Hadaate Ullah - Academia.edu (original) (raw)

Papers by Hadaate Ullah

2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML), 2021

Recent findings demonstrated that deep neural networks carry out features extraction itself to id... more Recent findings demonstrated that deep neural networks carry out features extraction itself to identify the electrocardiography (ECG) pattern or cardiac arrhythmias from the ECG signals directly and provided good results compared to cardiologists in some cases. But, to face the challenge of huge volume of data to train such networks, transfer learning is a prospective mechanism where network is trained on a large dataset and learned experiences are transferred to a small volume target dataset. Therefore, we firstly extracted 78,999 ECG beats from MIT-BIH arrhythmia dataset and transformed into 2D RGB images and used as the inputs of the DenseNet. The DenseNet is initialized with the trained weights on ImageNet and fine-tuned with the extracted beat images. Optimization of the pre-trained DenseNet is performed with the aids of on-the-fly augmentation, weighted random sampler, and Adam optimizer. The performance of the pre-trained model is assessed by hold-out evaluation and stratified 5-fold cross-validation techniques along with early stopping feature. The achieved accuracy of identifying normal and four arrhythmias are of 98.90% and 100% for the hold-out and stratified 5-fold respectively. The effectiveness of the pre-trained model with the stratified 5-fold by transfer learning approach is surpassed compared to the state-of-art-the approaches and models, and also explicit the maximum generalization of imbalanced classes.

2017 IEEE International Conference on Imaging, Vision & Pattern Recognition (icIVPR), 2017

In this paper, a new window function is proposed, which can be used to design an FIR filter. The ... more In this paper, a new window function is proposed, which can be used to design an FIR filter. The window is adjustable as by changing the value of a variable, the window can be adjusted accordingly. The proposed window is compared with hamming & Kaiser window. The result implies that, the proposed window has better side-lobe roll-off ratio (24.65 dB) than hamming (5.74 dB) & Kaiser (18.87 dB) window. Besides the ripple ratio of hamming & Kaiser window are −55.72 dB & −21.29 dB while the proposed window has −63.05 dB which is better than both windows. FIR filter, designed using the proposed window, has 24.78 dB side-lobe roll-off ratio, where hamming & Kaiser windows have 5.71 dB & 18.76 dB respectively.

Indian Journal of Science and Technology, 2016

Objective: Millions of people use wireless devices in their day to day diligences without knowing... more Objective: Millions of people use wireless devices in their day to day diligences without knowing the security facets of Wireless Technology. The aim of our research is to enhance the execution of widely used wireless devices's protocols by examining their behavior with Feed Forward Neural Network. Fundamentally, Neural Network is a multilayer perceptron network. It processes the records one at a time and "learn" by comparing the obtained output with the actual output. Hidden layer neurons play a cardinal role in the performance of Back Propagation. The process of determining the number of hidden layer neurons is still obscure. The work is focused on performance evaluation of the hidden layer neurons for WEP (Wired Equivalent Privacy) and WPA (Wi-Fi Protected Access) protocols. Methods/ Statistical Analysis: For this work, three network architectures have been picked out to perform the analysis. The research work is carried out by using Back Propagation Algorithm in Neural Network Toolbox on the data captured by using Wireshark tool. Findings: The behavior of various unlike hidden neurons is evaluated through simulation technique. Network performance is also diagnosed with the help of epochs and Mean Square Error (MSE). The performance of Neural Network is evaluated and outcomes indicate that hidden layer neurons affect the functioning of the network. Improvement: We would like to work with the parameter and learning of the Neural Network to achieve best results.

Journal of Science and Technology, 2017

Electric drives have numerous applications in diverse areas such as rolling mills, electric train... more Electric drives have numerous applications in diverse areas such as rolling mills, electric trains and robotic manipulators. Inefficient control of motor speed can destroy the equipment itself; even can cause a severe accident. In this project, a low-cost MOSFET based chopper drive DC motor speed control system was designed and implemented. In this system, the feedback voltage from the load controls the speed of the DC motor through speed controller. From the measurement, it is clear that for a maximum rpm of 210, the percentage of error varied from 1.10 to 14.28 for 12 V supply voltage. The total cost of the speed control scheme is only 18 USD. Such a drive will be appropriate for the speed control of DC motor in household to industrial appliances.

