Randomized Multimodal Authentication System for Secure File Storage

With the world becoming increasingly dependent on cyberspace, it has become alarming to experience an increase in threats, which requires effective security measures for protecting confidential information and user credentials in an organization. This study focuses on the development of a Multimodal Authentication system for securing files and folders through the incorporation of multiple modes of security, such as face recognition, image grid selection, and voice-based authentication. Deep learning-based face recognition is used here for facial Authentication along with the Histogram of Oriented Gradients used to detect the landmarks and extract the features. Algorithms such as Mel Frequency Cepstral Coefficients along with Dynamic Time Warping are used for voice authentication and Graphical password authentication for image grid selection. The proposed system uses role-based access control. More to that, it uses a file management component, which is utilized in allowing the data storage and retrieval systems to be secure. Advanced Encryption Standards have been utilized with RSA encryption on key exchange and also with hashing algorithms, such as SHA-256. By incorporating these techniques, this system provides the development of the secured file system.