This document describes the implementation of a back-propagation neural network for isolated Bangla speech recognition. The network was trained on Mel Frequency Cepstral Coefficient (MFCC) features extracted from recordings of 10 Bangla digits spoken by 10 speakers. The network architecture included an input layer of 250 neurons, a hidden layer of 16 neurons, and an output layer of 10 neurons. The network was trained using backpropagation and achieved a recognition rate of 96.3% for known speakers and 92% for unknown speakers. The system demonstrates the potential for developing speaker-independent isolated digit speech recognition in Bangla.
Implementation of Back-Propagation Neural Network using Scilab and its Conver...IJEEE
Artificial neural network has been widely used for solving non-linear complex tasks. With the development of computer technology, machine learning techniques are becoming good choice. The selection of the machine learning technique depends upon the viability for particular application. Most of the non-linear problems have been solved using back propagation based neural network. The training time of neural network is directly affected by convergence speed. Several efforts are done to improve the convergence speed of back propagation algorithm. This paper focuses on the implementation of back-propagation algorithm and an effort to improve its convergence speed. The algorithm is written in SCILAB. UCI standard data set is used for analysis purposes. Proposed modification in standard backpropagation algorithm provides substantial improvement in the convergence speed.
Implementation of Back-Propagation Neural Network using Scilab and its Conver...IJEEE
Artificial neural network has been widely used for solving non-linear complex tasks. With the development of computer technology, machine learning techniques are becoming good choice. The selection of the machine learning technique depends upon the viability for particular application. Most of the non-linear problems have been solved using back propagation based neural network. The training time of neural network is directly affected by convergence speed. Several efforts are done to improve the convergence speed of back propagation algorithm. This paper focuses on the implementation of back-propagation algorithm and an effort to improve its convergence speed. The algorithm is written in SCILAB. UCI standard data set is used for analysis purposes. Proposed modification in standard backpropagation algorithm provides substantial improvement in the convergence speed.
Classification by back propagation, multi layered feed forward neural network...bihira aggrey
Classification by Back Propagation, Multi-layered feed forward Neural Networks - Provides a basic introduction of classification in data mining with neural networks
Web spam classification using supervised artificial neural network algorithmsaciijournal
Due to the rapid growth in technology employed by the spammers, there is a need of classifiers that are more efficient, generic and highly adaptive. Neural Network based technologies have high ability of adaption as well as generalization. As per our knowledge, very little work has been done in this field using neural network. We present this paper to fill this gap. This paper evaluates performance of three supervised learning algorithms of artificial neural network by creating classifiers for the complex problem of latest web spam pattern classification. These algorithms are Conjugate Gradient algorithm, Resilient Backpropagation learning, and Levenberg-Marquardt algorithm.
Here is my class on the multilayer perceptron where I look at the following:
1.- The entire backproagation algorithm based in the gradient descent
However, I am planning the tanning based in Kalman filters.
2.- The use of matrix computations to simplify the implementations.
I hope you enjoy it.
Review: “Implementation of Feedforward and Feedback Neural Network for Signal...IJERA Editor
Main focus of project is on implementation of Neural Network Architecture (NNA) with on chip learning on
Analog VLSI Technology for signal processing application. In the proposed paper the analog components like
Gilbert Cell Multiplier (GCM), Neuron Activation Function (NAF) are used to implement artificial NNA.
Analog components used comprises of multiplier, adder and tan sigmoidal function circuit using MOS transistor.
This Neural Architecture is trained using Back Propagation (BP) Algorithm in analog domain with new
techniques of weight storage. Layout design and verification of above design is carried out using VLSI Backend
Microwind 3.1 software Tool. The technology used to design layout is 32 nm CMOS Technology.
Introduction to Artificial Neural NetworksAdri Jovin
This presentation describes the various components, classification and application of Artificial Neural Networks. It also gives an outline on the other soft computing techniques also.
Fundamental, An Introduction to Neural NetworksNelson Piedra
An introduction to Neural Networks, eight edition, 1996
Authors: Ben Krose, Faculty of Mathematics & Computer Science, University of Amsterdam. Patrick wan der Smagt, Institute of Robotics and Systems Dynamics, German Aerospace Research Establishment
Keynote: Nelson Piedra, Computer Sciences School - Advanced Tech, Technical University of Loja UTPL, Ecuador.
Adaptive modified backpropagation algorithm based on differential errorsIJCSEA Journal
A new efficient modified back propagation algorithm with adaptive learning rate is proposed to increase the convergence speed and to minimize the error. The method eliminates initial fixing of learning rate through trial and error and replaces by adaptive learning rate. In each iteration, adaptive learning rate for output and hidden layer are determined by calculating differential linear and nonlinear errors of output layer and hidden layer separately. In this method, each layer has different learning rate in each iteration. The performance of the proposed algorithm is verified by the simulation results.
Classification by back propagation, multi layered feed forward neural network...bihira aggrey
Classification by Back Propagation, Multi-layered feed forward Neural Networks - Provides a basic introduction of classification in data mining with neural networks
Web spam classification using supervised artificial neural network algorithmsaciijournal
Due to the rapid growth in technology employed by the spammers, there is a need of classifiers that are more efficient, generic and highly adaptive. Neural Network based technologies have high ability of adaption as well as generalization. As per our knowledge, very little work has been done in this field using neural network. We present this paper to fill this gap. This paper evaluates performance of three supervised learning algorithms of artificial neural network by creating classifiers for the complex problem of latest web spam pattern classification. These algorithms are Conjugate Gradient algorithm, Resilient Backpropagation learning, and Levenberg-Marquardt algorithm.
Here is my class on the multilayer perceptron where I look at the following:
1.- The entire backproagation algorithm based in the gradient descent
However, I am planning the tanning based in Kalman filters.
2.- The use of matrix computations to simplify the implementations.
I hope you enjoy it.
Review: “Implementation of Feedforward and Feedback Neural Network for Signal...IJERA Editor
Main focus of project is on implementation of Neural Network Architecture (NNA) with on chip learning on
Analog VLSI Technology for signal processing application. In the proposed paper the analog components like
Gilbert Cell Multiplier (GCM), Neuron Activation Function (NAF) are used to implement artificial NNA.
Analog components used comprises of multiplier, adder and tan sigmoidal function circuit using MOS transistor.
This Neural Architecture is trained using Back Propagation (BP) Algorithm in analog domain with new
techniques of weight storage. Layout design and verification of above design is carried out using VLSI Backend
Microwind 3.1 software Tool. The technology used to design layout is 32 nm CMOS Technology.
Introduction to Artificial Neural NetworksAdri Jovin
This presentation describes the various components, classification and application of Artificial Neural Networks. It also gives an outline on the other soft computing techniques also.
Fundamental, An Introduction to Neural NetworksNelson Piedra
An introduction to Neural Networks, eight edition, 1996
Authors: Ben Krose, Faculty of Mathematics & Computer Science, University of Amsterdam. Patrick wan der Smagt, Institute of Robotics and Systems Dynamics, German Aerospace Research Establishment
Keynote: Nelson Piedra, Computer Sciences School - Advanced Tech, Technical University of Loja UTPL, Ecuador.
Adaptive modified backpropagation algorithm based on differential errorsIJCSEA Journal
A new efficient modified back propagation algorithm with adaptive learning rate is proposed to increase the convergence speed and to minimize the error. The method eliminates initial fixing of learning rate through trial and error and replaces by adaptive learning rate. In each iteration, adaptive learning rate for output and hidden layer are determined by calculating differential linear and nonlinear errors of output layer and hidden layer separately. In this method, each layer has different learning rate in each iteration. The performance of the proposed algorithm is verified by the simulation results.
