Published Paper Details:

In today's exponentially growing data-centric industry security is the most important factor. Considering that, authentication that is based on biometrics traits is more secure for many applications. The aim is to build a model based on voice authentication which would recognize the verbal commands of the respective user to perform tasks on desktop applications, for example - sending emails and WhatsApp messages, performing file operations, etc. just like present systems, but there would be an additional layer of security in which it will only recognize the specific voice and thus follow the commands of the authenticated user. As the existing systems have the drawback of not authenticating the user's voice and directly functioning on any input voice, so to add a security aspect and maintain confidentiality we have proposed this model which would authenticate the user's voice and also address the issues where the existing voice-based systems fail such as authenticating pre-recorded voice of authenticated user. The main focus would be on increasing the security level and accuracy of biometric authentication. The phases involved in this system are Voice authentication and Desktop voice assistance. The implementation will consist of feature extraction, model building, and model evaluation. Next part is the Voice Assistant on Desktop Applications. This includes listening to spoken words and identifying them. The system takes voice and converts spoken words into text. This text is then fed as a query or the result is obtained and the system generates a reply. It uses a speech recognition library for performing speech recognition. This project would be the first step leading toward an era of secured voice-based assistance where confidentiality is a major requirement.