Short-Term Spectral Feature Extraction and Their Fusion in Text Independent Speaker Recognition: A Review

The paper gives an overview of Text-independent short-term-feature-extraction methods of Speaker Recognition System, for clean as well as noisy environment and their fusion at different levels. The basics of extracting feature, which is an imperative component for speaker recognition system, have been discussed along with their variants. The evolution of the conventional methods to ‘Stateof-the-Art’ feature extraction methods are also brought out. This review helps in understanding the developments, which have taken place at various stages, with their relative merits and demerits. A comparative study of different techniques has been done at the end of each section to justify the choice of techniques available in the ‘State-of-the-Art’ speaker recognition systems. This study quantifies the effectiveness of short-term features for speaker identification. Index Terms Short-term feature Extraction, Speaker Recognition, Mel-frequency-Cepstral-coefficients (MFCC), Fusion.