Biometric-key generation is a process to convert a piece of live biometric data into a key using an auxiliary information, which is also called a biometric helper. A biometric-key can be repetitively generated and the biometric need not be physically stored. In this paper, we introduce a novel biometric-key generation scheme which is based on a randomized biometric helper. This scheme consists of a randomized feature discretization process and a code redundancy construction. The former enables one to control the intra-class variations of biometric data to the minimal level and the latter further reduces the errors. The randomized biometric helper ensures that a biometric-key is easy to be revoked when the key is compromised. The proposed scheme is evaluated in the context of face data based on a subset of the FERET database.