Parallelization of feature extraction techniques on consumer-level multicore system

Three different feature extractions techniques Fast Fourier Transform (FFT), Mel-Frequency Ceptral Coefficient (MFCC), and Discrete Wavelet Transform (DWT) are parallelized in this study and used for classification. The Evolutionary Neural Network (ENN) is used as a classifier. In the scope of classification, ENN is a new technique that can be effectively used as a classifier. This research will help to extract the features in the most efficient way with less computation time in real life use. The parallel FFT, MFCC and DWT are developed in C# for multicore using .Net framework 4.0. The .Net framework offers comprehensive and flexible threads APIs that allow the efficient implementation of multithreaded application.