Optimal feature extraction and selection techniques for speech processing: A review

摘要

Today speech processing is one of the demanding applications among all others. This article highlights two important aspects of speech processing, namely which feature representation must be employed and what is their selection criteria. Depending on different application areas, speech processing needs different set of features and techniques to extract them. At the same time it is necessary to choose optimal set of features or optimal size of feature vector in order to reduce timing complexity. In this context we address various features suitable for different speech processing and how these features should be selected are discussed at the length. There is a need to reduce the size of feature vector, thus various techniques for selecting optimal set of features are also discussed in this article. This paper helps a reader to get a quick insight in the area of speech processing.