The problem of automatic language identification (LID) can be defined as the process of automatically identifying the language of a given spoken utterance. The need for reliable LID is continuously growing due to the technological trend toward increased human interaction with machines. In this paper a language identification system is implemented using Deep Neural Network (DNN) framework using MFCC features in context free and context dependent manner. DNN models for languages are also built using bottleneck features derived using an autoencoder structure. Performance of proposed method is validated using five Indian languages, namely, Gujarati, Hindi, Malayalam, Telugu, and Urdu
展开▼