Scholarlink

<p>This paper constructs a novel heterogeneous coupled neuron model that exhibits the coexistence of distinct chaotic attractors. By formulating the Hamilton energy function, the energy differences corresponding to different attractor structures are quantitatively demonstrated. The complex nonlinear dynamics of the model are further investigated using one-parameter and two-parameter bifurcation diagrams, Lyapunov exponent spectra, and spectral entropy complexity. In addition, the 0-1 test and NIST randomness test are employed to confirm its capability of generating robust chaotic sequences. Leveraging the chaotic characteristics of the model, a medical image compression and encryption scheme tailored for multi-image scenarios is proposed. Specifically, multiple medical images are first fused into a single image, which is then sparsified using discrete wavelet transform. 2D compressive sensing and quantization are subsequently applied to obtain a compressed representation. SHA-512 of the fused image is employed to initialize the model and generate random sequences. These sequences are then used to perform cross-plane scrambling, bit diffusion, index scrambling, and XOR diffusion operations on the compressed image, ensuring secure and efficient transmission. Extensive performance experiments validate the security and robustness of the proposed scheme, demonstrating strong resistance to various external attacks. The evaluated security performance metrics include a large key space of <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(2^{647}\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mn>2</mn> <mn>647</mn> </msup> </math></EquationSource> </InlineEquation>, near-zero pixel correlation, high information entropy of 7.9997, as well as NPCR and UACI values approaching the theoretical ideals.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel heterogeneous coupled neuron model and its application in medical image encryption

  • 期刊论文

  • 发表日期：

    2026年3月11日

查看全文
  • Shaocheng Qu,
  • Qianqian Shi,
  • Xinlei An

摘要

This paper constructs a novel heterogeneous coupled neuron model that exhibits the coexistence of distinct chaotic attractors. By formulating the Hamilton energy function, the energy differences corresponding to different attractor structures are quantitatively demonstrated. The complex nonlinear dynamics of the model are further investigated using one-parameter and two-parameter bifurcation diagrams, Lyapunov exponent spectra, and spectral entropy complexity. In addition, the 0-1 test and NIST randomness test are employed to confirm its capability of generating robust chaotic sequences. Leveraging the chaotic characteristics of the model, a medical image compression and encryption scheme tailored for multi-image scenarios is proposed. Specifically, multiple medical images are first fused into a single image, which is then sparsified using discrete wavelet transform. 2D compressive sensing and quantization are subsequently applied to obtain a compressed representation. SHA-512 of the fused image is employed to initialize the model and generate random sequences. These sequences are then used to perform cross-plane scrambling, bit diffusion, index scrambling, and XOR diffusion operations on the compressed image, ensuring secure and efficient transmission. Extensive performance experiments validate the security and robustness of the proposed scheme, demonstrating strong resistance to various external attacks. The evaluated security performance metrics include a large key space of \(2^{647}\) 2 647 , near-zero pixel correlation, high information entropy of 7.9997, as well as NPCR and UACI values approaching the theoretical ideals.