Answers are grounded in this paper and cite their sources.
The integration of complementary anatomical and functional information from various imaging modalities, such as CT, MRI, and PET, through multimodal medical image fusion has emerged as a potent technique for enhancing brain tumour diagnosis. Nevertheless, sharing and storage of fused medical images may pose serious privacy and security risks. Additionally, traditional methods for permutation diffusion encryption generally encounter sequential processing bottlenecks, making them unsuitable for real-time medical scenarios. Here we introduce a parallel cryptographic permutation network (PCPN) that blends DNA-assisted chaotic permutation and light diffusion to ensure secure and computationally efficient protection of fused medical images. At first, the fused image is divided into eight binary bit-planes. Next, a chaotic sequence generated by the logistic map is converted into DNA symbols to form distinct permutation tables for each bit-plane. The GPU-parallelised permutation allows the simultaneous scrambling of the bit planes. After that, the scrambled bit-planes are combined to form the encrypted image. Pixel-wise XOR diffusion is done by a chaotic keystream generated by a second logistic map. It not only offers avalanche characteristics and statistical randomness but also needs very low computation. By allowing independent bit-plane processing and precomputed permutation tables, the proposed system removes sequential feedback dependencies, which greatly diminishes encryption times. Experiments have indicated that our encryption scheme outperforms other schemes in cryptographic security. It accomplishes an NPCR of 99.92%, UACI of 33.48%, and entropy of 7.9992 bits/pixel. The scheme is highly resistant to differential and statistical attacks while preserving the integrity of the image. Our scheme strikes an ideal balance between protection and computational efficiency, which is a practical solution to ensure the confidentiality of multimodal medical image fusion for brain tumour diagnosis. And this method is also a very viable option for the implementation of real-time intelligent healthcare systems.
No takes yet. Share an insight, caveat, or question.
Ramasamy et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: