“…In [4] a novel method of feature extraction for speech recognition have been introduced based on root TDC and it has been modified by using LDA implementation [5]. In previously reported research, some of the extracted features commonly used were: energy function, average zero-crossing rate, fundamental frequency, spectral peak tracks brightness [4]. Also, another approach involves the statistical feature extraction from the Wigner-Ville distribution [5] as well as the combination of statistical features derived from both spectrogram and Wigner-Ville [6].…”