Carpathian Journal of Electronic and Computer Engineering, 2018

The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of... more The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of the transceiver greatly depends on the performance of the bandpass filter. A bandpass filter designed with passive inductors suffers from some drawbacks like large chip size, low-quality factor, less tenability etc. To prevail over these constraints, an active inductor-based bandpass filter circuit has been designed in GPDK-90nm CMOS technology utilizing cadence virtuoso environment. The simulation result shows that the active inductor-based bandpass filter circuit design achieves a gain of 6.79dB, a bandwidth of 5.05 GHz and a noise figure of 3.10dB. The circuit dissipates only 3.55mW power for its operation from a single 1.5V DC supply. By avoiding bulky inductor in the design helped to attain a very small chip area of 127.704μm2.

The main purpose of this project is to design an inverter that will enable the inversion of a DC ... more The main purpose of this project is to design an inverter that will enable the inversion of a DC power source, supplied by Photovoltaic (PV) Cells, to an AC power source that will be either used to supply a load or connected directly to the utility grid. The benefit of this project is to give access to an everlasting and pollution free source of energy. The future is looking towards alternative power sources all of which will need to be regulated in one form or another. To make this possible, a highly efficient low cost product will have to be ...

TELKOMNIKA Indonesian Journal of Electrical Engineering, 2015

Indium-tin oxide (ITO) which is optically transparent is referred as a “universal” electrode for ... more Indium-tin oxide (ITO) which is optically transparent is referred as a “universal” electrode for various optoelectronic devices such as organic light emitting diodes (OLEDs). It is scientifically proved that the performance of OLEDs raises up significantly by exposing the ITO surface to oxygen plasma. This study employs conducting atomic force microscopy (C-AFM) for unique nanometer-scale mapping of the local current density of a vapor-deposited ITO film. Indium Tin Oxide (ITO) thin films have been prepared by using the reactive evaporation method on glass substrates in an oxygen atmosphere. It is found that the deposition rate plays a vital role in controlling the electrical properties of the ITO thin films. The resistivity and the electrical conductivity were also investigated. The electrical resistivity of 3.10 x10 –6 Ωm has been obtained with a deposition rate of 2 nm/min.

Indonesian Journal of Electrical Engineering and Computer Science, 2017

Cadmium (Cd) is a soft, silver-white or blue lustrous metal typically found in mineral deposits w... more Cadmium (Cd) is a soft, silver-white or blue lustrous metal typically found in mineral deposits with lead, zinc and copper. Cadmium Oxide thin films have been prepared on a glass substrate at 350 0 C temperature by implementing the Spray Pyrolysis method. The direct and indirect band gap energies are determined using spectral data. The direct and indirect band gap energies decrease with the increasing film thickness. It is noted that for the same film thickness the direct band gap energy is greater than indirect band gap energy. The transmittance increases with the increasing wavelength for annealed and deposited films. It is also noted that for the same wavelength the transmittance for deposited films is greater than the transmittance for annealed films.

Journal of Healthcare Engineering

Recently, cardiac arrhythmia recognition from electrocardiography (ECG) with deep learning approa... more Recently, cardiac arrhythmia recognition from electrocardiography (ECG) with deep learning approaches is becoming popular in clinical diagnosis systems due to its good prognosis findings, where expert data preprocessing and feature engineering are not usually required. But a lightweight and effective deep model is highly demanded to face the challenges of deploying the model in real-life applications and diagnosis accurately. In this work, two effective and lightweight deep learning models named Deep-SR and Deep-NSR are proposed to recognize ECG beats, which are based on two-dimensional convolution neural networks (2D CNNs) while using different structural regularizations. First, 97720 ECG beats extracted from all records of a benchmark MIT-BIH arrhythmia dataset have been transformed into 2D RGB (red, green, and blue) images that act as the inputs to the proposed 2D CNN models. Then, the optimization of the proposed models is performed through the proper initialization of model lay...

