Search

An exact and quick finding of Covid-19 patients plays a major part in the initial period of medicinal treatment and prevention. Automatic recognition of COVID-19 cases utilizing lung X-ray images may help lessen the effect of this infection on human civilization. In recent years, computer vision is the major solution for diagnosing the covid-19 disease by utilizing X-ray images. Besides, many researchers had presented efficient Artificial Intelligence (AI)methods for identifying Covid-19 disease. Nevertheless, the accuracy and time consumption of the model is further to be improved. Thus, in this work, we proposed a novel Xception network with an optimized convolution neural network (OCNN). The CNN architecture is enhanced by using the adaptive seagull optimization (ASO) algorithm. There are three stages of the approach: pre-processing, feature extraction, and classification. At first, a median filter is applied to each image to reduce the noise present in the input image. Then, the noise-free images are fed to the Xception network to extract the features of the images. A convolution network is used to classify an image as positive or negative after feature extraction. The performance of the proposed approach is analyzed based on various metrics and performance compared with other techniques. [ABSTRACT FROM AUTHOR]

Brain signal analysis plays a significant role in attaining data related to motor activities. The parietal region of the brain plays a vital role in muscular movements. This approach aims to demonstrate a unique technique to identify an ideal region of the human brain that generates signals responsible for muscular movements; perform statistical analysis to provide an absolute characterization of the signal and validate the obtained results using a prototype arm. This can enhance the practical implementation of these frequency extractions for future neuro-prosthetic applications and the characterization of neurological diseases like Parkinson's disease (PD). To play out this handling method, electroencephalogram (EEG) signals are gained while the subject is performing different wrist and elbow movements. Then, the frontal brain signals and just the parietal signals are separated from the obtained EEG signal by utilizing a band pass filter. Then, feature extraction is carried out using Fast Fourier Transform (FFT). Subsequently, the extraction process is done by Daubechies (db4) and Haar wavelet (db1) in MATLAB and classified using the Levenberg-Marquardt Algorithm. The results of the frequency changes that occurred during various wrist movements in the parietal region are compared with the frequency changes that occurred in frontal EEG signals. This proposed algorithm also uses the deep learning pattern analysis network to evaluate the matching sequence for each action that takes place. Maximum accuracy of 97.2% and maximum error range of 0.6684% are achieved during the analysis. Results of this research confirm that the Levenberg-Marquardt algorithm, along with the newly developed deep learning hybrid PatternNet, provides a more accurate range of frequency changes than any other classifier used in previous works of literature. Based on the analysis, the peak-to-peak value is used to define the threshold for the prototype arm, which performs all the intended degrees of freedom (DOF), verifying the results. These results would aid the specialists in their decision-making by facilitating the analysis and interpretation of brain signals in the field of neuroscience, specifically in tremor analysis in PD. [ABSTRACT FROM AUTHOR]

Link failure between the nodes takes place due to reasons like channel interference and dynamic obstacles in mobile Ad Hoc networks, that give rise to severe performance degradation. In traditional AODV, the link failure is overcome by re-routing from the source node which is considered to be a time consuming process that increases the overhead of the nodes. Also, in case of multiple link failures, there are chances for loss of data packets. The performance of the network during link failure has to be maintained dynamically in circumstances such as long data transfer such as a stream of voice data which is an underlying challenge. Another challenging problem is to feasibly detect and defend possible attacks during data transmission. In order to overcome these performance related issues, we developed the Secured Local Link Failure Recovery Algorithm (SLLFR) for Adhoc networks that establishes recovery from link failures spontaneously at the point of link breakage. In such cases, a reliable link failure recovery is the main criteria that will determine the performance of the network in terms of quality of service (QoS). The SLLFR is deployed in each node collects RREP in the RREP Buffer Table (RBT) stack in the highest order of signal strength, which gets triggered during link failures. Once a link failure is detected, the intermediate node searches for an alternate path around the faulty area by choosing the first RREP that is stacked in the RBT and establishes a new route to the intended destination for sending the data packets without any time delay. The simulation results show that the performance parameters like packet delivery ratio, throughput, average end to end delay and routing overhead are better compared to the traditional AODV and other link failure recovery techniques. [ABSTRACT FROM AUTHOR]

