Speaker Recognition for Hindi Language Using Deep Neural Network

The work presents speaker identification using deep learning. Speaker identification is the process of identifying a speaker from the spoken utterances. The speaker recognition technology can be applied for authentication, surveillance, forensic speaker recognition, security and speaker tracking. The Text-to-speech (TTS) systems, also known as voice synthesis systems, can be implemented with artificial neural networks (ANNs). These systems produce corresponding speech outputs based on text input. ANNs are especially well suited for this task because of their capacity to produce realistic speech and learn intricate patterns. The speaker identification system was developed using deep learning for Hindi speech. The self-created database using Wave Surfer was applied to the experiment. The Mel Frequency Cepstral coefficients were applied. The comparative analysis of different speaker identification methods was also presented. The researcher can apply the research findings to develop a speaker identification system in low-resource languages.