Indonesian Journal of Electrical Engineering and Computer Science, 2016

In the case of medical science, one of the most restless researches is the identification of abno... more In the case of medical science, one of the most restless researches is the identification of abnormalities in brain. Electroencephalogram (EEG) is the main tool for determining the electrical activity of brain and it contains rich information associated to the varieties physiological states of brain. The purpose of this task is to identify the EEG signal as order or disorder. It is proposed to enrich an automated system for the identification of brain disorders. An EEG signal of a patient has been taken as a sample. The simulation has been done by MATLAB. The file which consists of the signal has been called in and plotted the signals in MATLAB. The proposed system covers pre-processing, feature extraction, feature selection and classification. By the pre-processing the noises are ejected. In this case the signal has been filtered using band pass filter. The Discrete Wavelet Transform (DWT) has been used to decompose the EEG signal into Sub-band signal. The feature extraction method...

Journal of theoretical and applied information technology, 2015

Muscle fatigue is often caused by unhealthy and irregular work practice. It is defined as a long ... more Muscle fatigue is often caused by unhealthy and irregular work practice. It is defined as a long lasting reduction of the ability to contract and it is the condition when produced force is reduced. Faster walking can cause muscle fatigue, which is unhealthy when the level of fatigue is high. There are many mathematical parameters that are suitable to assess the muscle fatigue during gait. Out of these parameters, the amplitude and frequency of the surface EMG signal (sEMG) reflects the more accurate physiological activity in the motor unit during contraction and at rest. In this research, Empirical mode decomposition (EMD) based filtering process is applied on sEMG signal for realizing the fatiguing contraction during subject walking exercise. The purpose of this research is to evaluate the surface electromyographic parameters (RMS, IAV and AIF) for addressing the effectiveness of the EMD method. In this study, RMS, IAV and AIF values were used as spectral variable, which extensivel...

2020 IEEE 2nd International Conference on Artificial Intelligence in Engineering and Technology (IICAIET), 2020

Radio frequency identification (RFID) technology is currently reader protocol specific. For a com... more Radio frequency identification (RFID) technology is currently reader protocol specific. For a common RFID standard to be used as the internet of things (IoT) devices, the reader should either be general or be avoided to facilitate tag communication with a common protocol. A complementery metal oxide semiconductor (CMOS) low noise amplifier (LNA) is recommended for 2.4 GHz IoT RFID. Spiral inductor based LNA cannot overcome the problems of bulky die area, lesser Q factor, limited tuning flexibility etc. Therefore, an LNA with an inductor less approach is designed in 90nm CMOS process cadence software. The post-layout simulation exhibits a 19 dB gain, a 164.2 MHz bandwidth and a 1.55 dB noise figure at 2.4 GHz. The LNA consumes very low power which is only 1.08 mW from a 1.5 V supply. A very compact layout of 127.7 µm2 has been achieved because of the inductor less approach.

Classification is one of the most hourly encountered problems in real world. Neural networks have... more Classification is one of the most hourly encountered problems in real world. Neural networks have emerged as one of the tools that can handle the classification problem. Feed-Forward Neural Networks (FFNN's) have been widely applied in many different fields as a classification tool. Designing an efficient FFNN structure with the optimum number of hidden layers and minimum number of layer's neurons for a given specific application or dataset, is an open research problem and more challenging depend on the input data. The random selections of hidden layers and neurons may cause the problem of either under fitting or over fitting. Over fitting arises because the network matches the data so closely as to lose its generalization ability over the test data. In this research, the classification performance using the Mean Square Error (MSE) of Feed-Forward Neural Network (FFNN) with back-propagation algorithm with respect to the different number of hidden layers and hidden neurons is...

2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML)

Recent findings demonstrated that deep neural networks carry out features extraction itself to id... more Recent findings demonstrated that deep neural networks carry out features extraction itself to identify the electrocardiography (ECG) pattern or cardiac arrhythmias from the ECG signals directly and provided good results compared to cardiologists in some cases. But, to face the challenge of huge volume of data to train such networks, transfer learning is a prospective mechanism where network is trained on a large dataset and learned experiences are transferred to a small volume target dataset. Therefore, we firstly extracted 78,999 ECG beats from MIT-BIH arrhythmia dataset and transformed into 2D RGB images and used as the inputs of the DenseNet. The DenseNet is initialized with the trained weights on ImageNet and fine-tuned with the extracted beat images. Optimization of the pre-trained DenseNet is performed with the aids of on-the-fly augmentation, weighted random sampler, and Adam optimizer. The performance of the pre-trained model is assessed by hold-out evaluation and stratified 5-fold cross-validation techniques along with early stopping feature. The achieved accuracy of identifying normal and four arrhythmias are of 98.90% and 100% for the hold-out and stratified 5-fold respectively. The effectiveness of the pre-trained model with the stratified 5-fold by transfer learning approach is surpassed compared to the state-of-art-the approaches and models, and also explicit the maximum generalization of imbalanced classes.

