Increased cyber threats require more secure methods for generating cryptographic keys. Conventional methods frequently use fixed keys that can be easily compromised, leading to the need to investigate other options. This study explores utilizing biometric data, specifically fingerprints, to create cryptographic keys that are both unique and secure against unauthorized access. This research seeks to improve the security of cryptographic systems by utilizing the natural uniqueness of biometric traits and implementing cancelable biometrics to address privacy issues. The research examines current literature on biometric cryptographic key generation, cancelable biometrics, and the related hurdles, suggesting a new method that merges cutting-edge feature extraction methods with dynamic key generation for a secure and user-friendly cryptographic system. The proposed system ensured robust security against brute-force attacks by effectively generating 2048-bit RSA keys with a high entropy (near 1). The average processing time per image was 71–87 s, which indicates that the method is feasible for real-world applications. The prime numbers generated were sufficiently distant and exhibited significant dissimilarity, reducing the risk of factorization.