A. Immanuel Selvakumar

Issues and research on foetal electrocardiogram signal elicitation

Abstract The existence of Electrocardiography (ECG) came to light over a century ago, yet the acquisition and elicitation of non-invasive foetal electrocardiogram (fECG) is still in infancy despite mammoth advances in clinical electrocardiography, advanced Biomedical signal processing techniques and fast growing engineering technology. The acquisition of foetal ECG becomes a challenging task since it is perilous for a direct contact over the foetus. Moreover, the non-invasive abdominal ECG (aECG) measurements obtained over the surface of a maternal abdomen contains several bioelectric potentials like maternal heart activity, foetal heart activity, maternal muscle activity, foetal movement activity, generated potentials by respiration and stomach activity, and noise (thermal noise, noise generated from electrode-skin contact). The strong Maternal Electrocardiogram (mECG) along with the weak fECG in the recordings obtained from a mother is overlapping in time as well as in frequency. Hence separation of these signals cannot be accomplished by simple windowing or filtering. Over the years numerous researchers have put enormous efforts in signal processing and biomedical engineering for humanizing the foetal ECG acquisition techniques used by the physiologists. The fECG signal is a vital information source to assist physicians for precise timely decisions during labor. Some of the recently developed algorithms by researchers have given remarkable results for fECG acquisition. The focus of this paper is to review these existing algorithms for non-invasive detection and elicitation of fECG in terms of their performance and capabilities with respect to standard databases available worldwide.

Performance Improved Iteration-Free Artificial Neural Networks for Abnormal Magnetic Resonance Brain Image Classification

Image classification is one of the typical computational applications widely used in the medical field especially for abnormality detection in Magnetic Resonance (MR) brain images. The automated image classification systems used for such applications must be significantly efficient in terms of accuracy since false detection may lead to fatal results. Another requirement is the high convergence rate which accounts for the practical feasibility of the system. Among the automated systems, Artificial Neural Network (ANN) is gaining significant positions for solving computational problems. Besides multiple advantages, there are also few drawbacks associated with the neural networks which are unnoticed for most of the applications. The main drawback is that the ANN which yields high accuracy requires high convergence time period and the ANN which are much quicker are usually inaccurate. Hence, there is a significant necessity for ANN which satisfies the criteria of high convergence rate and accuracy simultaneously. In this work, this drawback is tackled by proposing two novel neural networks namely Modified Counter Propagation Neural Network (MCPN) and Modified Kohonen Neural Network (MKNN). These networks are framed by performing modifications in the training methodology of conventional CPN and Kohonen networks. The main concept of this work is to make the ANN iteration-free which ultimately improves the convergence rate besides yielding accurate results. The performance of these networks are analysed in the context of abnormal brain image classification. Experimental results show promising results for the proposed networks in terms of the performance measures.

Modified Binary Particle Swarm optimization algorithm application to real-time task assignment in heterogeneous multiprocessor

Task assignment in a heterogeneous multiprocessor is a NP-hard problem, so approximate methods are used to solve the problem. In this paper the Modified Binary Particle Swarm Optimization (Modified BPSO) algorithm and Novel Binary Particle Swarm (Novel BPSO) Optimization are applied to solve the real-time task assignment in heterogeneous multiprocessor. The problem consists of a set of independent periodic task, which has to be assigned to a heterogeneous multiprocessor without exceeding the utilization bound. The objective is to schedule maximum number of tasks with minimum energy consumption. The execution times and deadlines of the tasks are assumed to be known. Here Modified BPSO performance is compared with Novel BPSO and Ant Colony Optimization algorithm (ACO). Experimental results show that Modified BPSO performs better than Novel BPSO and ACO for consistent utilization matrix and ACO performs better than Modified BPSO and Novel BPSO for inconsistent utilization matrix.

MPPT based stand-alone water pumping system

Renewable energy sources are becoming a viable substitute for conventional energy sources due to increases in worlds energy demand and scarce resources. Solar pump operated with AC drive offer better choice in terms of size, ruggedness, efficiency and maintainability. In this work, dc power from solar panel is boosted and fed to an inverter which gives ac output. Inverter drives the motor coupled to the water pump. To get the maximum power available at any instant an MPPT controller is used to control the converter. Of different types of MPPT algorithms artificial intelligence (AI) techniques are popular. Artificial neural networks (ANNs) & fuzzy logic (FL) two different types of AI techniques that are used to design the MPPT controller for PV system. In this proposed work, depending on solar radiation and temperature, the MPPT controller gives optimized duty cycle. Neural network and fuzzy logic are two MPPT controllers, simulated to give optimum duty cycle. These MPPT controllers are compared based on the power obtained from the boost converter. Simulation results are also presented

Distance metric-based time-efficient fuzzy algorithm for abnormal magnetic resonance brain image segmentation

Image segmentation is one of the significant computational applications of the biomedical field. Automated computational methodologies are highly preferred for medical image segmentation since these techniques are immune to human perception error. Artificial intelligence (AI)-based techniques are often used for this process since they are superior to other automated techniques in terms of accuracy and convergence time period. Fuzzy systems hold a significant position among the AI techniques because of their high accuracy. Even though these systems are exceptionally accurate, the time period required for convergence is exceedingly high. In this work, a novel distance metric-based fuzzy C-means (FCM) algorithm is proposed to tackle the low-convergence-rate problem of the conventional fuzzy systems. This modified approach involves the concept of distance-based dimensionality reduction of the input vector space that substantially reduces the iterative time period of the conventional FCM algorithm. The effectiveness of the modified FCM algorithm is explored in the context of magnetic resonance brain tumor image segmentation. Experimental results show promising results for the proposed approach in terms of convergence time period and segmentation efficiency. Thus, this algorithm proves to be highly feasible for time-oriented real-time applications.

