Vowel Classification and Identification Using the QOBE Algorithm

Speech classification and identification is a wide field of research that has many applications in digital systems employing speech processing. In this paper we explore the use of the quasi-optimal bounding ellipsoid (QOBE) algorithm in classifying and identifying several vowel signals from different speakers. QOBE algorithms have proved to be potential candidates in many digital signal processing applications for their efficiency in estimating the parameter set of an unknown system. We obtain vowels from the TIMIT database for both the training and testing sets. We conduct simulations to evaluate the performance of the QOBE in vowel classification and identification using central estimates. We conclude by comparing the use of QOBE estimates in accurate vowel identification.