Microfluidic channels play a vital role in separation of analytes of interest such as bacteria and platelet cells, etc., in various biochemical diagnosis procedures including urinary tract infections (UTI) and bloodstream infections. This paper presents the multi physics computational model specifically designed to study the effects of design parameters of a microfluidics channel for the separation of Escherichia coli (E. coli) from various blood constituents including red blood cells (RBC) and platelets. A standard two inlet and a two outlet microchannel of length 805 µm with a channel width of 40 µm is simulated. The effect of electrode potentials and the effect of electrode placement along the channel length and also the levitation of electrodes from the channel wall are studied to optimize the selective particle separation throughput. Simulated results show the efficient separation of E-coli with a mean diameter 0.68 µm is achieved at low voltages (less than 20 V) when electrodes placed near to the micro channel and also noticed that the applied electric potential is inversely proportional to the number of electrodes placed along the microfluidic channel. The computer aided multi physics simulations with multiple governing parameters could be advantage in design optimization of microfluidics channels and support precise bioparticle separation for better diagnosis. [ABSTRACT FROM AUTHOR]

Bio-impedance is the manner in which a person's body opposes the applied alternating current in a unique manner. Acquired bio-impedance data is being used for diagnosis of critical diseases condition of various systems in human body. Electrode placement on body surface to sense impedance signal for various systems is unique format for each system. This paper first part focus on design of electrical bio-impedance based acquisition system with patient identification and later part deals with visualizing, analyzing and prediction of disease condition using software tools. Impedance data from 100 patients were acquired using designed system made with surface electrodes, signal generator with variable frequency and amplitude, Microcontroller board to process acquired data. Obtained impedance values at low frequencies are a good source of cardiac related information. Data obtained from bio impedance based multi-parameter monitoring system was visualized and analyzed for the prediction of age and disease condition. The validity of this system was made using intelligent diagnosis system based on IBM SPSS computational algorithms. [ABSTRACT FROM AUTHOR]

This paper proposes a new algorithm for energy-resourceful attributeaware data aggregation and routing in wireless sensor network (WSN) named AERAA-DAR. The proposed AERAA-DAR combine advantages of both query and time-driven routing. In AERAA-DAR routing design, a joint K-means unequal-clustering scheme is applied to partition the nodes into unequal into clusters. Also, two weight functions, contending and proper weights are implemented to select the suitable cluster heads. A piece of link-state learning information and threshold function are combined to serves the CH nodes in close range establish vertical chain hierarchies. Chain leaders are chosen to allocate distinct duties dynamically, and cooperatively work as agents between other nodes and the sink. A cost-driven gradient-based routing scheme is implemented to realize multi-path from the chain leaders to the sink. The routing decisions are realized in consideration of the attenuation, residual energy, and shortest route conditions of the established paths. At the cluster heads and chain leaders, discrete wavelet transform (DWT) compression is carried out on the sensed data secured within the hierarchical sections. At the shared links of the paths to the sink, intersection linear network coding (LNC) is carried out on the extracted data sent from the chain regions. The obtained simulation results disclose that by consolidating the benefits of a cost-driven gradient-based multi-path routing scheme, LNC, and DWT in a cluster-chain topology, AERAA-DAR enhances the data reporting rate and accuracy. It is disclosed equally, that AERAA-DAR considerable lessens energy consumption, latency and increases network longevity compared to similar state-of-art routing schemes [ABSTRACT FROM AUTHOR]

The capacitive coupled wireless power transfer (CCWPT) operating at megahertz (MHz) frequency is broadly considered as the promising solution for low power biomedical implants. The class E power amplifier is attractive in MHz range wireless power transfer (WPT) applications due to zero voltage switching (ZVS) and zero voltage derivative switching (ZVDS) properties. The existing design of class E amplifier is investigated only for inductive resonant coupled (IRC) WPT systems; the modelling and optimization of the class E amplifier for CCWPT systems are not deliberated with load variation. Meanwhile, the variations in the coupling distance and load are common in real time applications, which could reduce the power amplifier (PA) efficiency. The purpose of this paper is to model and optimize the class E amplifier for CCWPT systems used in MHz range applications. The analytical model of PA parameters and efficiency are derived to determine the optimal operating conditions. Also, an inductive-capacitive-inductive (LCL) impedance matching network is designed for the robust operation of the PA, which improves the efficiency and maintains required impedance compression. The maximum efficiency of the proposed design reached up to 96.34% at 13.56 MHz and the experimental results are closely matched with the simulation. [ABSTRACT FROM AUTHOR]