Wavelet based feature extraction for the vowel sound

These days, how intelligent systems interact intuitively with human and its user friendly application can´t be separated from its user interface´s input method. Those systems as mentioned before, mostly use the human nature based communication method such as speech to make a seamless communication with the user. The main concern in this research is how speech as a input can be recognized and differentiated to be applied as an user-system communication method. That concern leads to this paper. This paper concerned in recognizing the vowel sound in Indonesian language. The method of recognition applied in this research is based on wavelet transform which has been proven on its reliability on feature extraction applications. Applying six different wavelet transform method, this research is set to find which one has the best properties in effectivity and efficiency on performing feature extraction of each vowel sound to be applied in the speech recognition method on the intelligent systems. The result shows the list of features of each vowel sound in Indonesian language that can be used for the next research, and some recommendations on the most effective and efficient wavelet transform to be used in performing vowel sound recognition in Indonesian language.