Sugata Munshi

Cross-correlation aided support vector machine classifier for classification of EEG signals

Over the last few decades pattern classification has been one of the most challenging area of research. In the present-age pattern classification problems, the support vector machines (SVMs) have been extensively adopted as machine learning tools. SVM achieves higher generalization performance, as it utilizes an induction principle called structural risk minimization (SRM) principle. The SRM principle seeks to minimize the upper bound of the generalization error consisting of the sum of the training error and a confidence interval. SVMs are basically designed for binary classification problems and employs supervised learning to find the optimal separating hyperplane between the two classes of data. The main objective of this paper is to introduce a most promising pattern recognition technique called cross-correlation aided SVM based classifier. The idea of using cross-correlation for feature extraction is relatively new in the domain of pattern recognition. In this paper, the proposed technique has been utilized for binary classification of EEG signals. The binary classifiers employ suitable features extracted from crosscorrelograms of EEG signals. These cross-correlation aided SVM classifiers have been employed for some benchmark EEG signals and the proposed method could achieve classification accuracy as high as 95.96% compared to a recently proposed method where the reported accuracy was 94.5%.

Cross-wavelet transform as a new paradigm for feature extraction from noisy partial discharge pulses

In this work a new approach based on cross-wavelet transform towards identification of noisy Partial Discharge (PD) patterns has been proposed. Different partial discharge patterns are recorded from the various samples prepared with known defects. A novel cross-wavelet transform based technique is used for feature extraction from raw noisy partial discharge signals. Noise is a significant problem in PD detection. The proposed method eliminates the requirement of denoising prior to processing and therefore it can be used to develop an automated and intelligent PD detector that requires minimal human expertise during its operation and analysis. A rough-set theory (RST) based classifier is used to classify the extracted features. Results show that the partial discharge patterns can be classified properly from the noisy waveforms. The effectiveness of the feature extraction methodology has also been verified with two other commonly used classification techniques: Artificial Neural Network (ANN) based classifier and Fuzzy classifier. It is found that the type of defect within insulation can be classified efficiently with the features extracted from cross-wavelet spectra of PD waveforms by all of these methods with a reasonable degree of accuracy.

Correlation technique and least square support vector machine combine for frequency domain based ECG beat classification.

The present work proposes the development of an automated medical diagnostic tool that can classify ECG beats. This is considered an important problem as accurate, timely detection of cardiac arrhythmia can help to provide proper medical attention to cure/reduce the ailment. The proposed scheme utilizes a cross-correlation based approach where the cross-spectral density information in frequency domain is used to extract suitable features. A least square support vector machine (LS-SVM) classifier is developed utilizing the features so that the ECG beats are classified into three categories: normal beats, PVC beats and other beats. This three-class classification scheme is developed utilizing a small training dataset and tested with an enormous testing dataset to show the generalization capability of the scheme. The scheme, when employed for 40 files in the MIT/BIH arrhythmia database, could produce high classification accuracy in the range 95.51-96.12% and could outperform several competing algorithms.

An adaptive bacterial foraging algorithm for fuzzy entropy based image segmentation

In this paper an Adaptive Bacterial Foraging is proposed for fuzzy entropy optimization when it is applied to the segmentation of gray images. The proposed algorithm represents the improved version of classical bacterial foraging algorithm which is a newly developed stochastic optimization tool. This optimization technique is applied for optimization of the fitness function which is fuzzy entropy. Classical bacterial foraging algorithm is improved by adaptively selecting the exploitation and exploration state in chemotaxis of E.coli. bacteria. The newly developed algorithm is applied on benchmark gray images and proved to be suitable for thresholding based image segmentation.

Rough-granular approach for impulse fault classification of transformers using cross-wavelet transform

A novel approach based on information granulation using Rough sets for impulse fault identification of transformers has been proposed. It is found that the location and type of fault within a transformer winding can be classified efficiently by the features extracted from cross-wavelet spectra of current waveforms, obtained from impulse test. Results show that the proposed methodology can localize the fault within 5% of the winding length with a high degree of accuracy. The basic concepts of feature extraction using cross-wavelet transform and the method of classification of those features by rough-granular method are also explained.

Linearization of NTC Thermistor Characteristic Using Op-Amp Based Inverting Amplifier

A low cost linearizing circuit is developed, placing the NTC thermistor in a widely used inverting amplifier circuit using operational amplifier. The performance of the system is verified experimentally. A linearity of approximately ± 1% is achieved over 30 °C -120 °C. When used for a narrower span, a much better linearity of ± 0.5% is obtained. The gain of the arrangement can be adjusted over a wide range by simply varying the feedback resistance. The simplicity of the configuration promises a greater reliability, and also curtails the deterioration in the stability of performance, by reducing the cumulation of drifts in the different circuit components and devices.

