Machine-learning based classification of speech and music

The need to classify audio into categories such as speech or music is an important aspect of many multimedia document retrieval systems. In this paper, we investigate audio features that have not been previously used in music-speech classification, such as the mean and variance of the discrete wavelet transform, the variance of Mel-frequency cepstral coefficients, the root mean square of a lowpass signal, and the difference of the maximum and minimum zero-crossings. We, then, employ fuzzy C-means clustering to the problem of selecting a viable set of features that enables better classification accuracy. Three different classification frameworks have been studied:Multi-Layer Perceptron (MLP) Neural Networks, radial basis functions (RBF) Neural Networks, and Hidden Markov Model (HMM), and results of each framework have been reported and compared. Our extensive experimentation have identified a subset of features that contributes most to accurate classification, and have shown that MLP networks are the most suitable classification framework for the problem at hand.     

Thermodynamics of ionic-association in aqueous solutions of Ca and Mg organic salts using ion-selective electrode technique—I. Formates, acetates, propionates and butyrates

The divalent selective electrode together with high precision solid state, digital pH -mv -meter makes broader application of potentiometry in physical and inorganic chemistry a certainty. The above set-up is used to determine the stoichiometric constants, K, for Ca and Mg ions association with formates, acetates, propionates and butyrates at 25°, 35° and 45°C in aqueous media. The K-values were converted to infinite dilution K_A values were found to be 8.4 LM⁻¹, 10.4 LM⁻¹, 19.1 LM⁻¹ and 19.3 LM⁻¹ for calcium salts of formate, acetate, propionate and butyrate respectively. Also K_A values for Mg salts of formate, acetate, propionate and butyrate were found to be 7.8 LM⁻¹, 9.5 LM⁻¹, 13.1 LM⁻¹ and 13.1 LM⁻¹ respectively. Other thermodynamic parameters such as ΔG°, ΔH° and ΔS° are also obtained from the variation of K_A with temperature for each salt. The data are interpreted relative to each other on basis of pK_a of the corresponding organic acid. Their temperature behaviour is similar to those salts derived from strong acids such as sulphates, rather than weak acids.     

Free vibration of a paraboloidal shell of revolution including shear deformation and rotary inertia

The governing strain-displacement and curvature-displacement equations for paraboloidal shells including shear deformation and rotary inertia are solved for free vibration of closed shells. The finite element method is used to obtain three-dimensional frequency of vibration solutions for a variety of boundary conditions, free, fixed and simply supported. Assumptions concerning the circumferential vibrational behavior are incorporated that reduce the analysis to a single coordinate and the element shape function is formulated using the meridional coordinate. The results for frequency of vibration compare favorably with the available literature. Selected results for frequency of vibration are presented in tabular form for several shell parameters, including free, pinned and fixed boundary conditions. Representative mode shapes are plotted for a fixed boundary condition.     

Adaptive collision-resolution algorithm and traffic-scheduling mechanism for DOCSIS-based HFC networks

A scheme is proposed that provides high QoS and collision-free data transmission in hybrid fiber-coax (HFC) networks. This scheme ensures effective medium access and creates an effective traffic-scheduling mechanism. To enhance the performance of the proposed scheme, a novel methodology has been adopted. Experiments have been performed to measure the effectiveness of the priority system that use the mean access delay, throughput, and channel utilization as figures of merit. Published in Russian in Radiotekhnika i Elektronika, 2007, Vol. 52, No. 3, pp. 469–479. The text was submitted by the authors in English.     

Structure-Preserving Dynamics of Stochastic Epidemic Model with the Saturated Incidence Rate

The structure-preserving features of the nonlinear stochastic models are positivity, dynamical consistency and boundedness. These features have a significant role in different fields of computational biology and many more. Unfortunately, the existing stochastic approaches in literature do not restore aforesaid structure-preserving features, particularly for the stochastic models. Therefore, these gaps should be occupied up in literature, by constructing the structure-preserving features preserving numerical approach. This writing aims to describe the structure-preserving dynamics of the stochastic model. We have analysed the effect of reproduction number in stochastic modelling the same as described in the literature for deterministic modelling. The usual explicit stochastic numerical approaches are time-dependent. We have developed the implicitly driven explicit approach for the stochastic epidemic model. We have proved that the newly developed approach is preserving the structural, dynamical properties as positivity, boundedness and dynamical consistency. Finally, convergence analysis of a newly developed approach and graphically illustration is also presented.     

Four-Step Iteration Scheme to Approximate Fixed Point for Weak Contractions

Fixed point theory is one of the most important subjects in the setting of metric spaces since fixed point theorems can be used to determine the existence and the uniqueness of solutions of such mathematical problems. It is known that many problems in applied sciences and engineering can be formulated as functional equations. Such equations can be transferred to fixed point theorems in an easy manner. Moreover, we use the fixed point theory to prove the existence and uniqueness of solutions of such integral and differential equations. Let X be a non-empty set. A fixed point for a self-mapping T on X is a point     

Improved Software Implementation for Montgomery Elliptic Curve Cryptosystem

The last decade witnessed rapid increase in multimedia and other applications that require transmitting and protecting huge amount of data streams simultaneously. For such applications, a high-performance cryptosystem is compulsory to provide necessary security services. Elliptic curve cryptosystem (ECC) has been introduced as a considerable option. However, the usual sequential implementation of ECC and the standard elliptic curve (EC) form cannot achieve required performance level. Moreover, the widely used Hardware implementation of ECC is costly option and may be not affordable. This research aims to develop a high-performance parallel software implementation for ECC. To achieve this, many experiments were performed to examine several factors affecting ECC performance including the projective coordinates, the scalar multiplication algorithm, the elliptic curve (EC) form, and the parallel implementation. The ECC performance was analyzed using the different factors to tune-up them and select the best choices to increase the speed of the cryptosystem. Experimental results illustrated that parallel Montgomery ECC implementation using homogenous projection achieves the highest performance level, since it scored the shortest time delay for ECC computations. In addition, results showed that NAF algorithm consumes less time to perform encryption and scalar multiplication operations in comparison with Montgomery ladder and binary methods. Java multi-threading technique was adopted to implement ECC computations in parallel. The proposed multithreaded Montgomery ECC implementation significantly improves the performance level compared to previously presented parallel and sequential implementations.     

Within-word pronunciation variation modeling for Arabic ASRs: a direct data-driven approach

Pronunciation variation is a major obstacle in improving the performance of Arabic automatic continuous speech recognition systems. This phenomenon alters the pronunciation spelling of words beyond their listed forms in the pronunciation dictionary, leading to a number of out of vocabulary word forms. This paper presents a direct data-driven approach to model within-word pronunciation variations, in which the pronunciation variants are distilled from the training speech corpus. The proposed method consists of performing phoneme recognition, followed by a sequence alignment between the observation phonemes generated by the phoneme recognizer and the reference phonemes obtained from the pronunciation dictionary. The unique collected variants are then added to dictionary as well as to the language model. We started with a Baseline Arabic speech recognition system based on Sphinx3 engine. The Baseline system is based on a 5.4 hours speech corpus of modern standard Arabic broadcast news, with a pronunciation dictionary of 14,234 canonical pronunciations. The Baseline system achieves a word error rate of 13.39%. Our results show that while the expanded dictionary alone did not add appreciable improvements, the word error rate is significantly reduced by 2.22% when the variants are represented within the language model.