TELKOMNIKA Indonesian Journal of Electrical Engineering, 2015

Malaysian Journal of Applied Sciences, 2018

Classification is one of the most hourly encountered problems in real world. Neural networks have... more Classification is one of the most hourly encountered problems in real world. Neural networks have emerged as one of the tools that can handle the classification problem. Feed-Forward Neural Networks (FFNN's) have been widely applied in many different fields as a classification tool. Designing an efficient FFNN structure with the optimum number of hidden layers and minimum number of layer's neurons for a given specific application or dataset, is an open research problem and more challenging depend on the input data. The random selections of hidden layers and neurons may cause the problem of either under fitting or over fitting. Over fitting arises because the network matches the data so closely as to lose its generalization ability over the test data. In this research, the classification performance using the Mean Square Error (MSE) of Feed-Forward Neural Network (FFNN) with back-propagation algorithm with respect to the different number of hidden layers and hidden neurons is computed and analyzed to find out the optimum number of hidden layers and minimum number of layer's neurons to help the existing classification concepts by MATLAB version 13a. By this process, firstly the random data has been generated using an suitable matlab function to prepare the training data as the input and target vectors as the testing data for the classification purposes of FFNN. The generated input data is passed on to the output layer through the hidden layers which process these data. From this analysis, it is find out from the mean square error comparison graphs and regression plots that for getting the best performance form this network, it is better to use the high number of hidden layers and more neurons in the hidden layers in the network during designing its classifier but so more neurons in the hidden layers and the high number of hidden layers in the network makes it complex and takes more time to execute. So as the result it is suggested that three hidden layers and 26 hidden neurons in each hidden layers are better for designing the classifier of this network for this type of input data features.

Carpathian Journal of Electronic and Computer Engineering, 2018

The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of... more The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of the transceiver greatly depends on the performance of the bandpass filter. A bandpass filter designed with passive inductors suffers from some drawbacks like large chip size, low-quality factor, less tenability etc. To prevail over these constraints, an active inductor-based bandpass filter circuit has been designed in GPDK-90nm CMOS technology utilizing cadence virtuoso environment. The simulation result shows that the active inductor-based bandpass filter circuit design achieves a gain of 6.79dB, a bandwidth of 5.05 GHz and a noise figure of 3.10dB. The circuit dissipates only 3.55mW power for its operation from a single 1.5V DC supply. By avoiding bulky inductor in the design helped to attain a very small chip area of 127.704μm 2 .

Artificial Neural Networks (ANNs) are one of the most comprehensive tools for classification. I... more Artificial Neural Networks (ANNs) are one of the most comprehensive tools for classification. In this study, the performance of Feed-Forward Neural Network (FFNN) with back-propagation algorithm is used to find out the appropriate activation function in the hidden layer using MATLAB 2013a. Random data has been generated and fetched to FFNN for testing the classification performance of this network. From the values of MSE, response graph and regression coefficients, it is clear that Tan sigmoid activation function is the best choice for the image classification. The FFNN with this activation function is better for any classification purpose of different applications such as aerospace, automotive, materials, manufacturing, petroleum, robotics, communication etc because to perform the classification the network designer have to choose an activation function.

Now-a-days online interviewing is preferred as it does reduce efforts & it is cost effective as w... more Now-a-days online interviewing is preferred as it does
reduce efforts & it is cost effective as well. Long distance is one of
the causes for drawing attention in online interview. In this paper
we have described a technique of the generation of automated data
(CV) to assist the interviewer having more precise knowledge on
the person talking to online.

