Dynamically generate a long-lived private key based on password keystroke features and neural network

It is well-known that the protection of long-lived private keys in cryptographic schemes is one of the most important issues for information security. Any cryptographic scheme that reveals private keys will soon have its security absolutely disintegrate. For example, in digital signature systems, anyone who obtains the victim's private key, authenticity and non-repudiation can no longer be claimed or proven. Because the private key is a long random bit string and should be stored securely, some special cryptographic hardware such as an IC (Integrated Circuit) card is needed to store and protect the private key. Unfortunately, the security of private keys solely depends on the vulnerable passwords. This study proposes combining a neural network technique and password keystroke features to dynamically generate a long-lived private key rather than statically stored in a storage unit. Compared with other traditional methods, even if the storage unit is lost or the password is revealed, the probability of exposing the private key is reduced.