Automatic Speech Recognition of Marathi isolated words using Neural Network

Abstract—Speech is the way of communication among the human beings and speech recognition is most interesting area of research from last five decades. Speech recognition is the problem of pattern matching so classification is important part of speech recognition. To develop database Voice signals are sampled directly from the microphone. The proposed method is implemented for Marathi isolated word. 100 words are collected from Marathi language. ASR implemented for 100 speakers give three utterances of 100 words. The features from the signals are extracted using Discrete Wavelet Transforms (DWT) because they are well suitable for processing non-stationary signals like speech because of their multi- resolutional, multi-scale analysis characteristics. Speech recognition is a multiclass classification problem. So, the feature vector set obtained are classified using Artificial Neural Networks (ANN). During classification stage, the input feature vector data is trained using information relating to known patterns and then they are tested using the test data set. Keywords- Speech Recognition, feature extraction, patter