2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML), 2021

Recent findings demonstrated that deep neural networks carry out features extraction itself to id... more Recent findings demonstrated that deep neural networks carry out features extraction itself to identify the electrocardiography (ECG) pattern or cardiac arrhythmias from the ECG signals directly and provided good results compared to cardiologists in some cases. But, to face the challenge of huge volume of data to train such networks, transfer learning is a prospective mechanism where network is trained on a large dataset and learned experiences are transferred to a small volume target dataset. Therefore, we firstly extracted 78,999 ECG beats from MIT-BIH arrhythmia dataset and transformed into 2D RGB images and used as the inputs of the DenseNet. The DenseNet is initialized with the trained weights on ImageNet and fine-tuned with the extracted beat images. Optimization of the pre-trained DenseNet is performed with the aids of on-the-fly augmentation, weighted random sampler, and Adam optimizer. The performance of the pre-trained model is assessed by hold-out evaluation and stratified 5-fold cross-validation techniques along with early stopping feature. The achieved accuracy of identifying normal and four arrhythmias are of 98.90% and 100% for the hold-out and stratified 5-fold respectively. The effectiveness of the pre-trained model with the stratified 5-fold by transfer learning approach is surpassed compared to the state-of-art-the approaches and models, and also explicit the maximum generalization of imbalanced classes.

2017 IEEE International Conference on Imaging, Vision & Pattern Recognition (icIVPR), 2017

In this paper, a new window function is proposed, which can be used to design an FIR filter. The ... more In this paper, a new window function is proposed, which can be used to design an FIR filter. The window is adjustable as by changing the value of a variable, the window can be adjusted accordingly. The proposed window is compared with hamming & Kaiser window. The result implies that, the proposed window has better side-lobe roll-off ratio (24.65 dB) than hamming (5.74 dB) & Kaiser (18.87 dB) window. Besides the ripple ratio of hamming & Kaiser window are −55.72 dB & −21.29 dB while the proposed window has −63.05 dB which is better than both windows. FIR filter, designed using the proposed window, has 24.78 dB side-lobe roll-off ratio, where hamming & Kaiser windows have 5.71 dB & 18.76 dB respectively.

Indian Journal of Science and Technology, 2016

Objective: Millions of people use wireless devices in their day to day diligences without knowing... more Objective: Millions of people use wireless devices in their day to day diligences without knowing the security facets of Wireless Technology. The aim of our research is to enhance the execution of widely used wireless devices's protocols by examining their behavior with Feed Forward Neural Network. Fundamentally, Neural Network is a multilayer perceptron network. It processes the records one at a time and "learn" by comparing the obtained output with the actual output. Hidden layer neurons play a cardinal role in the performance of Back Propagation. The process of determining the number of hidden layer neurons is still obscure. The work is focused on performance evaluation of the hidden layer neurons for WEP (Wired Equivalent Privacy) and WPA (Wi-Fi Protected Access) protocols. Methods/ Statistical Analysis: For this work, three network architectures have been picked out to perform the analysis. The research work is carried out by using Back Propagation Algorithm in Neural Network Toolbox on the data captured by using Wireshark tool. Findings: The behavior of various unlike hidden neurons is evaluated through simulation technique. Network performance is also diagnosed with the help of epochs and Mean Square Error (MSE). The performance of Neural Network is evaluated and outcomes indicate that hidden layer neurons affect the functioning of the network. Improvement: We would like to work with the parameter and learning of the Neural Network to achieve best results.

Journal of Science and Technology, 2017

Electric drives have numerous applications in diverse areas such as rolling mills, electric train... more Electric drives have numerous applications in diverse areas such as rolling mills, electric trains and robotic manipulators. Inefficient control of motor speed can destroy the equipment itself; even can cause a severe accident. In this project, a low-cost MOSFET based chopper drive DC motor speed control system was designed and implemented. In this system, the feedback voltage from the load controls the speed of the DC motor through speed controller. From the measurement, it is clear that for a maximum rpm of 210, the percentage of error varied from 1.10 to 14.28 for 12 V supply voltage. The total cost of the speed control scheme is only 18 USD. Such a drive will be appropriate for the speed control of DC motor in household to industrial appliances.

