Voice Based Dictionary Using DTW and MFCC

Digital processing of voice recognition algorithms is required for highly accurate and fast voice recognition technology. The voice or speech signal contains an information. To represent the voice signal, digital signal processes such as feature extraction technique and feature matching techniques are introduced. In this paper voice based dictionary has been developed, in voice based dictionary words are accessed through spoken alphabet from collected database or dictionary. The developed voice based dictionary system uses silence removal algorithm which remove silence from recorded alphabets, feature extraction technique i.e mel-frequency cepstral coefficients, gives good discrimination of the speech signal. Speech recognition approach i.e. dynamic time warping algorithm, measures similarity between the stored template and test template for speech recognition.