Review of Feature Extraction Techniques

Abstract

Speech is a form of vocal communication used by humans. It is made up of phonetic combinations of vowels and consonant sounds that form words. To analyse these signals we use speech processing which can be considered as a special case of digital signal processing that is specifically applied to speech signals. There are many aspects and parts in speech processing, in this paper we will be discussing about feature extraction techniques. There are various different methods that can be used for the purpose of feature extraction- Linear Predictive Coding, Mel Frequency Cepstral Coefficients, Discrete Wavelet Transforms. Determining which feature extraction technique to use is an extremely crucial step as it heavily impacts the accuracy of the project. In this paper we will be briefly discussing about the different feature extraction techniques and why Mel Frequency Cepstral Coefficients gives the most efficient and effective results.