Carpathian Journal of Electronic and Computer Engineering, 2018

The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of... more The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of the transceiver greatly depends on the performance of the bandpass filter. A bandpass filter designed with passive inductors suffers from some drawbacks like large chip size, low-quality factor, less tenability etc. To prevail over these constraints, an active inductor-based bandpass filter circuit has been designed in GPDK-90nm CMOS technology utilizing cadence virtuoso environment. The simulation result shows that the active inductor-based bandpass filter circuit design achieves a gain of 6.79dB, a bandwidth of 5.05 GHz and a noise figure of 3.10dB. The circuit dissipates only 3.55mW power for its operation from a single 1.5V DC supply. By avoiding bulky inductor in the design helped to attain a very small chip area of 127.704μm2.

The main purpose of this project is to design an inverter that will enable the inversion of a DC ... more The main purpose of this project is to design an inverter that will enable the inversion of a DC power source, supplied by Photovoltaic (PV) Cells, to an AC power source that will be either used to supply a load or connected directly to the utility grid. The benefit of this project is to give access to an everlasting and pollution free source of energy. The future is looking towards alternative power sources all of which will need to be regulated in one form or another. To make this possible, a highly efficient low cost product will have to be ...

TELKOMNIKA Indonesian Journal of Electrical Engineering, 2015

Indium-tin oxide (ITO) which is optically transparent is referred as a “universal” electrode for ... more Indium-tin oxide (ITO) which is optically transparent is referred as a “universal” electrode for various optoelectronic devices such as organic light emitting diodes (OLEDs). It is scientifically proved that the performance of OLEDs raises up significantly by exposing the ITO surface to oxygen plasma. This study employs conducting atomic force microscopy (C-AFM) for unique nanometer-scale mapping of the local current density of a vapor-deposited ITO film. Indium Tin Oxide (ITO) thin films have been prepared by using the reactive evaporation method on glass substrates in an oxygen atmosphere. It is found that the deposition rate plays a vital role in controlling the electrical properties of the ITO thin films. The resistivity and the electrical conductivity were also investigated. The electrical resistivity of 3.10 x10 –6 Ωm has been obtained with a deposition rate of 2 nm/min.

Indonesian Journal of Electrical Engineering and Computer Science, 2017

Cadmium (Cd) is a soft, silver-white or blue lustrous metal typically found in mineral deposits w... more Cadmium (Cd) is a soft, silver-white or blue lustrous metal typically found in mineral deposits with lead, zinc and copper. Cadmium Oxide thin films have been prepared on a glass substrate at 350 0 C temperature by implementing the Spray Pyrolysis method. The direct and indirect band gap energies are determined using spectral data. The direct and indirect band gap energies decrease with the increasing film thickness. It is noted that for the same film thickness the direct band gap energy is greater than indirect band gap energy. The transmittance increases with the increasing wavelength for annealed and deposited films. It is also noted that for the same wavelength the transmittance for deposited films is greater than the transmittance for annealed films.

Journal of Healthcare Engineering

Recently, cardiac arrhythmia recognition from electrocardiography (ECG) with deep learning approa... more Recently, cardiac arrhythmia recognition from electrocardiography (ECG) with deep learning approaches is becoming popular in clinical diagnosis systems due to its good prognosis findings, where expert data preprocessing and feature engineering are not usually required. But a lightweight and effective deep model is highly demanded to face the challenges of deploying the model in real-life applications and diagnosis accurately. In this work, two effective and lightweight deep learning models named Deep-SR and Deep-NSR are proposed to recognize ECG beats, which are based on two-dimensional convolution neural networks (2D CNNs) while using different structural regularizations. First, 97720 ECG beats extracted from all records of a benchmark MIT-BIH arrhythmia dataset have been transformed into 2D RGB (red, green, and blue) images that act as the inputs to the proposed 2D CNN models. Then, the optimization of the proposed models is performed through the proper initialization of model lay...

Indonesian Journal of Electrical Engineering and Computer Science, 2016

In the case of medical science, one of the most restless researches is the identification of abno... more In the case of medical science, one of the most restless researches is the identification of abnormalities in brain. Electroencephalogram (EEG) is the main tool for determining the electrical activity of brain and it contains rich information associated to the varieties physiological states of brain. The purpose of this task is to identify the EEG signal as order or disorder. It is proposed to enrich an automated system for the identification of brain disorders. An EEG signal of a patient has been taken as a sample. The simulation has been done by MATLAB. The file which consists of the signal has been called in and plotted the signals in MATLAB. The proposed system covers pre-processing, feature extraction, feature selection and classification. By the pre-processing the noises are ejected. In this case the signal has been filtered using band pass filter. The Discrete Wavelet Transform (DWT) has been used to decompose the EEG signal into Sub-band signal. The feature extraction method...

