Implementation of Voice Recognition-Based Cryptosystem

As user privacy becomes increasingly important, various encryption and decryption strategies have been developed to authenticate legitimate users and create privacy-preserving systems. However, in addition to security, user convenience is a crucial key performance metric in the current industry. This study aimed to enhance both security and user convenience by employing learning-based voice analysis techniques. The proposed voice recognition system employs Mel-frequency cepstral coefficients (MFCC) and cosine similarity to identify user voices and execute encryption and decryption services. Experimental results demonstrate that the proposed system effectively enhances both security and user convenience.