This is a CNN based model build with Keras. The solution here is to build a free application which can help children to be able to be fluent in speech and help them to overcome their fear, free of cost. The solution is divided into various phases: Speech feature extraction (mfcc) Convert the audio to a 2D MFCC Matrix then into an graphical image . Training the deep neural network. Testing and Predictions Checking if there is good probability of match and user said the word rightly.
priyasingh27 / sld-corrector Goto Github PK
View Code? Open in Web Editor NEWThis project forked from sans1012/sld-corrector
This is a CNN based model build with Keras.