Journal of theoretical and applied information technology, 2015

Muscle fatigue is often caused by unhealthy and irregular work practice. It is defined as a long ... more Muscle fatigue is often caused by unhealthy and irregular work practice. It is defined as a long lasting reduction of the ability to contract and it is the condition when produced force is reduced. Faster walking can cause muscle fatigue, which is unhealthy when the level of fatigue is high. There are many mathematical parameters that are suitable to assess the muscle fatigue during gait. Out of these parameters, the amplitude and frequency of the surface EMG signal (sEMG) reflects the more accurate physiological activity in the motor unit during contraction and at rest. In this research, Empirical mode decomposition (EMD) based filtering process is applied on sEMG signal for realizing the fatiguing contraction during subject walking exercise. The purpose of this research is to evaluate the surface electromyographic parameters (RMS, IAV and AIF) for addressing the effectiveness of the EMD method. In this study, RMS, IAV and AIF values were used as spectral variable, which extensivel...

2020 IEEE 2nd International Conference on Artificial Intelligence in Engineering and Technology (IICAIET), 2020

Radio frequency identification (RFID) technology is currently reader protocol specific. For a com... more Radio frequency identification (RFID) technology is currently reader protocol specific. For a common RFID standard to be used as the internet of things (IoT) devices, the reader should either be general or be avoided to facilitate tag communication with a common protocol. A complementery metal oxide semiconductor (CMOS) low noise amplifier (LNA) is recommended for 2.4 GHz IoT RFID. Spiral inductor based LNA cannot overcome the problems of bulky die area, lesser Q factor, limited tuning flexibility etc. Therefore, an LNA with an inductor less approach is designed in 90nm CMOS process cadence software. The post-layout simulation exhibits a 19 dB gain, a 164.2 MHz bandwidth and a 1.55 dB noise figure at 2.4 GHz. The LNA consumes very low power which is only 1.08 mW from a 1.5 V supply. A very compact layout of 127.7 µm2 has been achieved because of the inductor less approach.

Classification is one of the most hourly encountered problems in real world. Neural networks have... more Classification is one of the most hourly encountered problems in real world. Neural networks have emerged as one of the tools that can handle the classification problem. Feed-Forward Neural Networks (FFNN's) have been widely applied in many different fields as a classification tool. Designing an efficient FFNN structure with the optimum number of hidden layers and minimum number of layer's neurons for a given specific application or dataset, is an open research problem and more challenging depend on the input data. The random selections of hidden layers and neurons may cause the problem of either under fitting or over fitting. Over fitting arises because the network matches the data so closely as to lose its generalization ability over the test data. In this research, the classification performance using the Mean Square Error (MSE) of Feed-Forward Neural Network (FFNN) with back-propagation algorithm with respect to the different number of hidden layers and hidden neurons is...

2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML)

Recent findings demonstrated that deep neural networks carry out features extraction itself to id... more Recent findings demonstrated that deep neural networks carry out features extraction itself to identify the electrocardiography (ECG) pattern or cardiac arrhythmias from the ECG signals directly and provided good results compared to cardiologists in some cases. But, to face the challenge of huge volume of data to train such networks, transfer learning is a prospective mechanism where network is trained on a large dataset and learned experiences are transferred to a small volume target dataset. Therefore, we firstly extracted 78,999 ECG beats from MIT-BIH arrhythmia dataset and transformed into 2D RGB images and used as the inputs of the DenseNet. The DenseNet is initialized with the trained weights on ImageNet and fine-tuned with the extracted beat images. Optimization of the pre-trained DenseNet is performed with the aids of on-the-fly augmentation, weighted random sampler, and Adam optimizer. The performance of the pre-trained model is assessed by hold-out evaluation and stratified 5-fold cross-validation techniques along with early stopping feature. The achieved accuracy of identifying normal and four arrhythmias are of 98.90% and 100% for the hold-out and stratified 5-fold respectively. The effectiveness of the pre-trained model with the stratified 5-fold by transfer learning approach is surpassed compared to the state-of-art-the approaches and models, and also explicit the maximum generalization of imbalanced classes.