Multidimensional Perceptual Map for Project Prioritization and Selection - 20...Jack Zheng
Traditional perceptual maps are created using scatter charts or quadrant diagrams, which are based on two dimensions (X and Y axes). Then data items are plotted on the plane based on their values for the two attributes.
The multidimensional perceptual map does not rely on the definition of any fixed axes. The map is composed of smaller areas (cells), which are characterized by a vector of values that represent multiple attributes (dimensions). The positioning of data items in the map is determined by its calculated measure (usually Euclidean distance) again each cell. An unsupervised clustering technique called Self-Organizing Map (SOM) is used to generate such maps.
The multidimensional perceptual map ca be used in many areas including project portfolio management, project prioritization, marketing research, product evaluation, performance management, portfolio management, etc.
The project was started with a sole aim in mind that the design should be able to recognize the voice of a person by analyzing the speech signal. The simulation is done in MATLAB. The design of the project is based on using the Linear prediction filter coefficient (LPC) and Principal component analysis (PCA) on data (princomp) for the speech signal analysis. The Sample Collection process is accomplished by using the microphone to record the speech of male/female. After executing the program the speech is analyzed by the analysis part of our MATLAB program code and our design should be able to identify and give the judgment that the recorded speech signal is same as that of our desired output.
কারেন্ট ওয়ার্ল্ড বা কারেন্ট নিউজ বা কারেন্ট অ্যাফেয়ার্স বা সাম্প্রতিক সকল আপডেট তথ্য
গুরুত্বপূর্ণ সাম্প্রতিক সকল সাধারণ জ্ঞান (A-Z Recent & Latest General Knowledge)
(বিসিএস প্রিলিমিনারি ও যে কোন প্রতিযোগিতা মূলক পরীক্ষার প্রস্তুতির জন্য শেষ মুহূর্তের সাজেশন…
এছাড়া বাংলাদেশের যে কোনো প্রতিযোগিতা-মূলক পরীক্ষা, যেমন বিসিএস, ব্যাংক, কর্পোরেট প্রতিষ্ঠানে চাকরির জন্যে সাম্প্রতিক সাধারন জ্ঞান সম্পর্কে সম্যক জ্ঞান থাকা অত্যাবশ্যক।)
গুরুত্বপূর্ণ ফিচার সমূহঃ
বাংলাদেশের প্রশাসনিক ও ভৌগোলিক আপডেট
ব্যখ্যা সহ ৫০০টি গুরুত্বপূর্ণ সাম্প্রতিক সাধারণ জ্ঞান
বাংলাদেশের অর্থনৈতিক সমীক্ষা ২০১৬
জাতীয় বাজেট ২০১৬-১৭
রাষ্ট্র-সংস্থা-সংগঠনে নতুন যাঁরা (২০১৫-১৬)
সাম্প্রতিক চুক্তি ও সম্মেলন
সাম্প্রতিক আন্তর্জাতিক তথ্য
বাংলাদেশের সাম্প্রতিক আলোচিত বিষয়বলী
বাংলাদেশের ব্যাংক গুলোর আপডেট
বিগত বিসিএস পরীক্ষার প্রশ্নের আপডেট তথ্য
নোবেল পুরুস্কার-২০১৫
নোবেল পুরস্কার বিজয়ীদের সংক্ষিপ্ত পরিচিতি
নোবেল পুরস্কারের গুরুত্বপূর্ণ তথ্য
ফুটবল সম্পর্কিত আপডেট তথ্য
সালতামামি-২০১৫
বাংলাদেশে প্রাপ্ত ফল সমূহের পুষ্টিগুণ, ঔষধিগুণ ও স্বাস্থ্য উপকারিতা , 190 পৃষ্ঠার সম্পূর্ণ বাংলা বই ... খুব গুরুত্বপূর্ণ একটি বই
৫০টি বিস্তারিত ব্যখ্যা ছবি ...প্রতেকটি ফলের ----- চার্ট আকারে পুষ্টিগুণ বা উপাদান ... বিস্তারিত স্বাস্থ্য উপকারিতা...এবং এর ঔষধিগুণ বা ভেষজগুন বা আয়ুর্বেদিকগুণ ও তৈরি প্রক্রিয়া ... এবং রূপচর্চায় এর ব্যবহার ও অন্য কাজে এর ব্যবহার
Neural network based numerical digits recognization using nnt in matlabijcses
Artificial neural networks are models inspired by human nervous system that is capable of learning. One of
the important applications of artificial neural network is character Recognition. Character Recognition
finds its application in number of areas, such as banking, security products, hospitals, in robotics also.
This paper is based on a system that recognizes a english numeral, given by the user, which is already
trained on the features of the numbers to be recognized using NNT (Neural network toolbox) .The system
has a neural network as its core, which is first trained on a database. The training of the neural network
extracts the features of the English numbers and stores in the database. The next phase of the system is to
recognize the number given by the user. The features of the number given by the user are extracted and
compared with the feature database and the recognized number is displayed.
Web Spam Classification Using Supervised Artificial Neural Network Algorithmsaciijournal
Due to the rapid growth in technology employed by the spammers, there is a need of classifiers that are
more efficient, generic and highly adaptive. Neural Network based technologies have high ability of
adaption as well as generalization. As per our knowledge, very little work has been done in this field using
neural network. We present this paper to fill this gap. This paper evaluates performance of three supervised
learning algorithms of artificial neural network by creating classifiers for the complex problem of latest
web spam pattern classification. These algorithms are Conjugate Gradient algorithm, Resilient Backpropagation learning, and Levenberg-Marquardt algorithm.
Objective Evaluation of a Deep Neural Network Approach for Single-Channel Spe...csandit
Single-channel speech intelligibility enhancement is much more difficult than multi-channel
intelligibility enhancement. It has recently been reported that machine learning training-based
single-channel speech intelligibility enhancement algorithms perform better than traditional
algorithms. In this paper, the performance of a deep neural network method using a multiresolution
cochlea-gram feature set recently proposed to perform single-channel speech
intelligibility enhancement processing is evaluated. Various conditions such as different
speakers for training and testing as well as different noise conditions are tested. Simulations
and objective test results show that the method performs better than another deep neural
networks setup recently proposed for the same task, and leads to a more robust convergence
compared to a recently proposed Gaussian mixture model approach.
ON THE PERFORMANCE OF INTRUSION DETECTION SYSTEMS WITH HIDDEN MULTILAYER NEUR...IJCNCJournal
Deep learning applications, especially multilayer neural network models, result in network intrusion detection with high accuracy. This study proposes a model that combines a multilayer neural network with Dense Sparse Dense (DSD) multi-stage training to simultaneously improve the criteria related to the performance of intrusion detection systems on a comprehensive dataset UNSW-NB15. We conduct experiments on many neural network models such as Recurrent Neural Network (RNN), Long-Short Term Memory (LSTM), Gated Recurrent Unit (GRU), etc. to evaluate the combined efficiency with each model through many criteria such as accuracy, detection rate, false alarm rate, precision, and F1-Score.
On The Performance of Intrusion Detection Systems with Hidden Multilayer Neur...IJCNCJournal
Deep learning applications, especially multilayer neural network models, result in network intrusion detection with high accuracy. This study proposes a model that combines a multilayer neural network with Dense Sparse Dense (DSD) multi-stage training to simultaneously improve the criteria related to the performance of intrusion detection systems on a comprehensive dataset UNSW-NB15. We conduct experiments on many neural network models such as Recurrent Neural Network (RNN), Long-Short Term Memory (LSTM), Gated Recurrent Unit (GRU), etc. to evaluate the combined efficiency with each model through many criteria such as accuracy, detection rate, false alarm rate, precision, and F1-Score.
