Speech-related research has a wide range of applications. Most speech-related researches employ Mel-frequency cepstral coefficients (MFCCs) as acoustic features. However, finding the optimum number MFCCs is an active question. MFCC-based speech classification was performed for both vowels and words in Bengali language. As model, deep neural network (DNN) with Adam optimizer used. The performances were measured five different performance metrics, namely confusion matrix, accuracy, area under...

Autism spectrum disorder (ASD) is a neuro-developmental that results in behavioural retardation verbal communications and social interactions. Traditional ASD detection methods involve assessing patients' patterns by medical practitioners, which often lack credibility precision. The contribution of the current study involves 3D-CNN (convolutional neural network) model to diagnose patients from healthy individuals using functional magnetic resonance imaging (fMRI) brain. We utilised publicly...

In this paper, an automatic speech recognition system has been proposed for isolated Bangla word using Support vector machine with Dynamic time warping (DTW). For training purposes, we have collected data from 40 speakers five different words. All the was in a highly acoustic and noise-proof environment. Mel frequency cepstrum coefficients (MFCC's) are considered as static features signal. dynamic features, first second derivatives of MFCC utilized. After determining feature vectors,...

This paper deals with the step-by-step implementation of a Bengali speech classification model variations Mel-Frequency Cepstral Coefficient (MFCC). A deep convolutional neural network is developed for purpose. vowels are representative lower MFCC variational group, and words higher group. The necessary dataset derived from vowel word tokens was utilized analysis network. Then trained validated both isolated separately same number data. performance measured according to four different...

Abstract Although speech recognition has achieved significant success using integrated and efficient models, still some series of challenges remain as linguistic‐acoustic patterns are perturbed by speakers' individual articulation gestures environmental noises. Due to dynamic changes in the vocal tract cavity, word utterances yield temporal features, whereas vowel less‐perturbed quasi‐stationary features. To recognize vowels words, basic feedforward neural network (NN), among other methods,...

The Raman scattering study of vibrational modes and hole concentration in a ferromagnetic semiconductor Ga1−xMnxSb grown by Mn ion implantation, deposition post-annealing has been presented. experiments are performed both implanted unimplanted regions before after etching the samples. spectra measured from region show only GaSb-like phonon modes. On other hand, additional approximately at 115, 152, 269, 437 659 cm−1. experimental results demonstrate that extra associated with surface...

A numerical approach is presented to study the crystallographic orientation-dependent performance of 445nm InGaN/GaN quantum well (QW) Blue Laser by solving a six-band k.p Hamiltonian for wurtzite crystal using finite difference method. Euler's rotation technique used modify wave vector and matrix in conventional (0001) orientation. It found that there substantial correlation energy band dispersion profile, peak gain emission wavelength with change The regular optical gains are inspected as...

In this paper, the crystal orientation-dependent electronic and optical performance of a compressively strained quantum well (QW) blue-violet Vertical Cavity Surface Emitting Laser (VCSEL) is numerically simulated using MATLAB. It found that there strong correlation between orientation energy band dispersion hence emission spectrum laser. The peak gain wavelength depend on fundamental gap, separation HH1 LH1 subbands, momentum matrix element. observed maximum minimum VCSEL 1150 920 cm-1...

The optical properties of compressively strained GaInSb quantum well (QW) laser fabricated in (001), (110), (111), (112) and (113) directions are theoretically investigated by solving one dimensional conduction valence bands Schrödinger equation using finite difference method. band mixing effect is found minimum direction maximum (111) orientation. results obtained from the present study demonstrate that there a strong correlation between gain its emission wavelength with crystal orientation...

For the field of security concerned bounds human identification has fallen out a titanic promising biological or biometric specification based system. Biometric systems worship different kinds properties, e.g., characteristics fingerprint, face, iris posture. The motive this research is to assess mortal psychic behavior by Electroencephalogram (EEG) signal with time-frequency exploration acquiring information from interior changes brain territory. This paperwork represents conditions emerged...

A numerical approach is presented to calculate crystal orientation-dependent effective mass of heavy hole, light hole and spin-orbit split-off III-V semiconductors (GaAs, InAs, InP, GaSb, InSb etc.) at Gamma-point by envelope approximation using k.p method. Unitary transformation used modify the wave vector eight-band Hamiltonian applied in conventional (100) orientation. The simulation carried out for orientations (100), (110) (111) as well non-conventional (113), (131) (201). It found that...

Moiré images that are generally termed as moiré fringes have been generated due to the interference of two repetitive gratings. These patterns can be applied many uses in metrology such measurements surface profilometry aerofoil, stress-strain effects, thermal deformation and so on. Moreover, 3D reconstruction well movement characterization linearly rotary moving objects visualized identified by imaging technique. Recently it is approached an emerging tool fields biotechnology-particularly...

The optical properties of compressively strained GaInSb/GaInAlSb mid-infrared quantum well light sources are numerically studied solving one-dimensional Schrödinger equation using finite difference method. simulation results demonstrate that band-mixing effects and effective mass hole reduced when the is highly strained. strain-dependent differential gains evaluated for different values strain found maximum 1.52% well. emission wavelength proposed laser can be tuned from 2.39 μm to 2.27 due...

This paper represents an assistive robotic technology to assist elderly people who are unable move frequently and walk properly also forget take medicine. The goal is implement a prototype test its capability for helping the elder make their life easy live better. working has been designed four features. First, it heartbeat sensor where doctors or nurses will easily measure patient's rate. second one common box that open by touch using obstacle detector which known as sonar sensor. Third,...

In this paper, the crystallographic orientation-dependent performances of 635nm red quantum well (QW) laser are numerically studied. The electronic and optical characteristics investigated by solving time independent Schrodinger equation using finite difference method. results obtained in (001), (101), (111), (113), (211) orientations found that there is a substantial correlation peak gain emission wavelength with crystal orientations. maximum minimum gains estimated (113) wavelengths 642nm...

In this article, the statistical moments of local spectrum obtained from Gabor transform technique for chirp signal are presented both analytically and numerically. The estimation considering only ridge frequency component electrically band limited components is also discussed.

In this article, impersonation experiments were conducted utilizing natural morphed speeches between (/a/-/b/-/a/) and (/a/-/g/-/a/) as stimuli. The sound stimuli are produced the glottal source linear predictive coding (LPC) filtering coefficients, which represent vocal tract states. An algorithm has been proposed for determination of quality based on redundancies canonical variates sounds. redundancy is defined by variate loading factor corresponding correlation coefficient. imitated...

This study is concentrated on the investigation of neuronal responses induced by Bangla vowels. Neuronal are estimated Auditory Model, which computational representative human auditory nerve fiber that determines neural spike rates elicited any sound wave. rate profiles two Vowels used to determine response dissimilarity proposed here as a vowel discriminating scale. The formulated using cross-correlation profiles. As unequal in length, Fourier transform coefficients calculate...