Improved MFCC Feature Extraction Combining Symmetric ICA Algorithm for Robust Speech Recognition

Abstract

<span style="font-family: Times New Roman; font-size: small;">&nbsp;</span><p class="Abstract" style="margin: 0cm 0cm 0pt; layout-grid-mode: char;"><span style="font-size: x-small;"><span style="font-family: Times New Roman;"><span style="font-weight: normal;">Independent component analysis (ICA), instead of </span><span style="font-weight: normal; mso-fareast-language: ZH-CN;">the </span><span style="font-weight: normal;">traditional discrete cosine transform (DCT), is often used to project log Mel spectrum in robust speech feature extraction. The paper proposed using symmetric orthogonalization in ICA for projecting log Mel spectrum into a new feature space as a substitute in extracting speech features to solve the problem of cumulative error and unequal weights that deflation orthogonalization brings, so as to improve the robustness of speech recognition systems, and increase the efficiency of estimation at the same time. Furthermore, the paper studied the nonlinearities of the objective function in ICA and their coefficients, tested them in all kinds of environments</span><span style="font-weight: normal; mso-fareast-language: ZH-CN;">,</span><span style="font-weight: normal;"> finding that they influenced the recognition rate greatly in speech recognition systems, and applied a new coefficient in the proposed method. Experiments based on HMM and Aurora-2 speech corpus suggested that the new method was superior to deflation-based ICA and MFCC.</span></span></span></p><p class="Abstract" style="margin: 0cm 0cm 0pt; layout-grid-mode: char;"><em style="mso-bidi-font-style: normal;"><span style="font-weight: normal; mso-bidi-font-size: 9.0pt;"></span></em></p

Computer Aided Diagnosis System Feature Extraction of Alzheimer Disease Using MFCC

“…The MFCC feature extraction steps can be summarized as follows (Ittichaichareon et al;Loughran et al, 2008;Molau et al, 2001;Tawfik et al;Zhao et al, 2012):…”

“…In Molau, Pitz, Schlüter, and Ney (2001), it presents a method to derive Mel-frequency cepstral coefficients directly from the power spectrum of a speech signal and it used linear discriminant analysis (LDA) for feature extraction. In Zhao, Zhao, Liu, and Yu (2012) improve the robustness of speech recognition systems and increase the efficiency of estimation at the same time using the Independent component analysis (ICA) for feature extraction. In Loughran, Walker, O'Neill, and O'Farrell (2008), it determined how many coefficients are suitable for musical sound identification using MFCC, next it used Principal Component Analysis (PCA) to reduce the dimensionality of the data.…”

Analysis Of Inertial Stability Dynamical For Robot Based On Lie Algebra

“…From purely mathematical point of view, affine Lie algebras are interesting because their representation theory, like representation theory of finite dimensional, semisimple Lie algebras is much better understood than that of general Kac-Moody algebras. As observed by Victor Kac, the character formula for representations of affine Lie algebras implies certain combinatorial identities, the Macdonald identities [1,2]. Artificial potential field method is a kind of virtual force method proposed by Khatib in 1986, the method regards mobile robot's movement in the environment as a virtual manual movement of the force field, and has a path planning model through environmental potential field.…”

Multi-agent Remote Sensing Image Segmentation Algorithm

“…High spatial resolution remote sensing image (HSRI) has rich texture information and shape information, compared with the low resolution image, can get more sophisticated surface feature information [7][8][9]. However, with the improvement of spatial resolution images, spectrum distribution of features in images is more complex, and of different feature spectrum overlapping, feature information present highly details, the feature classes internal variance in images increases, variance in different features classes decreases, and these characteristics make computer interpretation way must be transform from the traditional way of pixels oriented spectrum processing to object oriented multi features approach [10][11][12].…”