Text independent speaker recognition using combined lpc and mfc coefficientseSAT Publishing House
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology
Comparison of Neural Network Training Functions for Hematoma Classification i...IOSR Journals
Classification is one of the most important task in application areas of artificial neural networks
(ANN).Training neural networks is a complex task in the supervised learning field of research. The main
difficulty in adopting ANN is to find the most appropriate combination of learning, transfer and training
function for the classification task. We compared the performances of three types of training algorithms in feed
forward neural network for brain hematoma classification. In this work we have selected Gradient Descent
based backpropagation, Gradient Descent with momentum, Resilence backpropogation algorithms. Under
conjugate based algorithms, Scaled Conjugate back propagation, Conjugate Gradient backpropagation with
Polak-Riebreupdates(CGP) and Conjugate Gradient backpropagation with Fletcher-Reeves updates (CGF).The
last category is Quasi Newton based algorithm, under this BFGS, Levenberg-Marquardt algorithms are
selected. Proposed work compared training algorithm on the basis of mean square error, accuracy, rate of
convergence and correctness of the classification. Our conclusion about the training functions is based on the
simulation results
International Refereed Journal of Engineering and Science (IRJES)irjes
The core of the vision IRJES is to disseminate new knowledge and technology for the benefit of all, ranging from academic research and professional communities to industry professionals in a range of topics in computer science and engineering. It also provides a place for high-caliber researchers, practitioners and PhD students to present ongoing research and development in these areas.
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
Implementation of recurrent neural network for the forecasting of USD buy ra...IJECEIAES
This study implements a recurrent neural network (RNN) by comparing two RNN network structures, namely Elman and Jordan using the backpropagation through time (BPTT) programming algorithm in the training and forecasting process in foreign exchange forecasting cases. The activation functions used are the linear transfer function, the tan-sigmoid transfer function (Tansig), and the log-sigmoid transfer function (Logsig), which are applied to the hidden and output layers. The application of the activation function results in the log-sigmoid transfer function being the most appropriate activation function for the hidden layer, while the linear transfer function is the most appropriate activation function for the output layer. Based on the results of training and forecasting the USD against IDR currency, the Elman BPTT method is better than the Jordan BPTT method, with the best iteration being the 4000th iteration for both. The lowest root mean square error (RMSE) values for training and forecasting produced by Elman BPTT were 0.073477 and 122.15 the following day, while the Jordan backpropagation RNN method yielded 0.130317 and 222.96 also the following day
Image Steganography Using Wavelet Transform And Genetic AlgorithmAM Publications
This paper presents the application of Wavelet Transform and Genetic Algorithm in a novel
steganography scheme. We employ a genetic algorithm based mapping function to embed data in Discrete Wavelet
Transform coefficients in 4x4 blocks on the cover image. The optimal pixel adjustment process is applied after
embedding the message. We utilize the frequency domain to improve the robustness of steganography and, we
implement Genetic Algorithm and Optimal Pixel Adjustment Process to obtain an optimal mapping function to
reduce the difference error between the cover and the stego-image, therefore improving the hiding capacity with
low distortions. Our Simulation results reveal that the novel scheme outperforms adaptive steganography technique
based on wavelet transform in terms of peak signal to noise ratio and capacity, 39.94 dB and 50% respectively.
Efficient And Improved Video Steganography using DCT and Neural NetworkIJSRD
As per the demand of modern communication it is important to establish secret communication which is obtain by seganography .Video Steganography is the technique of hiding some covert message inside a video. The addition of this information to the video is not recognizable through the human eye as modify of a pixel color is negligible. In the proposed method Discrete Cosine Transform (DCT) and neural network is used. Input image is divided into blocks and is processed to generate quantization matrix of cover and stego images by using Discrete Cosine Transform (DCT).And using neural network performance of this method can be further improved. The neural network is trained and on the basis of training and segmentation done, neural network provide efficient positions where data can be merge. The performance and efficiency is measured by PSNR and MSE value.
Similar to Implementation Of Back-Propagation Neural Network For Isolated Bangla Speech Recognition (20)
Submit Your Research Articles - International Journal of Information Sciences...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
INFORMATION THEORY BASED ANALYSIS FOR UNDERSTANDING THE REGULATION OF HLA GEN...ijistjournal
Considering information entropy (IE), HLA surface expression (SE) regulation phenomenon is considered as information propagation channel with an amount of distortion. HLA gene SE is considered as sink regulated by the inducible transcription factors (TFs) (source). Previous work with a certain number of bin size, IEs for source and receiver is computed and computation of mutual information characterizes the dependencies of HLA gene SE on some certain TFs in different cells types of hematopoietic system under the condition of leukemia. Though in recent time information theory is utilized for different biological knowledge generation and different rules are available in those specific domains of biomedical areas; however, no such attempt is made regarding gene expression regulation, hence no such rule is available. In this work, IE calculation with varying bin size considering the number of bins is approximately half of the sample size of an attribute also confirms the previous inferences.
Call for Research Articles - 5th International Conference on Artificial Intel...ijistjournal
5th International Conference on Artificial Intelligence and Machine Learning (CAIML 2024) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of Artificial Intelligence and Machine Learning. The Conference looks for significant contributions to all major fields of the Artificial Intelligence, Machine Learning in theoretical and practical aspects. The aim of the Conference is to provide a platform to the researchers and practitioners from both academia as well as industry to meet and share cutting-edge development in the field.
Authors are solicited to contribute to the conference by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the areas of Computer Science, Engineering and Applications.
Online Paper Submission - International Journal of Information Sciences and T...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
SYSTEM IDENTIFICATION AND MODELING FOR INTERACTING AND NON-INTERACTING TANK S...ijistjournal
System identification from the experimental data plays a vital role for model based controller design. Derivation of process model from first principles is often difficult due to its complexity. The first stage in the development of any control and monitoring system is the identification and modeling of the system. Each model is developed within the context of a specific control problem. Thus, the need for a general system identification framework is warranted. The proposed framework should be able to adapt and emphasize different properties based on the control objective and the nature of the behavior of the system. Therefore, system identification has been a valuable tool in identifying the model of the system based on the input and output data for the design of the controller. The present work is concerned with the identification of transfer function models using statistical model identification, process reaction curve method, ARX model, genetic algorithm and modeling using neural network and fuzzy logic for interacting and non interacting tank process. The identification technique and modeling used is prone to parameter change & disturbance. The proposed methods are used for identifying the mathematical model and intelligent model of interacting and non interacting process from the real time experimental data.
Call for Research Articles - 4th International Conference on NLP & Data Minin...ijistjournal
4th International Conference on NLP & Data Mining (NLDM 2024) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of Natural Language Computing and Data Mining.
Authors are solicited to contribute to the conference by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the following areas, but are not limited to.
Research Article Submission - International Journal of Information Sciences a...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
Call for Papers - International Journal of Information Sciences and Technique...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
Implementation of Radon Transformation for Electrical Impedance Tomography (EIT)ijistjournal
Radon Transformation is generally used to construct optical image (like CT image) from the projection data in biomedical imaging. In this paper, the concept of Radon Transformation is implemented to reconstruct Electrical Impedance Topographic Image (conductivity or resistivity distribution) of a circular subject. A parallel resistance model of a subject is proposed for Electrical Impedance Topography(EIT) or Magnetic Induction Tomography(MIT). A circular subject with embedded circular objects is segmented into equal width slices from different angles. For each angle, Conductance and Conductivity of each slice is calculated and stored in an array. A back projection method is used to generate a two-dimensional image from one-dimensional projections. As a back projection method, Inverse Radon Transformation is applied on the calculated conductance and conductivity to reconstruct two dimensional images. These images are compared to the target image. In the time of image reconstruction, different filters are used and these images are compared with each other and target image.
