A privacy protection scheme for biological characteristics based on 4D hyperchaos and matrix transformation

In recent years, biometrics have been widely used in areas such as access control, healthcare, finance and the Internet of Things (IoT). However, due to the uniqueness and immutability of biometric data, it poses a serious privacy risk once leaked. To address these challenges, this paper proposes an improved biometric image encryption scheme. We enhance the classical three-dimensional Chen’s chaotic system into a four-dimensional model to take full advantage of its high sensitivity and stochasticity. By integrating Latin matrices and semi-tensor products, we develop a novel encryption algorithm designed to protect multimodal biometrics. The method overcomes the instability of traditional cryptographic algorithms and ensures robust protection of biometric data when processing different images such as face, fingerprint, palmprint and iris. Various performance evaluations are also conducted, in which the image encryption time reaches 0.071s, the UACI values of the ciphertext images are close to 99.6094%, and the information entropy of the ciphertext images reaches 7.9980. The experimental results show that the algorithm has excellent encryption, security, and efficiency. This method provides a reliable solution for securing biometric data in an increasingly complex digital environment.