TELKOMNIKA Indonesian Journal of Electrical Engineering, 2015

Malaysian Journal of Applied Sciences, 2018

Classification is one of the most hourly encountered problems in real world. Neural networks have... more Classification is one of the most hourly encountered problems in real world. Neural networks have emerged as one of the tools that can handle the classification problem. Feed-Forward Neural Networks (FFNN's) have been widely applied in many different fields as a classification tool. Designing an efficient FFNN structure with the optimum number of hidden layers and minimum number of layer's neurons for a given specific application or dataset, is an open research problem and more challenging depend on the input data. The random selections of hidden layers and neurons may cause the problem of either under fitting or over fitting. Over fitting arises because the network matches the data so closely as to lose its generalization ability over the test data. In this research, the classification performance using the Mean Square Error (MSE) of Feed-Forward Neural Network (FFNN) with back-propagation algorithm with respect to the different number of hidden layers and hidden neurons is computed and analyzed to find out the optimum number of hidden layers and minimum number of layer's neurons to help the existing classification concepts by MATLAB version 13a. By this process, firstly the random data has been generated using an suitable matlab function to prepare the training data as the input and target vectors as the testing data for the classification purposes of FFNN. The generated input data is passed on to the output layer through the hidden layers which process these data. From this analysis, it is find out from the mean square error comparison graphs and regression plots that for getting the best performance form this network, it is better to use the high number of hidden layers and more neurons in the hidden layers in the network during designing its classifier but so more neurons in the hidden layers and the high number of hidden layers in the network makes it complex and takes more time to execute. So as the result it is suggested that three hidden layers and 26 hidden neurons in each hidden layers are better for designing the classifier of this network for this type of input data features.

Carpathian Journal of Electronic and Computer Engineering, 2018

The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of... more The bandpass filter is one of the essential blocks of every modern RF transceiver. Performance of the transceiver greatly depends on the performance of the bandpass filter. A bandpass filter designed with passive inductors suffers from some drawbacks like large chip size, low-quality factor, less tenability etc. To prevail over these constraints, an active inductor-based bandpass filter circuit has been designed in GPDK-90nm CMOS technology utilizing cadence virtuoso environment. The simulation result shows that the active inductor-based bandpass filter circuit design achieves a gain of 6.79dB, a bandwidth of 5.05 GHz and a noise figure of 3.10dB. The circuit dissipates only 3.55mW power for its operation from a single 1.5V DC supply. By avoiding bulky inductor in the design helped to attain a very small chip area of 127.704μm 2 .

Artificial Neural Networks (ANNs) are one of the most comprehensive tools for classification. I... more Artificial Neural Networks (ANNs) are one of the most comprehensive tools for classification. In this study, the performance of Feed-Forward Neural Network (FFNN) with back-propagation algorithm is used to find out the appropriate activation function in the hidden layer using MATLAB 2013a. Random data has been generated and fetched to FFNN for testing the classification performance of this network. From the values of MSE, response graph and regression coefficients, it is clear that Tan sigmoid activation function is the best choice for the image classification. The FFNN with this activation function is better for any classification purpose of different applications such as aerospace, automotive, materials, manufacturing, petroleum, robotics, communication etc because to perform the classification the network designer have to choose an activation function.

Now-a-days online interviewing is preferred as it does reduce efforts & it is cost effective as w... more Now-a-days online interviewing is preferred as it does
reduce efforts & it is cost effective as well. Long distance is one of
the causes for drawing attention in online interview. In this paper
we have described a technique of the generation of automated data
(CV) to assist the interviewer having more precise knowledge on
the person talking to online.

The main purpose of this project is to design an inverter that will enable the inversion of a DC ... more The main purpose of this project is to design an inverter that will enable the inversion of a DC power source, supplied by Photovoltaic (PV) Cells, to an AC power source that will be either used to supply a load or connected directly to the utility grid. The benefit of this project is to give access to an everlasting and pollution free source of energy. The future is looking towards alternative power sources all of which will need to be regulated in one form or another. To make this possible, a highly efficient low cost product will have to be designed. Among all the different converter designs only a few are capable of providing high power with high efficiency. This will be tested experimentally, first by computer simulation, and then in the laboratory Implementation.