Online Paper Submission - 6th International Conference on Machine Learning & ...ijistjournal
6th International Conference on Machine Learning & Applications (CMLA 2024) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of on Machine Learning & Applications.
Authors are solicited to contribute to the conference by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the following areas, but are not limited to.
Submit Your Research Articles - International Journal of Information Sciences...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
BER Performance of MPSK and MQAM in 2x2 Almouti MIMO Systemsijistjournal
Almouti published the error performance of the 2x2 space-time transmit diversity scheme using BPSK. One of the key techniques employed for correcting such errors is the Quadrature amplitude modulation (QAM) because of its efficiency in power and bandwidth.. In this paper we explore the error performance of the 2x2 MIMO system using the Almouti space-time codes for higher order PSK and M-ary QAM. MATLAB was used to simulate the system; assuming slow fading Rayleigh channel and additive white Gaussian noise. The simulated performance curves were compared and evaluated with theoretical curves obtained using BER tool on the MATLAB by setting parameters for random generators. The results shows that the technique used do find a place in correcting error rates of QAM system of higher modulation schemes. The model can equally be used not only for the criteria of adaptive modulation but for a platform to design other modulation systems as well.
Online Paper Submission - International Journal of Information Sciences and T...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
Call for Papers - International Journal of Information Sciences and Technique...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
International Journal of Information Sciences and Techniques (IJIST)ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
BRAIN TUMOR MRIIMAGE CLASSIFICATION WITH FEATURE SELECTION AND EXTRACTION USI...ijistjournal
Feature extraction is a method of capturing visual content of an image. The feature extraction is the process to represent raw image in its reduced form to facilitate decision making such as pattern classification. We have tried to address the problem of classification MRI brain images by creating a robust and more accurate classifier which can act as an expert assistant to medical practitioners. The objective of this paper is to present a novel method of feature selection and extraction. This approach combines the Intensity, Texture, shape based features and classifies the tumor as white matter, Gray matter, CSF, abnormal and normal area. The experiment is performed on 140 tumor contained brain MR images from the Internet Brain Segmentation Repository. The proposed technique has been carried out over a larger database as compare to any previous work and is more robust and effective. PCA and Linear Discriminant Analysis (LDA) were applied on the training sets. The Support Vector Machine (SVM) classifier served as a comparison of nonlinear techniques Vs linear ones. PCA and LDA methods are used to reduce the number of features used. The feature selection using the proposed technique is more beneficial as it analyses the data according to grouping class variable and gives reduced feature set with high classification accuracy.
Research Article Submission - International Journal of Information Sciences a...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
A MEDIAN BASED DIRECTIONAL CASCADED WITH MASK FILTER FOR REMOVAL OF RVINijistjournal
In this paper A Median Based Directional Cascaded with Mask (MBDCM) filter has been proposed, which is based on three different sized cascaded filtering windows. The differences between the current pixel and its neighbors aligned with four main directions are considered for impulse detection. A direction index is used for each edge aligned with a given direction. Minimum of these four direction indexes is used for impulse detection under each masking window. Depending on the minimum direction indexes among these three windows new value to substitute the noisy pixel is calculated. Extensive simulations showed that the MBDCM filter provides good performances of suppressing impulses from both gray level and colored benchmarked images corrupted with low noise level as well as for highly dense impulses. MBDCM filter gives better results than MDWCMM filter in suppressing impulses from highly corrupted digital images.
DECEPTION AND RACISM IN THE TUSKEGEE SYPHILIS STUDYijistjournal
During the twentieth century (1932-1972), white physicians representing the United States government
conducted a human experiment known as the Tuskegee Syphilis Study on black syphilis patients in Macon
County, Alabama. The creators of the study, who supported the idea of black inferiority and the concept
that black people’s bodies functioned differently from white people’s, observed the effects of a disease
called syphilis on untreated black patients in order to collect data for further research on syphilis. Black
individuals involved with the study believed that they were receiving treatment, although in truth,
treatments for syphilis were purposely held back from them. Not only this, but fluids from their bodies, such
as blood and spinal fluid, were extracted to serve as research material without their awareness of the
purpose of the collection. The physicians justified their approach by positioning it as mere observation,
asserting that they were not actively intervening with the patients participating in the experiment. However,
despite their claims of passivity, these white physicians engaged in various morally improper actions,
including deceit, which ultimately resulted in the deaths of numerous black patients who might have had a
chance at survival.
Online Paper Submission - International Journal of Information Sciences and T...ijistjournal
The International Journal of Information Science & Techniques (IJIST) focuses on information systems science and technology coercing multitude applications of information systems in business administration, social science, biosciences, and humanities education, library sciences management, depiction of data and structural illustration, big data analytics, information economics in real engineering and scientific problems.
This journal provides a forum that impacts the development of engineering, education, technology management, information theories and application validation. It also acts as a path to exchange novel and innovative ideas about Information systems science and technology.
Removing Uninteresting Bytes in Software FuzzingAftab Hussain
Imagine a world where software fuzzing, the process of mutating bytes in test seeds to uncover hidden and erroneous program behaviors, becomes faster and more effective. A lot depends on the initial seeds, which can significantly dictate the trajectory of a fuzzing campaign, particularly in terms of how long it takes to uncover interesting behaviour in your code. We introduce DIAR, a technique designed to speedup fuzzing campaigns by pinpointing and eliminating those uninteresting bytes in the seeds. Picture this: instead of wasting valuable resources on meaningless mutations in large, bloated seeds, DIAR removes the unnecessary bytes, streamlining the entire process.
In this work, we equipped AFL, a popular fuzzer, with DIAR and examined two critical Linux libraries -- Libxml's xmllint, a tool for parsing xml documents, and Binutil's readelf, an essential debugging and security analysis command-line tool used to display detailed information about ELF (Executable and Linkable Format). Our preliminary results show that AFL+DIAR does not only discover new paths more quickly but also achieves higher coverage overall. This work thus showcases how starting with lean and optimized seeds can lead to faster, more comprehensive fuzzing campaigns -- and DIAR helps you find such seeds.
- These are slides of the talk given at IEEE International Conference on Software Testing Verification and Validation Workshop, ICSTW 2022.
Building RAG with self-deployed Milvus vector database and Snowpark Container...Zilliz
This talk will give hands-on advice on building RAG applications with an open-source Milvus database deployed as a docker container. We will also introduce the integration of Milvus with Snowpark Container Services.
LF Energy Webinar: Electrical Grid Modelling and Simulation Through PowSyBl -...DanBrown980551
Do you want to learn how to model and simulate an electrical network from scratch in under an hour?
Then welcome to this PowSyBl workshop, hosted by Rte, the French Transmission System Operator (TSO)!
During the webinar, you will discover the PowSyBl ecosystem as well as handle and study an electrical network through an interactive Python notebook.
PowSyBl is an open source project hosted by LF Energy, which offers a comprehensive set of features for electrical grid modelling and simulation. Among other advanced features, PowSyBl provides:
- A fully editable and extendable library for grid component modelling;
- Visualization tools to display your network;
- Grid simulation tools, such as power flows, security analyses (with or without remedial actions) and sensitivity analyses;
The framework is mostly written in Java, with a Python binding so that Python developers can access PowSyBl functionalities as well.
What you will learn during the webinar:
- For beginners: discover PowSyBl's functionalities through a quick general presentation and the notebook, without needing any expert coding skills;
- For advanced developers: master the skills to efficiently apply PowSyBl functionalities to your real-world scenarios.
