Abstract

In the today’s stage of increasing digital interactions and the need for secure authentication methods, traditional password-based systems face numerous challenges, such as susceptibility to theft, hacking, and the inconvenience of remembering complex passwords. In retaliation to these challenges, the paper presents a novel approach to authentication called the Silent Password, which proposed lip synchronization analysis and Convolution Neural Networks (CNNs) for user verification. The Silent Password system harnesses the unique characteristics of an individual’s lip movements during speech, capitalizing on the fact that these movements are both distinct and difficult for impostors to replicate. Furthermore, the CNN model extracts spatial and temporal features from the lip movement data, which are then compared with the user’s enrollment data. If the lip synchronization patterns match within an acceptable threshold, access is granted. This study demonstrates the feasibility of silent password as an innovative, secure and user-friendly authentication method, paving the way for future advancements in biometric-based user verification systems.