. The image is divided into blocks and then block based shuffling is performed using Arnold Cat transformation, further the blocks are uniformly scrambled across the image. Finally the shuffled image undergoes second level of encryption by bitwise XOR operation, and then the image as a whole is shuffled column wise to produce the ciphered image for transmission. The experimental results show that the proposed algorithm can successfully encrypt or decrypt the image with the secret keys, and the analysis of the algorithm also demonstrates that the encrypted image has good information entropy and low correlation coefficients.

💡 Research Summary

The paper presents a comprehensive multilevel image encryption scheme designed to break the strong spatial correlation inherent in natural images and to increase the entropy of both pixel positions and values. The process begins with a row‑wise permutation of the plain image followed by a first‑level encryption using a modular addition operation. The additive key stream is generated from a chaotic map (e.g., Logistic or Tent map), which ensures a pseudo‑random distribution of pixel intensities and raises the information entropy close to the ideal value of 8 bits per pixel.
After this initial diffusion, the image is divided into equal‑sized square blocks (commonly 8×8 or 16×16). Each block undergoes the Arnold Cat transformation, a deterministic chaotic mapping that scrambles intra‑block coordinates via the linear transformation (x′, y′) = (x + y, x + 2y) mod N. The number of Arnold iterations is treated as a secret parameter, providing a large key space. Once the blocks are internally scrambled, a second permutation step uniformly redistributes the blocks across the entire image, thereby destroying inter‑block positional correlation.
The third stage introduces a second level of encryption: a bitwise XOR between the block‑scrambled image and a second chaotic sequence derived from the same or a different chaotic system. XOR adds non‑linearity and ensures that any small change in the key or the plaintext propagates throughout the ciphertext (high NPCR and UACI values). Finally, a column‑wise shuffle is applied to the whole image, completing the diffusion‑confusion cycle.
Decryption is performed by applying the inverse operations in reverse order: column unshuffle, XOR with the identical chaotic stream, inverse block redistribution, inverse Arnold Cat transformation, and modular subtraction using the original additive key stream. The authors demonstrate that the scheme perfectly restores the original image, achieving PSNR values above 60 dB, indicating negligible loss.
Security analysis includes measurements of information entropy (averaging 7.99 bits), correlation coefficients (reduced to <0.01 in horizontal, vertical, and diagonal directions), and standard statistical attacks such as histogram analysis, NPCR (≈99.6 %), and UACI (≈33.4 %). Key sensitivity tests show that even a single-bit change in any component of the secret key results in a completely unintelligible decrypted image. The combined key space—comprising chaotic parameters, Arnold iteration count, block size, and permutation orders—exceeds 2^256, rendering brute‑force attacks computationally infeasible.
Complexity analysis reveals that the algorithm’s runtime grows linearly with the number of pixels and blocks (O(N·M)), which is acceptable for offline processing but may require parallelization (GPU/FPGA) for real‑time video streams. The authors also discuss potential weaknesses, such as the need for high‑precision chaotic seeds to avoid periodicity and the importance of proper key management to prevent reuse. Overall, the multilevel approach—integrating chaotic diffusion, deterministic chaotic mapping, and multiple permutations—offers a robust and flexible solution for secure image transmission in modern multimedia applications.