GridMate - End to end testing is a critical piece to ensure quality and avoid...ThomasParaiso2
End to end testing is a critical piece to ensure quality and avoid regressions. In this session, we share our journey building an E2E testing pipeline for GridMate components (LWC and Aura) using Cypress, JSForce, FakerJS…
Communications Mining Series - Zero to Hero - Session 1DianaGray10
This session provides introduction to UiPath Communication Mining, importance and platform overview. You will acquire a good understand of the phases in Communication Mining as we go over the platform with you. Topics covered:
• Communication Mining Overview
• Why is it important?
• How can it help today’s business and the benefits
• Phases in Communication Mining
• Demo on Platform overview
• Q/A
zkStudyClub - Reef: Fast Succinct Non-Interactive Zero-Knowledge Regex ProofsAlex Pruden
This paper presents Reef, a system for generating publicly verifiable succinct non-interactive zero-knowledge proofs that a committed document matches or does not match a regular expression. We describe applications such as proving the strength of passwords, the provenance of email despite redactions, the validity of oblivious DNS queries, and the existence of mutations in DNA. Reef supports the Perl Compatible Regular Expression syntax, including wildcards, alternation, ranges, capture groups, Kleene star, negations, and lookarounds. Reef introduces a new type of automata, Skipping Alternating Finite Automata (SAFA), that skips irrelevant parts of a document when producing proofs without undermining soundness, and instantiates SAFA with a lookup argument. Our experimental evaluation confirms that Reef can generate proofs for documents with 32M characters; the proofs are small and cheap to verify (under a second).
Paper: https://eprint.iacr.org/2023/1886
Climate Impact of Software Testing at Nordic Testing DaysKari Kakkonen
My slides at Nordic Testing Days 6.6.2024
Climate impact / sustainability of software testing discussed on the talk. ICT and testing must carry their part of global responsibility to help with the climat warming. We can minimize the carbon footprint but we can also have a carbon handprint, a positive impact on the climate. Quality characteristics can be added with sustainability, and then measured continuously. Test environments can be used less, and in smaller scale and on demand. Test techniques can be used in optimizing or minimizing number of tests. Test automation can be used to speed up testing.
DevOps and Testing slides at DASA ConnectKari Kakkonen
My and Rik Marselis slides at 30.5.2024 DASA Connect conference. We discuss about what is testing, then what is agile testing and finally what is Testing in DevOps. Finally we had lovely workshop with the participants trying to find out different ways to think about quality and testing in different parts of the DevOps infinity loop.
GraphSummit Singapore | The Art of the Possible with Graph - Q2 2024Neo4j
Neha Bajwa, Vice President of Product Marketing, Neo4j
Join us as we explore breakthrough innovations enabled by interconnected data and AI. Discover firsthand how organizations use relationships in data to uncover contextual insights and solve our most pressing challenges – from optimizing supply chains, detecting fraud, and improving customer experiences to accelerating drug discoveries.
Maruthi Prithivirajan, Head of ASEAN & IN Solution Architecture, Neo4j
Get an inside look at the latest Neo4j innovations that enable relationship-driven intelligence at scale. Learn more about the newest cloud integrations and product enhancements that make Neo4j an essential choice for developers building apps with interconnected data and generative AI.
Observability Concepts EVERY Developer Should Know -- DeveloperWeek Europe.pdfPaige Cruz
Monitoring and observability aren’t traditionally found in software curriculums and many of us cobble this knowledge together from whatever vendor or ecosystem we were first introduced to and whatever is a part of your current company’s observability stack.
While the dev and ops silo continues to crumble….many organizations still relegate monitoring & observability as the purview of ops, infra and SRE teams. This is a mistake - achieving a highly observable system requires collaboration up and down the stack.
I, a former op, would like to extend an invitation to all application developers to join the observability party will share these foundational concepts to build on:
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...James Anderson
Effective Application Security in Software Delivery lifecycle using Deployment Firewall and DBOM
The modern software delivery process (or the CI/CD process) includes many tools, distributed teams, open-source code, and cloud platforms. Constant focus on speed to release software to market, along with the traditional slow and manual security checks has caused gaps in continuous security as an important piece in the software supply chain. Today organizations feel more susceptible to external and internal cyber threats due to the vast attack surface in their applications supply chain and the lack of end-to-end governance and risk management.
The software team must secure its software delivery process to avoid vulnerability and security breaches. This needs to be achieved with existing tool chains and without extensive rework of the delivery processes. This talk will present strategies and techniques for providing visibility into the true risk of the existing vulnerabilities, preventing the introduction of security issues in the software, resolving vulnerabilities in production environments quickly, and capturing the deployment bill of materials (DBOM).
Speakers:
Bob Boule
Robert Boule is a technology enthusiast with PASSION for technology and making things work along with a knack for helping others understand how things work. He comes with around 20 years of solution engineering experience in application security, software continuous delivery, and SaaS platforms. He is known for his dynamic presentations in CI/CD and application security integrated in software delivery lifecycle.
Gopinath Rebala
Gopinath Rebala is the CTO of OpsMx, where he has overall responsibility for the machine learning and data processing architectures for Secure Software Delivery. Gopi also has a strong connection with our customers, leading design and architecture for strategic implementations. Gopi is a frequent speaker and well-known leader in continuous delivery and integrating security into software delivery.
Generative AI Deep Dive: Advancing from Proof of Concept to ProductionAggregage
Join Maher Hanafi, VP of Engineering at Betterworks, in this new session where he'll share a practical framework to transform Gen AI prototypes into impactful products! He'll delve into the complexities of data collection and management, model selection and optimization, and ensuring security, scalability, and responsible use.
UiPath Test Automation using UiPath Test Suite series, part 6DianaGray10
Welcome to UiPath Test Automation using UiPath Test Suite series part 6. In this session, we will cover Test Automation with generative AI and Open AI.
UiPath Test Automation with generative AI and Open AI webinar offers an in-depth exploration of leveraging cutting-edge technologies for test automation within the UiPath platform. Attendees will delve into the integration of generative AI, a test automation solution, with Open AI advanced natural language processing capabilities.
Throughout the session, participants will discover how this synergy empowers testers to automate repetitive tasks, enhance testing accuracy, and expedite the software testing life cycle. Topics covered include the seamless integration process, practical use cases, and the benefits of harnessing AI-driven automation for UiPath testing initiatives. By attending this webinar, testers, and automation professionals can gain valuable insights into harnessing the power of AI to optimize their test automation workflows within the UiPath ecosystem, ultimately driving efficiency and quality in software development processes.
What will you get from this session?
1. Insights into integrating generative AI.
2. Understanding how this integration enhances test automation within the UiPath platform
3. Practical demonstrations
4. Exploration of real-world use cases illustrating the benefits of AI-driven test automation for UiPath
Topics covered:
What is generative AI
Test Automation with generative AI and Open AI.
UiPath integration with generative AI
Speaker:
Deepak Rai, Automation Practice Lead, Boundaryless Group and UiPath MVP
Sudheer Mechineni, Head of Application Frameworks, Standard Chartered Bank
Discover how Standard Chartered Bank harnessed the power of Neo4j to transform complex data access challenges into a dynamic, scalable graph database solution. This keynote will cover their journey from initial adoption to deploying a fully automated, enterprise-grade causal cluster, highlighting key strategies for modelling organisational changes and ensuring robust disaster recovery. Learn how these innovations have not only enhanced Standard Chartered Bank’s data infrastructure but also positioned them as pioneers in the banking sector’s adoption of graph technology.
Implementation Of Back-Propagation Neural Network For Isolated Bangla Speech Recognition
1. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
DOI : 10.5121/ijist.2013.3401 1
Implementation Of Back-Propagation Neural
Network For Isolated Bangla Speech Recognition
Md. Ali Hossain1
, Md. Mijanur Rahman2
, Uzzal Kumar Prodhan3
, Md.