A Linearization Scheme for Thermistor-Based Sensing in Biomedical Studies

Temperature is one of the basic biophysical quantity monitored for various biomedical systems. Moreover, the variation of temperature is also an important parameter which can be used for estimating other measurands, such as respiratory airflow. This paper proposes a simple operational amplifier-based astable multivibrator circuit for linearization of the characteristic of a negative temperature coefficient thermistor constituting one of the timing resistors. The circuit has been combined with a lookup table to get the unknown temperature value from multivibrator output. Moreover, the same system topology can be used as a linearizer for measurement of respiratory airflow. The performance of the composite system has been verified experimentally. A linearity of approximately ±0.75% has been achieved over 30 °C-110 °C in the case of temperature measurement and ±1.2% for airflow of 10-60 LPM. Better results can also be achieved with the introduction of interpolation algorithms, but at a higher computational and component cost. The compactness of the complete system makes it a good candidate for embedded sensing applications in biomedical systems, such as point-of-care monitoring or in sleep study.

A Hybrid Filtering Scheme for Proper Denoising of Real-time Data in Dielectric Spectroscopy

Condition monitoring of different power equipment with noninvasive and nondestructive techniques, such as dielectric spectroscopy, has become one of the important facets of power system reliability. This is increasingly becoming essential for predicting the insulation condition and probable future failure of power apparatus. Noise contamination, being one of the major problems in these measurements, should be treated judiciously. The paper proposes a simple but versatile real-time denoising scheme for this kind of signals of dielectric spectroscopy in time domain namely, polarization-depolarization current (PDC) and recovery voltage (RV), through a hybrid filtering technique using weighted median (WM) filter along with low pass digital infinite impulse response (IIR) filter. The proposed filter is able to remove real- life, random, high frequency noises as well as spurious impulse signals while preserving any step change in the signal. The performance of the proposed scheme is judged through experimentation using a real-time, automated computer controlled test setup for condition monitoring of transformers and the results show that it performs efficiently on different types of signals of dielectric spectroscopy in time domain, namely, PDC and RV.

Obstructive sleep apnoea detection using convolutional neural network based deep learning framework

This letter presents an automated obstructive sleep apnoea (OSA) detection method with high accuracy, based on a deep learning framework employing convolutional neural network. The proposed work develops a system that takes single lead electrocardiography signals from patients for analysis and detects the OSA condition of the patient. The results show that the proposed method has some advantages in solving such problems and it outperforms the existing methods significantly. The present scheme eliminates the requirement of separate feature extraction and classification algorithms for the detection of OSA. The proposed network performs both feature learning and classifies the features in a supervised manner. The scheme is computation-intensive, but can achieve very high degree of accuracy—on an average a margin of more than 9% compared to other published literature till date. The method also has a good immunity to the contamination of the signals by noise. Even with pessimistic signal to noise ratio values considered here, the methods already reported are not able to outshine the present method. The software for the algorithm reported here can be a good contender to constitute a module that can be integrated with a portable medical diagnostic system.

Modified Bacterial Foraging Optimization Technique for Vector Quantization-Based Image Compression

Vector quantization (VQ) techniques are well-known methodologies that have attracted the attention of research communities all over the world to provide solutions for image compression problems. Generation of a near optimal codebook that can simultaneously achieve a very high compression ratio and yet maintain required quality in the reconstructed image (by achieving a high peak-signal-to-noise-ratio (PSNR)), to provide high fidelity, poses a real research challenge. This chapter demonstrates how such efficient VQ schemes can be developed where the near optimal codebooks can be designed by employing a contemporary stochastic optimization technique, namely bacterial foraging optimization (BFO), that mimics the foraging behavior of a common type of bacteria, Escherichia coli, popularly known as E. coli. An improved methodology is proposed here, over the basic BFO scheme, to perform the chemotaxis procedure within the BFO algorithm in a more efficient manner, which is utilized to solve this image compression problem. The codebook design procedure has been implemented using a fuzzy membership-based method, and the optimization procedure attempts to determine suitable free parameters of these fuzzy sets. The usefulness of the proposed adaptive BFO algorithm, along with the basic BFO algorithm, has been demonstrated by implementing them for a number of benchmark images, and their performances have been compared with other contemporary methods, used to solve similar problems.