Comparative Study of Phoneme Recognition Techniques

Abstract

Automatic Speech Recognition is the most popular and demanding field in the research area. For most of the real world applications, Phoneme recognition is important for successful development of ASR. This Paper gives an overview of the techniques and systems for the Phoneme recognition based on three categories-Vector Quantization, HMM, Neural Network followed by comparative study of different techniques. This paper helps in selecting the appropriate technique along with its feature description. Also gives the generalized approach of the phoneme recognition technique to understand their working. This paper concludes with the decision that the present phoneme recognition techniques work better for isolated words then continues speech.