Farukuzzaman Khan4
1
Dept. of Computer Science & Engineering, Bangladesh University, Bangladesh.
ali.cse.bd@gmail.com
2,3
Dept. of Computer Science & Engineering, Jatiya Kabi Kazi Nazrul Islam University, Bangladesh.
mijan_cse@yahoo.com2
and uzzal_bagerhat@yahoo.com3
4
Dept. of Computer Science & Engineering, Islamic University, Bangladesh.
mfkhanbd2@gmail.com
ABSTRACT
This paper is concerned with the development of Back-propagation Neural Network for Bangla Speech
Recognition. In this paper, ten bangla digits were recorded from ten speakers and have been recognized.
The features of these speech digits were extracted by the method of Mel Frequency Cepstral Coefficient
(MFCC) analysis. The mfcc features of five speakers were used to train the network with Back propagation
algorithm. The mfcc features of ten bangla digit speeches, from 0 to 9, of another five speakers were used
to test the system. All the methods and algorithms used in this research were implemented using the
features of Turbo C and C++ languages. From our investigation it is seen that the developed system can
successfully encode and analyze the mfcc features of the speech signal to recognition. The developed
system achieved recognition rate about 96.332% for known speakers (i.e., speaker dependent) and 92% for
unknown speakers (i.e., speaker independent).
Keywords
Back-propagation,Feedforward Neural Networks, MFCC, Perceptrons, Speech Recognition.
1. Introduction
To communicate with each other, Speech is probably the most efficient way. It is possible to use
speech as a useful interface to interact with machines[1]. Speech recognition research work has
been started since 1930. However Bangla speech recognition research work has been started
since around 2000 [2]. Besides English language there is a lot of research experiment and
achieved result in various languages throughout the world. But in Bangla language, early
researchers in this field have limited success in phonemes [3,4], letters [5], words [6] or small
vocabulary continuous speech[7] for single speaker. In our system, we have captured speech from
ten different speakers, which may an early attempt for developing speaker independent isolated
Bangla digit speech recognition system in Bangla language. With a rich heritage, Bangla is an
important language . It is spoken by approximately 8% of the world population [8]. But, for the
computerization of this language, a systematic and scientific effort has not started yet. To support
rapidly developing computerization of Bangla Language, one of the most important issues is
Bangla speech recognition. Accordingly, we have developed a Bangla speech recognition system.
A Neural Network is an information Processing Paradigm and it is stimulated by the way
biological nervous systems, like the brain process Information[9]. Simple computational elements
operating in parallel are included in Neural networks [1]. The network function is determined
largely by the connections between elements. A neural network can be trained so that a particular
2. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
2
input guides to a specific target output [1]. Neural network can be used in different sector. There
are many applications of NNs limited only by our thoughts. Innovation is a solution to success, so
dude use NNs to generate something which will modernize the world! For the sake of writing, a
few Neural Networks applications are in Speech Recognition, Optical Character Recognition
(OCR), Modelling human behaviour, Classification of patterns, Loan risk analysis, music
generation, Image analysis, Creating new art forms, Stock market prediction, etc. In our research
work, Multilayer Feed-forward Network with Back-propagation algorithm is used to recognize
isolated Bangla speech digits from 0 to 9. All the methods and algorithms discussed in this paper
were implemented using the features of C and C++ languages.
2. Feed forward Networks
The simplest type of feed-forward network that use supervised learning is Perceptron. Binary
threshold units arranged into layers made a perceptron shown in Figure-1[10]. It is trained by the
Delta Rule or variations thereof. The Delta Rule can be applied directly for the case of a single
layer perceptron, as shown in Figure-1(a). For being perceptron’s activations are binary, reduces
this general learning rule to the Perceptron Learning Rule. This rules says, if an input is yi =1
(Active) and the output yj is incorrect , then depending on the desired output , the weight wji need
to be either decreased(if desired output is 0) or increased(if desired output is 1) by a small
amount ε [10]. To find a set of weights to accurately organize the patterns in any training set, This
procedure is insured, if the patterns are linearly distinguishable ( i.e. by a straight line, they can be
split into two classes). However, most training sets are not linearly distinguishable (For instance,
the simple XOR function); for these cases we require multiple layers.
Figure-1(b) shows Multi-layer perceptrons (MLPs). It can theoretically learn any function, but
they are very complex to train. It is not possible to apply The Delta Rule directly to MLPs
because in the hidden layer, there are no targets. If an MLP don’t use discrete activation functions
(i.e., threshold functions) that is it uses continuous activation functions (i.e., sigmoids functions),
then it turn out to be possible to use partial derivatives and the chain rule .These rules is used to
derive the influence of any weight on any output activation, that indicates, to reduce the
network’s error, how to modify that weight [10]. This generalized Delta Rule is known as back-
propagation.
Any number of hidden layers can have in MLPs, although for many applications, a single
hidden layer is sufficient, and more hidden layers tend to make training slower, for this
reason the terrain in weight space becomes more complicated. There are many ways that
MLPs can also be architecturally constrained, such as by limiting the weights values, or
by limiting their geometrically local areas connectivity, or tying different weights
together.
Figure 1. Perceptrons. (a) Single layer perceptron; (b) multi-layer perceptron.
3. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
3
2.1 Generalized Delta Rule (Back-Propagation)
The generalized delta rule is a supervised learning algorithm [11]. It is used to train a multilayer
neural network that maps the relation between the target output and actual output. During the
training period, the input pattern is passed through the network with network connection weights
and biases of the activation or transfer functions. Initially, the values and biases are assigned by a
small random numbers. In this rule, repeatedly presenting input-output pairs and then modifying
weights; the modification of weight reduces the network error. The training of the network
proceeds as follows:
• First, the input pattern is assigned to input neurons; and send to hidden neurons with
weights and computed its activation; and then send this to the output neurons with
weights and computed its activation, which represent the network’s output response of
the input pattern.
• Second, the output responses are compared to the desired output and an error term is
computed.
• Third, the error information is used to update the network weights and biases. The
modification of weight is computed by the four different parameters: a learning rate, the
derivative of activation function, the error term, and the current activity at the input layer.
• Fourth, each hidden unit computes its error. This is done with an output unit’s error,
which is sending it backwards as a signal to a hidden unit.
• Fifth, after computing the hidden unit error, the weights of the input-to-hidden units are
updated using the same equation that was used in the output layer.
If there is more than one layer of hidden units, then this procedure can be repeated iteratively.
That is, each hidden unit error in one layer, as an error signal, can be propagated backwards to an
adjacent layer once the hidden unit weights have been modified. Then the next training pattern
can be presented to the input units, and the learning process occurs again [12].
2.2 Working with back-propagation Algorithm
The application of the generalized delta rule involves two phases. At the first phase, to compute
the output values yp o for each output unit , the input x is presented and propagated forward
through the network. Desired output value do is compared with this output, for each output unit
resulting in an error signal δp o. To calculate the appropriate weight changes, a backward pass is
involved by the second phase through the network during which the error signal is passed to each
unit in the network [13]. Figure-2 shows the flowchart of the training of the neural network with
Back-propagation algorithm.
Weight adjustments with sigmoid activation function:
The weight of a connection is adjusted by an amount proportional to the product of an error signal
δ, on the unit k receiving the input and the output of the unit j sending this signal along the
connection: ∆pw jk = γ ᵟ
P
h yP
h
• The error signal for an output unit is given by: ᵟ
ᵨ
o = (d
ᵨ
o- y
ᵨ
o) (s
ᵨ
o)
4. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
4
• If the activation function F, the 'sigmoid' function defined as:
)))) ====
• In this case the derivative is equal to
)))) ====
==== (- )
====
==== yP
(1- yP
).
