Speech recognition systems are ubiquitous and find its application in automated voice control, voice dialling and automated directory assistance. This paper aims at implementing a neural network based isolated spoken word recognition system on an embedded board-Raspberry Pi using open source software called octave. Mel-Frequency Cepstral Coefficient (MFCC) features are extracted from speech signal and given as input to the neural network. The Feed Forward Multi-Layer Perceptron Neural Network trained with back propagation rule is implemented using Octave in Raspberry Pi. TIDIGITS corpus is used for the experiment. Speaker dependent speech recognition results in 100% accuracy but the speaker independent recognition system shows less accuracy. © 2017 IEEE.

John Sahaya Rani Alex

Electronics

School of Electronics Engineering

Chennai Campus

Pranjali p patange

Vellore Institute of Technology (VIT) is a private university located in&nbsp;Tamil Nadu, India. Founded in 1984, as Vellore Engineering College, the institution offers 20 undergraduate, 34 postgraduate, four integrated and four research programs. It has campuses in Vellore, Amravati, Bhopal and Chennai.

VIT is one of the top ranked private universities in India according to NIRF, THE and QS Rankings.&nbsp;Govt. of India has recognized&nbsp;VIT, Vellore as an&nbsp;Institution of Eminence. This has allowed VIT to take independent quality initiatives and move up in world ranking.

&nbsp;

&nbsp;

VIT University

Implementation of ANN based speech recognition system on an embedded board

2017 International Conference on Nextgen Electronic Technologies: Silicon to Software (ICNETS2)

Background: Speech recognition is an active area of research, used to transliterate words vocalized by individuals in order to make them machine recognizable. For real time speech recognition applications the response time, size of training data and recognition accuracy are the important aspects. Methods: A Hybrid speech recognition system is proposed on the basis on Artificial Neural Network (ANN) in this research. The Self Organising Feature Map (SOFM) is used to reduce the feature vector dimensions which are extracted using the Mel-Frequency Cepstrum Coefficients (MFCC), Perceptual Linear Predictive (PLP) and Discrete Wavelet Transform (DWT) methods. The Back Propagation Network (BPN) algorithm is used for training the Artificial Neural Network for pattern classification. Findings: The proposed method is tested with TIDIGITS data. Results indicate that despite of the large reduction in the feature vector dimensions the recognition accuracy obtained using SOFM technique is same as that of the recognition accuracy of the conventional methods. The response time is also fast and the data size of the input data is reduced considerably. The proposed hybrid system is further tested using multilingual isolated digit data.

Indian Journal of Science and Technology

Performance Analysis of SOFM based Reduced Complexity Feature Extraction Methods with back Propagation Neural Network for Multilingual Digit Recognition

Background/Objectives: Recent research focuses on low-power design techniques. This has been mainly inspired by the demand of hand-held electronic devices which have to consume less power. Neural network based classifiers are widely used in speech recognition which needs higher power. Methods/Statistical Analysis: A hybrid approach for designing architecture of Multi-Layer Perceptron (MLP) based Neural Network (NN) for speech recognition is proposed using bipartite tabular method and banking organization method. This approach is prototyped in Xilinx xc3s1200. The Back propagation neural network is trained in MATLAB using TIDIGITS corpus. Using the optimized weights from MATLAB, the proposed prototyped low-power architecture is evaluated in Xilinx. Findings: The outcome shows a considerable reduction both in area and power. Conclusion/Improvements: The reduction in switching power is by 33% and the average power is by 25% are noted. There was 2% reduction in the resources used by the proposed architecture.

Low-Power Multi-Layer Perceptron Neural Network Architecture for Speech Recognition

Neurocomputing

A comparative study on selecting acoustic modeling units in deep neural networks based large vocabulary Chinese speech recognition

2015 IEEE Photonics Conference (IPC)

Phase change dispersion during surface plasmon coupling via nano-objects

International Journal of Computer and Electrical Engineering

Design and Implementation of Neural Network Based Controller for Mobile Robot Navigation in Unknown Environments

Journal	Data powered by Typeset2017 International Conference on Nextgen Electronic Technologies: Silicon to Software (ICNETS2)
Publisher	Data powered by TypesetIEEE
Open Access	No