As a powerful cryptographic primitive, indistinguishable obfuscation has been widely used to protect data privacy on the Internet of Medical Things (IoMT) systems. Basically, the cryptographic technique protects data privacy using a function to obfuscate medical applications to perform outputs computationally indistinguishable. The state-of-the-art obfuscation technique (GGH13) utilizes a variant of the multilinear map to enhance security. However, in such schemes, it can be observed that noise lies in each element of the matrix, which means the matrix is a full rank matrix with a probability of almost 1 and results that it is unable to establish the relationship between the matrix determinant and rank. In this paper, we propose an attack to break such obfuscator. Specifically, we use approximate eigenvalues to remove the influence of noise on the matrix eigenvalues and build a specific relationship between the determinant and matrix rank. Our analysis shows the structural weakness of the state-of-the-art indistinguishable obfuscation mechanism, and we further discuss the future direction to resolve such privacy issues for IoMT applications.

作为强大的加密原语，难以区分的混淆已广泛用于保护医疗物联网（IoMT）系统上的数据隐私。基本上，密码技术使用混淆医疗应用程序的功能来保护数据隐私，以执行计算上无法区分的输出。最新的混淆技术（GGH13）利用多线性映射的变体来增强安全性。但是，在这种方案中，可以观察到噪声位于矩阵的每个元素中，这意味着该矩阵是具有几乎为1的概率的满秩矩阵，并且导致无法建立矩阵行列式与矩阵之间的关系。秩。在本文中，我们提出了一种攻击来打破这种混淆器。特别，我们使用近似特征值消除噪声对矩阵特征值的影响，并在行列式和矩阵等级之间建立特定的关系。我们的分析显示了最先进的，难以区分的混淆机制的结构弱点，并且我们进一步讨论了解决IoMT应用程序中此类隐私问题的未来方向。