• Such that the output unit error signal can be Written as:
ᵟ
P
o = (d
P
o- y
P
o) y
P
o(1- y
P
o).
• The error signal for a hidden unit is determined recursively in terms of error signals of the
units to which it directly connects and the weights of those connections. For the sigmoid
activation function:
ᵟ
P
h = h) ᵟ
P
o who= yP
h(1- yP
h) ᵟ
P
o who
Figure 2. Back-propagation training algorithm
5. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
5
3. IMPLEMENTATION
3.1 Speech Data Acquisition
In a sound proof laboratory environment, Bangla speech words recording was completed with the
help of close-talking microphone, sound recorder software and high quality sound card. To
represent a signal, Wave form is the most general way [14-16]. To make a sample database, The
10 Bangla digits originated from ten speakers were recorded as wav file. At a sampling rate of
8.00 KHz and coded in 8 bits PCM [17], The utterances were recorded. At first, we discard 58
bytes (file header) from the beginning of the wave file for extracting wave data. After then wave
data read as character. Required voiced data is extracted from the input speech ( may containing
unvoice, voice and silence signal) by the data extraction process. By using the start and end-point
detection technique [8,18], the voiced data is extracted from the speech file and stored in a text
file as integer data.
3.2 Feature Extraction
The most important part of all recognition systems is the feature extraction that translates the
speech signal to some digital form having meaningful features. Obviously, for any recognition
system, a good feature may produce a good result. Frame blocking, Preemphasis, Windowing
and the computation of Mel Frequency Cepstrum Coefficient (MFCC) are the some signal
processing steps included in Feature extraction process, as shown in Figure-3 [8,19]. Initially,
each speech word was segmented into frame (a set of samples). A frame represent typically 16 to
32 ms of speech. Then apply the further computation on these speech frames. In our research
work, Hamming window function is used for feature extraction, which used in speech recognition
as one of the most popular windows. To extract a set of features that represents Mel Frequency
Cepstrum Coefficients (MFCC) of the signal, The pre-processed and windowed speech signal is
then passed through some computational steps. The Discrete Fourier Transform (DFT),
computation of first two formant frequencies, Mel frequency warping, Discrete Cosine Transform
(DCT) and finally the computation of Mel Frequency Cepstrum Coefficient (MFCC) are included
in the computation steps of MFCC, as shown in Figure-3 [8, 19-21].
Figure 3. Computation of MFCC features.
3.3 Training and Testing the Network
In the past, Neural networks were used by many authors for speech recognition [22-25]. For our
implementation, Multilayer Feed forward Network (three layer neural network) has been created
using C & C++. For training the network, Back Propagation algorithm was used. The network
consists of an input layer of 250 neurons, one hidden layer of 16 neurons and an output layer
contains 10 neurons used to recognize 10 speech words. We used a set of 250 mfcc (Mel
Frequency Cepstrum coefficients) feature values as input pattern for the neural network. The
input values are in a range of -7 up to 1.5. In our design we put all these input values in a 'Input
Layer' variable matrix. Linear activation function elements are included in the output layer. As 10
output neurons, output matrix contains 10x10 unit matrix. Non-linear sigmoidal activation
function is included in the hidden layer. For training the network, at first Randomize weights and
biases are set using random function ranging values from -1 to 1. Last layer does not require
6. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
6
weights and first layer does not need biases. For each training pattern, network layers weights and
biases are updated using back-propagation algorithms to reach the target, with the following
equations.
For output layer:
Delta = (TargetO - ActualO) * ActualO * (1 - ActualO)
Weight = Weight + LearningRate * Delta * Input
For hidden layers:
Delta = ActualO * (1-ActualO) * Summation(Weight_from_current_to_next AND
Delta_of_next)
Weight = Weight + LearningRate * Delta * Input
Where, TargetO= Target Output (This outputs are set to recognize ten difference words )
ActualO= Actual Output (Calculated output for training phase)
Delta= Bias of Neuron
LearningRate = Learning Rate to Update network weight.
A set of 15 mfcc features pattern of each speech word are used as training pattern to reach its
target. The weights and biases of the network are updated until the network error reaches almost
zero. At testing phase, the trained network was simulated with unknown speech pattern. It was
observed that the trained network performs very well and more than ten words can be recognized
by using the developed system.
4. Results and Discussion
In this experiment, 300 samples of ten Bangla digits, from 0 to 9, (i.e., 30 samples for each digit)
were recorded from ten speakers and then extracted mfcc features from these Bangla speech
words. The MFCC features were used with 8-mfcc/frame. 150 mfcc features of speech words
were used for training the network and another 150 mfcc features were used for testing the
network in the recognition phase. The developed system achieved 96.332% recognition accuracy
for known speakers (i.e., speaker dependent) and 92% accuracy for unknown speakers (i.e.,
speaker independent). The detailed results are shown in Table-1.
In this paper, an attempt was made to develop Back-propagation Neural Network for Isolated
Bangla Speech Recognition system. This is a single word recognition system, where ten Bangla
digits are recognized at a time. All the test patterns were conducted with ten different Speakers
from different age. Where fives are trained speakers and fives are test speakers. Instruments
should have constant settings. The speaker was a 25-aged male person. Performance with some
speakers from different age group was also tested. It was observed that speaking habit or style of
speaker affects the performance of the system, so the speakers should be well trained. The
sources of errors also include speeds of utterance and loudness variation. Also, the characteristics
of microphone and other recording instruments and background noises affect the system
performance.
7. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
7
Table 1. Details Recognition Results
5. CONCLUSION
The developed system with Back-propagation Neural Network achieved the reasonable results for
isolated Bangla digit speech recognition. The system can be used to recognize ten Bangla digits at
a time and achieved 92% recognition accuracy for multiple speakers. With properly trained
speakers and noise free environment, the developed system will produce better recognition
results. The variability in various parameters, like speed, noise, and loudness will properly handle
in our future research. A well-organized system should be completely speaker independent. So
speakers of different ages and genders should be employed in the future researchers. In Future,
speech recognition for continuous Bangla speech with more powerful recognition tools, like the
Hidden Markov Model (HMM), Time Delay Neural Network (TDNN) and Gaussian Mixture
Model (GMM) will be established.
References
[1] Wouter Gevaert, Georgi Tsenov, Valeri Mladenov, “Neural Networks used for Speech
Recognition”, Journal of Automatic Control, University of Belgrade, VOL. 20:1-7, 2010.
[2] Md. Abul Hasnat Jabir Mowla Mumit Khan, “Isolated and Continuous Bangla Speech Recognition:
Implementation, Performance and application perspective”, Conference Papers (Centre for Research
on Bangla Language Processing) Department of Computer Science and Engineering, BRAC
University, Bangladesh, 2007.
[3] S. M. Jahangir Alam, “System Development for Bangla Phoneme Recognition”, M.Sc. Thesis, Dept.
of Computer Science & Engineering, Islamic University, Kushtia-7003, July-2004.
[4] Karim, A.H. M. Rezaul, “ Bangla Phoneme Recognition”, Graduation Thesis report, Shahjalal
University of Science & Technology, Sylhet, 2002.
[5] Abul Hasanat Md. Rezaul Karim, Md. Shahidur Rahman and Md. Zafar Iqbal, “Recognition of
Spoken letters in Bangla”, 5th ICCIT 2002, East West University, Dhaka, Bangladesh, 27-28
December 2002.
[6] Kaushik Roy, Dipankar Das and M. Ganjer Ali, “Development of the Speech Recognition System
using Artificial Neural Network”, 5th ICCIT 2002, East West University, Dhaka, Bangladesh, 27-28,
December 2002.
