Text And Language Independent Speaker Recognition System for Robust Environment
Pasumarthi Soumya Sri, P. Seetha, S. Emalda Roslin, N. M. Nandhitha · 2023
Forensic speaker recognition is required when an individual voice is presented as evidence, such as in telephone recordings, wiretaps, audio surveillance, or informant recordings. Speech is the primary means of communication in our daily lives, with numerous languages used for this purpose. Speech technologies are widely utilized and have countless applications, allowing machines to respond accurately and dependably to human voices and provide helpful and valuable services. The goal of this project is to create a model that can identify speakers regardless of the language they speak and the volume of their voice, and can even recognize speakers when there is noise present in the speech. In our proposed work, Discrete wavelet transform (DWT) is used for feature extraction and Back Propagation Network (BPN) and Adaptive Resonance Theory (ART) is used for classification. This speaker recognition system can be applied in various fields, particularly in forensics, security, and biometrics where speech can be used as a password or for other purposes.