#Speaker Recognition
This purpose of this project is to identify the people by their audio and recognize authonticity. For the implementation i used CNN and for feature extraction i used MFCC and spectrogram. I used my own dataset to train the model. There is a 10 presons' audio and 10 sample of individual. I got 78 % validation accuracy.Moreover i also worked on a opensource dataset called TIMIT courpus and i got 82% accuracy on that. In this system there a speech_recognition library is also used for the speech to text conversion to identify a command.