Bangla
Digit
No. of
samples
for
testing
Speaker Dependent (Known speaker’s
speech were used for testing)
Speaker Independent (Unknown speaker’s
speech were used for testing)
No. of Properly
Recognized Digit
Recognition Rate
(%)
No. of Properly
Recognized Digit
Recognition
Rate (%)
0 15 14 93.33 14 93.33
1 15 15 100 14 93.33
2 15 14 93.33 13 86.67
3 15 15 100 14 93.33
4 15 15 100 14 93.33
5 15 14 93.33 14 93.33
6 15 15 100 14 93.33
7 15 13 86.67 13 86.67
8 15 14 93.33 14 93.33
9 15 14 93.33 14 93.33
Total 150 143 96.332 138 92
8. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
8
[7] Md. Saidur Rahman, “Small Vocabulary Speech Recognition in Bangla Language”, M.Sc. Thesis,
Dept. of Computer Science & Engineering, Islamic University, Kushtia-7003, July-2004.
[8] Md. Mijanur Rahman, Fatema Khatun, and Dr. Md. Al-Amin Bhuiyan “Development of Isolated
Speech Recognition System for Bangla Words”, Vol. 1, Issue. 3, September - December, 2010
(IJARITAC), pp.272-278.
[9] Preet Ind, Gour Sundar Mitra Thakurer Singh, “Enhanced Password Based Security System Based
on User Behavior using Neural Networks” I.J. Information Engineering and Electronic Business,
Vol. 2, 29-35, 2012.
[10] Joe Tebelskis “Speech Recognition using Neural Networks” School of Computer Science Carnegie
Mellon University Pittsburgh, Pennsylvania 15213-3890, May 1995, CMU-CS-95-142
[11] Rumelhart, D. E., Hinton, G. E., and Williams, R. J., “Learning representations by back- propagating
Errors Nature”, 323, 533-536, 1986.
[12] http://www.bcp.psych.ualberta.ca/~mike/Pearl_Street/Dictionary/contents/G/generalizeddr.html
[13] R.Vijaya arjunan and Dr.V.Vijaya Kumar, “A Novel framework for Voice Signal recognition using
RASTA and Authentication by Neural Networks”, International Journal of Recent Trends in
Engineering, Vol 2, No. 3, November 2009.
[14] Rabiner L., Bing_Hwang J.,”Fundamentals of Speech Recognition”, Prentice_Hall, 1993.
[15] Jurafsky, Daniel and Martin, James H, “Speech and Language Processing: An Introduction to
Natural Language Processing, Computational Linguistics, and Speech Recognition”, Prentice Hall,
1st ed.1996.
[16] DSG Pollock, “A handbook of Time-series analysis, signal processing and dynamics”, Academic
press London, 1999.
[17] S. Gokul, “Multimedia Magic”, BPB Publications, B-14, Connaught Place, New Delhi-110001.
[18] Dr. Ramesh Chandra Debnath and Md. Farukuzzaman Khan, “Bangla Sentence Recognition Using
End- Point Detection”, Rajshahi University Studies: Part B, Journal of Science, Vol 32, 2004.
[19] Jean-Claude Junqua and Jean-Paul Haton, “Robustness in Automatic Speech Recognition:
Fundamentals and Applications”, Kluwer Academic Publishers, Dordrecht, Netherlands, 1997.
[20] F. Jelinek, L. R. Bahl, and R. L. Mercer, “Design of a Linguistic Statistical Decoder for the
Recognition of Continuous Speech”, IEEE Trans. Information Theory, IT-21, pp 250-256, 1975.
[21] Md. Farukuzzaman Khan and Dr. Ramesh Chandra Debnath, “Comparative Study of Feature
Extraction Methods for Bangla Phoneme Recognition”, 5th ICCIT 2002, East West University,
Dhaka, Bangladesh, PP.27-28, December 2002.
[22] Akram M. Othman, and May H. Riadh, “Speech Recognition Using Scaly Neural Networks”, World
Academy of Science, Engineering and Technology, vol. 38, 2008.
[23] Mohamad Adnan Al-Alaoui, Lina Al-Kanj, Jimmy Azar, and Elias Yaacoub, “Speech Recognition
using Artificial Neural Networks and Hidden Markov Models”, IEEE Multidisciplinary Engineering
Education Magazine, VOL. 3, 2008.
[24] Dou-Suk Kim and Soo-Young Lee, “Intelligent judge neural network for speech recognition”,
Neural Processing Letters, Vol.1, No.1, PP.17-20, 1994.
[25] Chee Peng Lim, Siew Chan Woo, Aun Sim Loh, Rohaizan Osman, “Speech Recognition Using
Artificial Neural Networks”, 1st International Conference on Web Information Systems Engineering
(WISE'00),Vol.1, 2000.
Biographies of Authors
Md. Ali Hossain
Mr. Hossain was born in Manikganj, Bangladesh. He received the B.Sc. and M.Sc.
degrees from the Department of Computer Science and Engineering, University of
Islamic University, Kushtia, Bangladesh, in 2008 and 2009, respectively. He is
serving as a Lecturer with the Department of Computer Science and Engineering
(CSE), Bangladesh University, Dhaka. He has got a number of research articles
published in different international journals. His current research interests include
speech processing, biomedical imaging, biomedical signal, bioinformatics, neural
networks and AI. Mr. Ali Hossain is an Associate Member of the Bangladesh
Computer Society and Executive Member of Islamic University Computer
Association (IUCA).
9. International Journal of Information Sciences and Techniques (IJIST) Vol.3, No.4, July 2013
9
Md. Mijanur Rahman
Mr. Rahman is working as an assistant professor of the department of Computer
Science and Engineering in Jatiya Kabi Kazi Nazrul Islam University, Trishal,
Mymensingh, Bangladesh. He completed his B Sc (Hons) and M Sc in CSE degree
from Islamic University, Kushtia, Bangladesh. At present he is continuing his PhD
research work in the department of Computer Science and Engineering, Jahangirnagar
University, Savar, Dhaka, Bangladesh. He has got a number of research articles
published in different local and international journals. His current research interests
include the fields of pattern recognition, image and speech processing, neural networks,
fuzzy logics and AI.
Uzzal Kumar Prodhan
Assistant Professor, Department of Computer Science & Engineering, Jatiya Kabi
Kazi Nazrul Islam University, Trishal, Mymensingh, Bangladesh. He has completed
his M.Sc. and B.Sc. from the department of Computer Science & Engineering, Islamic
University, Bangladesh. He got first class in both exams. He passed S.S.C. and H.S.C
with star marks. After completing M.Sc in CSE, he joined Bangladesh University as a
Lecturer & joined in Jatia Kabi Kazi Nazrul Islam University as an Assistant
Professor. In his long teaching life he was appointed as a head examiner in Computer
Technology by Bangladesh Technical Education Board, Dhaka. Due to his teaching
interest he was selected as a Book reviewer of National Curriculum of Textbook
Board, Dhaka. He has successfully completed Microsoft Certified IT Professional (MCITP) on Server 2008
platform. His research interest includes Artificial Intelligence, Neural Network, Cryptography, Computer
Architecture and Organization and Pattern Recognition. He has many international and national research
publications. His email addresses are uzzal_bagerhat@yahoo.com, uzzal.prodhan@bu.edu.bd.
Prof. Md. Farukuzzaman Khan
Prof. Khan is working as professor of the department of Computer Science and
Engineering in Islamic University, Kushtia, Bangladesh. He completed his B Sc
(Hons), M Sc and M. Phil degree from Rajshahi University, Rajshahi, Bangladesh. He
is a PhD researcher in the department of Computer Science and Engineering, Islamic
University, Kushtia, Bangladesh. He has has got a number of research articles
published in different local and international journals. His current research interests
pattern recognition, image and speech processing, DSP and AI.