Fuzzy-logic based hill-climbing method for maximum power point tracking in PV systems

A fuzzy-logic controller based on hill-climbing method for maximum power point tracking of photovoltaic system is proposed. The features of maximum power point tracker based on conventional hill-climbing algorithm are studied. The new controller incorporates the advantages of hill-climbing algorithm and eliminates its drawbacks. This technique provides fast and accurate convergence to the maximum power point during steady-state and gradually changing weather conditions. The hill-climbing algorithm is implemented using 17 fuzzy control rules. Simulation results are provided to illustrate the validity of the proposed hill-climbing based fuzzy-logic controller.

Superior foetal electrocardiogram signal elicitation using a novel artificial intelligent Bayesian methodology

BANFIS - a novel hybrid methodology.Faster convergence rate within 10s.Above par SNR evaluation methodology - towards efficiency of the algorithm.State of art performance evaluation - towards extorted signal quality.Premier quality of foetal signals elicitated - precious for STAN. Innumerable casualties due to intrauterine hypoxia are a major worry during prenatal phase besides advanced patient monitoring with latest science and technology. Hence, the analysis of foetal electrocardiogram (fECG) signals is very vital in order to evaluate the foetal heart status for timely recognition of cardiac abnormalities. Regrettably, the latest technology in the cutting edge field of biomedical signal processing does not seem to yield the desired quality of fECG signals required by physicians, which is the major cause for the pathetic condition. The focus of this work is to extort non-invasive fECG signal with highest possible quality with a motive to support physicians in utilizing the methodology for the latest intrapartum monitoring technique called STAN (ST analysis) for forecasting intrapartum foetal hypoxia. However, the critical quandary is that the non-invasive fECG signals recorded from the maternal abdomen are affected by several interferences like power line interference, baseline drift interference, electrode motion interference, muscle movement interference and the maternal electrocardiogram (mECG) being the dominant interference. A novel hybrid methodology called BANFIS (Bayesian adaptive neuro fuzzy inference system) is proposed. The BANFIS includes a Bayesian filter and an adaptive neuro fuzzy filter for mECG elimination and non-linear artefacts removal to yield high quality fECG signal. Kalman filtering frame work has been utilized to estimate the nonlinear transformed mECG component in the abdominal electrocardiogram (aECG). The adaptive neuro fuzzy filter is employed to discover the nonlinearity of the nonlinear transformed version of mECG and to align the estimated mECG signal with the maternal component in the aECG signal for annulment. The outcomes of the investigation by the proposed BANFIS system proved valuable for STAN system for efficient prediction of foetal hypoxia.

Model-Based MPPT for Shaded and Mismatched Modules of Photovoltaic Farm

Photovoltaic modules (PVM) of same rating and manufacturer have unique characteristics in real time due to manufacturing dispersion. This unique maximum power point (MPP) of each module is tracked by the proven distributed module level maximum power point tracking (DMMPPT). Since the conventional slow iterative MPPT algorithms in DMMPPT fail under the rapidly varying environmental conditions, the model-based algorithm (MBA) is used in this paper due to its swift tracking and character leaning nature. By identifying the actual behavior of each module, the new MBA is implemented here in compensation power dc–dc converter for the distributed model-based maximum power point tracking (CPDC-DMBMPPT). The flyback converter-based CPDC converter, which provides the current or voltage compensation is used to operate the module in its own MPP without any compromise even at the mismatched or partial shaded condition. As the individual module parameters used in MBA are estimated during the installation, the costly high precision measuring instruments are avoided in the consumer site. Also, due to the LabVIEW-based centralized control, updates in the MBA becomes easy without changing the individual controllers in Photovoltaic farms. The proposed methodology and its proven outcomes are discussed through the simulation and hardware outputs.

Fault detection in PV systems

A practical fault detection approach for PV systems intended for online implementation is developed. The fault detection model here is built using artificial neural network. initially the photovoltaic system is simulated using MATLAB software and output power is collected for various combinations of irradiance and temperature. Data is first collected for normal operating condition and then four types of faults are simulated and data are collected for faulty conditions. Four faults are considered here and they are: Line to Line faults with a small voltage difference, Line to line faults with a large voltage difference, degradation fault and open-circuit fault. This data is then used to train the neural network and to develop the fault detection model. The fault detection model takes irradiance, temperature and power as the input and accurately gives the type of fault in the PV system as the output. This system is a generalized one as any PV module datasheet can be used to simulate the Photovoltaic system and also this fault detection system can be implemented online with the use of data acquisition system.

Automated multi-level pathology identification techniques for abnormal retinal images using artificial neural networks

Aim To automatically classify abnormal retinal images from four different categories using artificial neural networks with a high degree of accuracy in minimal time to assist the ophthalmologist in subsequent treatment planning. Methods We used 420 abnormal retinal images from four different categories (non-proliferative diabetic retinopathy, central retinal vein occlusion, central serous retinopathy and central neo-vascularisation membrane). Green channel extraction, histogram equalisation and median filtering were used as image pre-processing techniques, followed by texture-based feature extraction. The application of Kohonen neural networks for pathology identification was also explored. Results The approach described yielded an average classification accuracy of 97.7% with ±0.8% deviation for individual categories. The average sensitivity and the specificity values are 96% and 98%, respectively. The time taken by the Kohonen neural network to achieve these accurate results was 300±40 s for the 420 images. Conclusion This study suggests that the approach described can act as a diagnostic tool for retinal disease identification. Simultaneous multi-level classification of abnormal images is possible with high accuracy using artificial neural networks. The results also suggest that the approach is time-efficient, which is essential for ophthalmologic applications.