Speaker Recognition System using CNN in Noisy Environment

The conventional speaker identification systems are not giving hundred percent result. So lots of people are still finding the best way or algorithm to enhance the the accuracy rate of speaker recognition system. There are many traditional methods to recognize the speaker. For example, using Gaussian Mixture model (GMM), using wavelet, wavelet packet transform, using vector quantization and etc. Now a days Neural network system is known for its great accuracy. So, in this article We are going to implement, convolution Neural Network (CNN) for speaker recognitions, in order to obtain best accuracy using suitable activation function and morphological filtering technique.