speak

Speaker Recognition System

This is a text-independent speaker recognition system built using deep learning. The model and test code is in Python. It is intended to be used as an authentication technique in IoT systems. The IoT niche kept in mind while building this solution was a home automation system that does not require near perfect accuracy but instead requires extremely simple, robust, and hassle-free authentication that enables people with disabilities like paralysis and dementia to use it without difficulty. It is designed to register a new user profile within 7 seconds and authenticate within 3 seconds.

I have included screenshots of a command line interface to demonstrate its functionality, although it is not supposed to be used in this manner.