Audio Fingerprinting to Achieve Greater Accuracy and Maximum Speed with Multi-Model CNN-RNN-LSTM in Speaker Identification

The process of matching speech data with database records is known as speaker identification. The major objective of this paper is to find the accuracy and speed in comparison of training set database from RAVDESS with the test signal using neural network methods of Convolutional Neural Network (CNN), Recurrent Neural Network (RNN) along Long Short-Term Memory (LSTM) with combination of audio fingerprinting technique. Speech is most fundamental form of human communication and language is the primary means of exchange among humans. An essential component of social interaction pitch and tone changes are grouped together while accounting for a wide range of issues. The audio fingerprint of voice was produced after the background noise was eliminated. Dataset of RAVDESS using multilayer perception, Audio fingerprinting and CNN, RNN with LSTM to contrast the results with speed and accuracy measures. The machine will ultimately display the gender determination in relation to words per second and accuracy in terms of no of epochs has been observed .and the results show that every classifier for the dataset performs faster and with higher accuracy.